Skip to main content

Full text of "Deficits : how big and how bad?"

See other formats


Deficits 



Edited by 
David W. Conklin 
and 

Thomas J. Courchene 


Ontario Economic Council 


How Big and Hov 





















Deficits: 


How Big and How Bad? 


Ontario Economic Council 
Special Research Report 


ONTARIO MINISTRY OF 
TREASURY AND ECONOMICS 

JAN 1 6 1984 

niv 

LIBRARY 




©1983 


ONTARIO ECONOMIC COUNCIL 
81 Wellesley Street East 
Toronto, Ontario 
M4Y1H6 

Printed in Canada 


Canadian Cataloguing in Publication Data 

Main entry under title: 

Deficits 

(Special research report / Ontario Economic Council, 
ISSN 0225-591X) 

Proceedings of a conference held in Toronto, Mar. 
1983. 

Includes bibliographies. 

ISBN 0-7743-8497-2 

1. Debts, Public--Congresses. 2. Deficit financing— 

Congresses. I. Conklin, D.W. (David W.) II. 
Courchene, Thomas J., 1940- III. Ontario Economic 

Council. IV. Series: Special research report 
(Ontario Economic Council) 

HJ8513.D4 1983 336.3'4 83-093040-X 


These papers reflect the views of the authors and not necessari 
those of the Ontario Economic Council. The Council establis' 
policy questions to be investigated and commissions resean 
projects, but it does not ihftuence the conclusions or recomn 
dations of authors. The decision to sponsor publication of th 
papers was based on their competence and relevance to 
public policy. 


HC 
117 
. S 741 
. D31 3 


Contents 


Preface 

INTRODUCTION 1 

by David Conklin 

OVERVIEW OF THE DEFICIT DEBATE 12 
by David Conklin and Adil Sayeed 

GOVERNMENT DEFICITS, INFLATION, AND FUTURE GENERATIONS 
by Franco Modigliani 

Discussion 71 

THE ONTARIO BUDGET DEFICIT: A CAUSE FOR CONCERN? 78 
by D.A.L. Auld 

Comments by W.R. White 107 

Discussion 114 

DEFICITS AND THE ECONOMY TO 1990: 

PROJECTIONS AND ALTERNATIVES 116 
by D.P. Dungan and T.A. Wilson 

Comments by Leo de Bever 148 

WHAT CAN MACROECONOMIC THEORY TELL US ABOUT THE WAY 
DEFICITS SHOULD BE MEASURED? 150 
by Michael Parkin 

Comments by Brian L. Scarfe 177 

Comments by John Bossons 181 

WHAT DOES THE PUBLIC THINK ABOUT DEFICITS? 

WHAT DOES BAY STREET THINK ABOUT DEFICITS? 189 
by Ian McKinnon 


55 


FISCAL POLICY 'CROWDING-OUT' OF PRIVATE INVESTMENT IN AN OPEN 
ECONOMY: THE CASE OF CANADA 215 
by R.G. Wirick 

Comments by Peter Andersen 249 
Discussion 253 

DEFICITS AND CAPITAL MARKETS 261 
by John Grant 

Comments by Roger Keane 279 

GOVERNMENT DEFICITS: HISTORICAL ANALYSIS AND PRESENT POLICY 
ALTERNATIVES 284 
by John McCallum 

Comments by Robert W. Baguley 317 
Discussion 319 

FISCAL DISCIPLINE AND RULES FOR CONTROLLING THE DEFICIT: 

SOME UNPLEASANT KEYNESIAN ARITHMETIC 323 
by Neil Bruce and Douglas D. Purvis 

Comments by William Mackness 340 

RAPPORTEUR'S REMARKS 346 
by David Laidler 

Discussion 358 


MEMBERS OF THE ONTARIO ECONOMIC COUNCIL 361 


Preface 


Part of the Council's mandate, as outlined in the Ontario Economic Council 
Act (1970), is to 'create an awareness and public understanding of provin¬ 
cial socio-economic issues by holding seminars and conferences'. Given the 
rising levels of provincial and federal deficits, the Council had little diffi¬ 
culty in selecting Deficits: How Big and How Bad? as the topic of our 
1983 Conference. Motivating our selection were two other important as¬ 
pects of deficits: 1) large deficits will be with us for a while yet and 
2) there is a wide spectrum of views relating to the seriousness of the 
current deficit levels. 

Our objective in mounting this Conference was to inform rather than 
to persuade. Accordingly, the Council strove to ensure that the persons 
chosen to present the papers covered a wide spectrum of views. Most but 
not all of the paper presenters were academics and most, but not all, of 
the discussants were practitioners. 

We wish to thank the Council staff both for running the conference 
and overseeing the production of the volume. 

Needless to say the views expressed in this volume do not necessarily 
coincide with those of the Council. On a later occasion the Council may 
issue a position paper on deficits. We take this opportunity to express 
our gratitude to all those who participated in the Conference and we hope 
that the volume of studies and commentaries will provide a timely and 
useful focal point for a continuing public discussion of government 
deficits. 



Thomas J. Courchene 
Chairman 

Ontario Economic Council 



Digitized by the Internet Archive 
in 2018 with funding from 
Ontario Ministry of Finance 


https://archive.org/details/deficitshowbighoOOconk 


Introduction 

David Conklin* 


The papers in this volume were presented at a Conference sponsored by 
the Ontario Economic Council, March 8-9, 1983, in Toronto, Canada. The 
question, Deficits: How Big and How Bad? was interpreted broadly and 
from many perspectives. Calculations and computer projections provided 
estimates of Canadian federal deficit levels at various rates of capacity 
utilization and with various policy options over the years 1930-1990. 
Special attention was paid to the nature and role of deficits in the 
province of Ontario. Authors analysed the relationships between deficits, 
inflation, interest rates, monetary policy, capital markets, intergenerational 
transfers, international capital flows, and exchange rates. The opinions of 
the public and of financial leaders were examined. Recommendations con¬ 
cerning fiscal policy were advocated. In this process, deficits provided a 
focus for examining our economic system as a whole: for considering the 
recent history of the economy, for predicting future economic develop¬ 
ments, and for formulating policy prescriptions. Furthermore, the dis¬ 
cussion of deficits involved the exploration of much of our current under¬ 
standing of economic theory. Throughout the Conference, appointed 
discussants provided comments on each paper. Questions and opinions 
were presented from the floor, and these participants' remarks have been 
included in the volume to the extent that they cast new light on the 
issues. The purpose of this chapter is to sketch briefly some of the 
themes and methodologies pursued in the various papers. 

Conklin and Sayeed compiled a summary of the literature, an Over ¬ 
view of the Deficit Debate , which was circulated to participants prior to 
the Conference. The public accounting framework was described. Cana¬ 
dian ratios of deficits/GNP were compared with those of other countries. 


Research Director, Ontario Economic Council 





Full-employment budgets, structural deficits, and inflation adjustments 
were analysed, together with specific estimates for Canada. Separation of 
capital spending from current expenditures was discussed. Attention was 
given to Keynesian theory, crowding-out, international capital flows, and 
the Ricardian equivalence theorem. The interactions among deficits, mone¬ 
tary policy, and inflation were explored. The importance of forecasting 
techniques in the prescription of policies was emphasized. 

The keynote speaker at the Conference was Professor Franco 
Modigliani, who entitled his remarks, Government Deficits, Inflation and 
Future Generations . Modigliani focused on three main points: 

The first is that current deficits are not a major cause of inflation; on 
the contrary, I shall argue that inflation is a major cause of deficits. 
The causation goes from stagflation to deficit, not the other way around. 

Next, I want to argue that deficits are, nonetheless, a bad thing. 
They may not be bad in the short run, under some circumstances, but if they 
continue in the long run, they have serious consequences. 

Finally, I propose to establish that deficits are not the cause of our 
current trouble, despite the fact that they are bad, for the simple reason 
that there are scarcely any deficits. That is, in terms of that deficit 
which is harmful, there are very few deficits of any significance at this 
time (page 55). 

In supporting these points, Modigliani emphasized that there is no 
mechanical connection between running a deficit and creating money. The 
reason that deficits are bad is not that they cause inflation, but rather 
that they can crowd out desirable investment. Modigliani compared four 
common views of deficits: the 'naive' view that deficits are a burden 
because government debt is just like the debt of a family; the 'naive 
no-burden' view that we owe the government debt to ourselves, so that it 
is a simple transfer; the 'sophisticated burden' view that deficits reduce 
the amount of capital we bequeath to future generations; and the 'super- 
sophisticated no-burden' view that tax financing and debt financing are 
equivalent because citizens, realizing that current deficits mean higher 
taxes in the future, increase current saving by the amount of the deficits. 
Modigliani rejected the latter argument and supported the 'sophisticated 
burden' view that deficits crowd out tangible capital, reducing the stock of 
capital below what it otherwise would be, and transferring the burden to 
future generations. Nevertheless, Modigliani emphasized that the deficit 
that creates crowding-out is the rise in the real debt rather than the cash 
deficit, and that until 1982 there were no deficits in this sense. Large 
measured deficits, he argued, have been caused by stagflation and high 


2 




nominal interest payments on outstanding debt. 

Professor Doug Auld examined the question, The Ontario Budget 
Deficit: A Cause for Concern? As a percentage of total provincial expen¬ 
diture, current deficits are not particularly high compared with deficits of 
previous years. 'In absolute terms, however, deficits have grown con¬ 
siderably in recent years, reflecting the growth of the public sector' (page 
81). Auld emphasized the impact of inflation on debt charges and the 
impact of unemployment on both expenditures and receipts. After con¬ 
sidering policy-induced changes in the Ontario deficit, as opposed to those 
caused by variations in the level of economic activity, he concluded that 
'on balance, discretionary policy over this period appears to have been 
counter-cyclical' (page 93). Auld examined the issue of how capital 
spending should be dealt with in the budget. He discussed the argument 
that capital spending should be financed over the lifetime of the relevant 
assets, and he indicated the extent to which Ontario deficits would be 
reduced by such an accounting approach. Auld's calculations indicated 
that total debt as a percentage of GPP, or in real per capita terms, grew 
until the late 1970s and has since declined. Consequently, he was optim¬ 
istic that 'there is very little reason to be concerned about Ontario's debt 
at this time. The precipitously high interest rates of 1979-82 have created 
problems of debt management, but these problems should disappear with a 
return to more moderate interest rate levels' (page 100). Auld does 
caution however that if the working definition of full employment is above 
6.5 per cent, 'it may be necessary to sustain a small structural deficit or 
realign expenditures and revenues' (page 100). 

In the paper, Deficits and the Economy to 1990: Projections and 
Alternatives , Professors Dungan and Wilson used the econometric FOCUS 
model of the Canadian economy in order to develop projections based upon 
several different sets of policies. Important assumptions were made at the 
outset, both in terms of specific relationships among economic variables 
and also in terms of exogenous parameters. For example, they assumed 
that the Bank of Canada would 'conduct monetary policy to accommodate 
about 6 per cent inflation with about 4 per cent real growth' (page 118). 
They drew 'heavily upon a long-term trend projection for the United States 
that was released by Data Resources Inc. in December 1982' (page 117). 
Their base case projection resulted in a federal deficit of $28 billion in 
1983 on a National Income Accounts basis, gradually declining to $18 billion 
in 1990. They noted that these absolute figures give a misleading picture. 


3 






The inflation component of interest payments on the public debt makes up 
a large part of each number. In addition, the $18 billion figure for the 
1990 deficit must be considered in the context of a nominal GNP that will 
have more than doubled over the projection interval. As a percentage of 
GNP, the deficit will have fallen substantially by 1990. They concluded: 


The existence of considerable excess supply in the near term, as evidenced 
by both unemployment and capacity utilization rates, indicates that there 
is room for near-term fiscal stimulus with little danger of significant 
short-term crowding-out through financial market or price level effects, 
(page 124) 

Dungan and Wilson stressed, however, that such stimulus should be 
temporary, so that it would have no significant effects on the medium-term 
structural deficit. 

Dungan and Wilson then conducted a series of policy experiments. 
This approach permitted a quantitative analysis of the unemployment-infla¬ 
tion trade-off through a series of fiscal policy experiments, with a variety 
of companion monetary policies. They examined the impact of temporary 
expenditure increases, the indexation of capital cost allowances, and the 
implementation of a new method for determining unemployment insurance 
premiums. These experiments reinforced their base case conclusion that 'a 
carefully designed fiscal stimulus could help to speed recovery from the 
recent deep recession' (page 137). It should be noted that each such 
experiment had a temporary positive impact on inflation, but over the 
longer term some experiments yielded a reduced rate of inflation. 

Professor Michael Parkin examined the question, What Can Macroeco ¬ 
nomic Theory Tell Us About the Way Deficits Should Be Measured? He re¬ 
viewed various ways in which the conventional accounting deficit may be 
adjusted to arrive at alternatively defined deficits. In this analysis, he 
pointed out that unfunded pension liabilities are a significant factor that 
should be considered in correcting the conventionally measured deficit. 

The choice of a particular definition must depend 'on which of various 
questions concerning macroeconomic performance we wish to address, and 
also on which model of economic behaviour best describes the world in 
which we live' (page 152). A major conclusion reached by Parkin was 
that: 


macroeconomic theory is much less clear-cut in its implications for how the 
deficit should be measured than most economists seem ready to believe. 
However, to the extent that macroeconomic theory does offer guidance, it 
suggests that we face a potentially serious problem arising from an evol- 


4 




ving deficit that has been too big, is too big, and promises to remain too 
big...The rosy picture painted by the real full-employment deficit calcula¬ 
tions is clearly misleading and probably seriously so. (page 153) 

Parkin examined the relationship between the deficit and aggregate 
demand, and emphasized the role of changes in the deficit rather than 
simply the size of the deficit itself. Concerning the impact of the deficit 
on aggregate supply, Parkin stressed the importance of unanticipated fiscal 
stimulus, stating that 'no particular measure of the deficit has any major 
role to play in indicating what determines the level of economic activity. 
It is only the current level of the deficit relative to its previously 
expected (cyclically adjusted) level that is of concern' (page 165). When 
considering the deficit and real interest rates, Parkin noted Canada's 
special circumstances as a small, open economy able to import capital. His 
detailed discussion of the relationships between deficits and inflation 
focused on the concept of a deficit process over time and the growth of 
the money supply over time. Furthermore, Parkin emphasized the dangers 
of over-correcting the cyclically adjusted or structural deficit when con¬ 
sidering the impact of the deficit on inflation. This risk of over¬ 
correction occurs because the traditional relationship between unemployment 
and inflation may no longer hold. There are two reasons why the rate of 
unemployment that is compatible with zero inflation may have risen. First, 
the amplitude of cyclical fluctuations has increased. The inflation rate has 
been more variable, and this has shifted the short-run trade-off between 
inflation and unemployment. Second, in times of exceptional structural 
change and reallocation of productive factors the natural rate of unemploy¬ 
ment may be higher than in more normal times. Parkin criticized 
unemployment-adjusted deficits that are based on 7 or T\ per cent un¬ 
employment rates, and suggested 10 per cent unemployment as a more 
appropriate rate for such adjustments. Finally, Parkin recommended 
further research on these issues. 

Mr. Ian McKinnon, Vice-President of Decima Research, examined the 
questions What Does the Public Think About Deficits? What Does Bay Street 
Think About Deficits? The former question was answered by analysing 
Decima's national public opinion quarterly surveys, particularly that of 
September, 1982. The latter question was answered through interviews 
with leaders of the Canadian financial community, specially conducted for 
this paper. 

McKinnon concluded that, for Canadians, 'the economy now dominates 


5 




the political agenda' (page 190) and that there has been 'a precipitous de¬ 
cline in the public's assessment of the economy' (page 190). That is, the 
federal government was perceived as having a responsibility to act, parti¬ 
cularly to decrease interest rates. The survey 

certainly does not depict a Canadian public given to Keynesian visions of 
governments spurring aggregate demand through deficit spending. Instead, it 
strongly indicates a 'public household' view of what governments' actions 
should be. As individuals should try to economize and save in hard times, 
so too should governments. (page 195-6) 

McKinnon's survey of financial leaders revealed that they under¬ 
estimated this fiscal frugality of the average Canadian. Within the finan¬ 
cial community, there was no consensus in regard to altering the size of 
the deficit. The fear of crowding-out as a result of the large deficits 
seemed prevalent, particularly when future recovery would bring private 
requirements to the capital markets. 

Professor Ron Wirick examined Fiscal Policy 'Crowding-Out' of Private 
Investment in an Open Economy: The Case of Canada . Wirick began with 
'the fact that in a creditworthy country such as Canada, gaps between 
domestic savings and investment can be financed by virtually unlimited 
international capital flows at an exogenous world interest rate' (page 
215). Wirick focused his principal attention on the question of whether 
fiscal policy crowds out private investment by causing 'an increase in 
borrowing costs to Canadian corporations' (page 216). The bulk of his 
paper consisted of calculating and analysing three components of the gap 
between the Canadian corporate and U.S. treasury bond rates. He labeled 
these components the corporate premium, the currency-related differential, 
and the sovereign-risk premium. He concluded that 'there is absolutely no 
indication of any recent increase in the sovereign-risk premium, despite 
the ballooning deficits of the Canadian federal government' (page 228). He 
also argued that stimulative fiscal policy 'could, in principle, reduce the 
magnitude of the corporate interest premium' (page 220). The fluctuations 
between Canadian and American bond yields have been caused, to a large 
extent, by differences in the expected inflation rates in the two countries. 
Consequently, comparing yields in the two countries one finds less varia¬ 
tion in the real than in the nominal differences. Therefore, 'there is very 
little reason to believe that bond-financed fiscal deficits of the magnitude 
presently contemplated will exert any upward pressure on domestic real 
interest costs' (page 232). 


6 





Wirick cautioned, however, that deficits could put upward pressure 
on the Canadian exchange rate if they were financed through capital 
inflows. With such a currency appreciation, Canadian exports would 
decrease and imports would increase. In this way, deficits could crowd 
out net exports rather than investment. Wirick's analysis of this type of 
crowding out was based largely on the theory developed originally by 
Fleming and Mundell. 

Wirick concluded that 

there is little doubt in the author's mind that a package of fiscal stim¬ 
ulation could be designed, even one of substantial size, that would have no 
adverse effect on domestic interest rates or investment. Furthermore, if 
the Bank of Canada cooperated by preventing any significant appreciation of 
the Canadian dollar, there should not be any major crowding-out of the net 
export sector. In short, it should be possible for the fiscal and monetary 
authorities to provide a large aggregate demand stimulus to the Canadian 
economy, if this were deemed desirable . (page 244) 

Like many other'authors, Wirick saw the demand stimulus of greater 
deficits as carrying with it the risk of renewed inflation. This spectre of 
another burst of inflation leading to a subsequent, possibly even worse, 
recession haunted much of the Conference, and Wirick typically stressed 
the need to avoid the twin risks of 'a downward spiral of plunging 
business and consumer confidence and an upward spiral of escalating price 
expectations' (page 245). 

In his examination of Deficits and Capital Markets , Dr. John Grant 
emphasized the financial community's great fear of deficits. He suggested 
several reasons for this fear. Most important is a concern that deficits 
will be financed through the creation of money, thereby causing inflation. 
Expectations of price stability cannot be restored easily or quickly: 'A 
government that has once destroyed social trust in price stability must 
work painfully and hard to restore it' (page 262). Grant pointed to the 
threat of crowding-out as a second cause for dislike of deficits. Some 
private borrowers may be crowded out just as effectively as they would be 
by a tax increase. Grant stressed that monetary policy is an important 
determinant of the nature and extent of crowding-out, and that 'there has 
been no simple relationship between deficits and the rate of monetary 
expansion' (page 272). In addition, the ability of Canadian borrowers to 
have access to foreigners' savings can modify tendencies for deficits to 
crowd out private investment. 

Some members of the financial community oppose deficits because they 


7 




prefer current to deferred taxation, perhaps because of the interest bur¬ 
den. Some may expect to receive less government services than they pay 
for. Overall, however, Grant saw the inflation theme as the important 
element in the financial community's concern. The high level of current 
real interest rates is a result of the high risk premium demanded because 
of expectations of renewed inflation. Grant expressed his view that gov¬ 
ernment deficits will fall in the late 1980s, as demographic changes reduce 
the demand for government services relative to the size of the economy. 
Returning to a stable price level will be a difficult, but important process. 
In this, Grant felt that steadfast monetary restraint will be essential. 

In his paper. Government Deficits: Historical Analysis and Present 
Policy Alternatives , Professor John McCallum addressed a series of 
questions: 

What was the stance of fiscal policy, and how much did it contribute to (or 
alleviate) the Depression (of the 1930s)? How much less severe would the 
Depression have been if today’s automatic stabilizers had been in place 
and/or today’s approach to discretionary fiscal policy had then prevailed? 
Or, to turn the question around, how much worse would our present recession 
be if we had followed 1930s-style fiscal policy in 1982? (page 284) 

With his computer simulations, McCallum concluded that, 'had the 
automatic stabilizers of today existed during the thirties, the muliplier and 
hence the depth of the Depression would have been reduced by just over 
40 per cent' (page 294). Furthermore, his calculations suggested that 'the 
shock to the Canadian economy in 1982 may have been of the same order of 
magnitude as the shocks of the early thirties and that the difference in 
overall fiscal policy has been a key factor in explaining why the economy 
has not this time fallen into major depression' (page 295). McCallum 
analysed federal fiscal policy throughout the 1970s and concluded that it 
exerted 'a helpful stabilizing influence' (page 305) and that 'a key point to 
emphasize is that over the period 1971-81 discretionary federal fiscal policy 
contributed at least as much to stability as did the automatic stabilizers of 
all levels of government combined' (page 307). However, the extremely 
severe 1982 recession was not met with appropriately large deficits. In 
fact, 'the small increase in the adjusted federal deficit in 1982 was mainly 
the result of a non-discretionary increase in real interest rates' (page 
305). Consequently, McCallum recommended 'a temporary increase in 
automatic stabilizers or a fiscal stimulus whose size is made contingent on 
the strength of the economy over the coming months' (page 309). 


8 




Professors Neil Bruce and Doug Purvis discussed Fiscal Discipline and 
Rules for Controlling the Deficit: Some Unpleasant Keynesian Arithmetic . 

They stressed the need for 'commitment in the form of fiscal policy rules. 
These are promises to pursue certain courses of action, particularly with 
respect to the deficit, even when it is politically opportune to renege' 
(page 323). They advocated fiscal prudence, meaning that 'in some sense 
and in some time frame, the government must balance its budget' (page 
326). The common argument that the budget should be balanced over the 
cycle, with deficits in recession offset by surpluses in expansions relies 
upon a combination of automatic stabilizers and discretionary policies. 
Bruce and Purvis doubted that this goal could be attained through reliance 
on discretionary policies. They noted that the cycle consists of complex 
and uncertain fluctuations, and that policymakers cannot easily predict 
future deficits against which the present position should be balanced. 
During expansions, a strong temptation exists to forgo surpluses. 
Furthermore, lags occur in knowing the economy's position and in deciding 
what to do about it. For a variety of reasons, they 'believe that fiscal 
discretion has harmful effects and that adopting and following fiscal rules 
can lead to an improvement in macroeconomic performance' (page 328). 

Bruce and Purvis advocated the establishment of a tax and expendi¬ 
ture - structure that would balance the budget over the medium term but 
would automatically be counter-cyclical in response to short-run fluctua¬ 
tions. They explained the mechanics of various types of automatic stabi¬ 
lizers, indicating the differences among proportional, integral, and deriva¬ 
tive automatic stabilizers. They recommended greater attention be paid to 
the role of integral stabilizers, through which 'the stimulus (contraction) 
becomes greater the longer the economy remains away from its normal or 
average state;' and derivative stabilizers, through which stimulus would 
occur 'if the economy grew more slowly than trend, regardless of whether 
the output gap were positive or negative' (page 330). 

Bruce and Purvis discussed the proper role of fiscal policy in achiev¬ 
ing an optimal pace of transition from high to low inflation. They stressed 
that the rate of monetary expansion will likely be affected by the rate of 
growth of the nominal value of liquid government liabilities. The increas¬ 
ing ratio of government debt to GNP has probably affected real interest 
rates and has reduced public confidence in the policies being pursued. 
Consequently, they argued that any additional fiscal stimulus at the 
present time should automatically be self-eliminating as recovery develops. 


9 





Of crucial importance is the commitment to reduce the deficit with re¬ 
covery. Whether or not the government fulfills this commitment will 
seriously affect the confidence of the private sector. 

Professor David Laidler acted as rapporteur for the conference. He 
emphasized that the major cause for concern about deficits is the fear that 
deficits will lead to higher inflation. While current levels of Canadian 
deficits have not been produced by irresponsible fiscal policies, any sub¬ 
stantial increases in deficits could be dangerous, in that they could re¬ 
kindle inflation. In contemplating the burden imposed by government 
debt, Laidler noted that the private sector may wish to hold a roughly 
constant fraction of its wealth in government debt. As the economy 
grows, the government must run a deficit if that demand is to be satis¬ 
fied. The debt-to-income ratio is an important statistic from this 
perspective. 

In describing the current state of the economy, Laidler argued that 
excessively tight monetary policy underlay the stagflation which, in turn, 
led to high deficits. He criticized the failure of federal authorities to 
coordinate fiscal and monetary policies properly. In analysing the deficits, 
Laidler agreed with inflation adjustments as an accounting procedure that 
is useful in order to calculate the rate at which real government debt is 
being accumulated by the public. However, he felt that the cyclical ad¬ 
justments to the measured deficit were more controversial. In particular, 
he believed that unemployment rates of between 7 and 8 per cent were not 
attainable over the medium-term horizon. Consequently, fiscal policies 
based on those over-optimistic forecasts could rekindle inflation and lead to 
even worse recession in the future. 

In discussing crowding-out, Laidler acknowledged that the ability of 
Canadians to borrow abroad, essentially at interest rates established in the 
United States capital market, would ensure that the Canadian deficit would 
be unlikely to crowd out private investment. Nevertheless, he noted that 
interest payments will result from such international borrowing and these 
will become a real drain on Canadian production. Furthermore, Canadian 
deficits could undermine confidence in the stability of the dollar and could 
threaten exchange rate depreciation. 

Laidler discussed the important role of public expectations in deter¬ 
mining the effectiveness of fiscal policies. In particular, he felt that 
public trust in the federal economic authorities has been eroded and that 
this erosion has made the size of the deficit a more serious problem than it 


10 


would otherwise have been. Inflation expectations have become deeply 
embedded in the public's mind, and these must be reduced. The only 
cure for inflation is slack in the country's economic activity. The current 
recession is providing this necessary slack, and, consequently, only 
short-term token stimulative policy should be pursued. 'If the Canadian 
economy is to be put upon a stable long-term footing, we must let the 
current recession take its course. The risks of doing otherwise are too 
great' (page 358). 

Throughout this volume, discussants' remarks provide criticisms and 
alternative points of view. It is hoped that the variety of insights and 
perspectives will assist the reader in analysing this complex issue. An¬ 
swers to the question 'Deficits: How Big and How Bad?' appear to depend 
upon the economic circumstances of a particular time and place. Although 
the authors responded in the context of Canada and Ontario in 1983, the 
material that follows has been presented in such a way that it should be 
useful in other places and at other times. The search for appropriate 
analytical methodology provides the general theme throughout the book, 
while the specific references and calculations are useful both in themselves 
and as illustrations of the methodology. 


11 


Overview of the deficit debate 

David Conklin* and Adil Sayeed** 


This background paper provides a broad overview of some of the issues 
pertaining to government deficits. It is our hope that readers who are not 
completely familiar with all of the basic analytical concepts involved in the 
measurement and interpretation of deficits will find this paper a useful 
introduction to the deficit debate. The issues discussed in this paper are, 
of course, covered in greater depth in the conference papers reprinted in 
this volume. 

The first section of this paper presents a brief outline of the budget¬ 
ing procedures conventionally used to measure government deficits. It 
also compares recent Canadian deficits with deficits in other countries and 
with past Canadian deficits. The second section discusses various propos¬ 
als for adjustments to the conventional measurement procedures. The 
third section covers the different theoretical views of how government 
deficits affect the economy. The concluding section explores the linkages 
between views of how deficits should be measured and views of the eco¬ 
nomic impacts of deficits. 

WHAT IS THE DEFICIT? 

People usually refer to a particular figure to describe a deficit. As with 
many things in life, however, a single word or number conveys very little 
meaning. This section of our paper will indicate some of the complexities 
involved in descriptions of deficits. 

A government budget furnishes a detailed accounting of revenues and 
expenditures. A government deficit is simply the excess of expenditures 


Research Director, Ontario Economic Council. 
Research Assistant, Ontario Economic Council. 


12 


over revenues. However, the precise definitions of 'revenue' and 
'expenditure' depend on the particular accounting framework used to draw 
up the budget. Thus, a deficit may appear smaller or larger simply 
because of a shift from one accounting system to another. 

These accounting differences can be illustrated by referring to the 
various measures of the federal deficit that are reported. The public is 
probably most familiar with the deficit calculated on a Public Accounts 
basis. The Public Accounts include revenues and expenditures of the 
various government departments and agencies. 

However, the Public Accounts deficit is not a measure of the total 
borrowing needs of the federal government, which are also affected by 
other, 'non-budgetary' transactions. 'Non-budgetary' expenditures include 
loans and investments made by the federal government as well as any net 
deficits recorded by 'specified purpose funds' such as the Unemployment 
Insurance fund. Net surpluses reported by public employee pension funds 
managed by the federal government are recorded as 'non-budgetary' rev¬ 
enues. The 'financial requirements' of the federal government are a 
measure of the excess of total expenditures, budgetary and non- 
budgetary, over total revenues. An additional measure, the 'net cash 
requirements', differs slightly in that the net effect of government trans¬ 
actions on the foreign exchange market is also included. The differences 
between these measures are set out below: 


Federal deficit on a 
Public Accounts basis 


Federal financial 
requirements 
+ or - 

net requirements due 
to foreign exchange 
transactions 

Federal net cash 
requirements 

The revenues and expenditures of federal ministries and agencies are 
recorded in the Public Accounts. The financial and net cash requirements 
measure the federal government's demand for loanable funds. However, 
economists generally prefer yet another measure, the deficit computed on a 


Budgetary 

expenditures 

+ 

'Non-budgetary' 
expenditures 

Total outlays 


Budgetary 

revenues 

+ 

'Non-budgetary' 
revenues 

Total receipts 


13 





National Accounts basis, for the purpose of gauging the overall impact of 
government on the rest of the economy. 

The National Accounts deficit differs from the Public Accounts deficit 
in its treatment of revenues and expenditures and in that it includes net 
changes in the 'specified purpose funds.' The National Accounts deficit 
measures the increase in the net liabilities of the federal government. 
Purely financial transactions that affect the borrowing requirements, but 
not the net liability position, of the federal government are not included in 
the National Accounts. Because of the widespread acceptance of the 
system of National Accounts among economists, future references to deficits 


in this paper will be 

Federal deficit on 
a Public Accounts = 
basis 


Federal deficit 
on a 

National Accounts 
basis 

+ 

Deficit of the 
provincial-local- = 
hospital sector on a 
National Accounts basis 
+ 

Net change in = 

CPP & QPP balances 

Deficit of the = 

consolidated gov't, 
sector on a National 
Accounts basis 


deficits computed on 


Federal expenditures 
on a PA basis 

+ or - 

Adjustment to 
NA basis 

+ 

Net deficits 
reported by 'specified 
purpose funds' 

Federal expenditures 
on a NA basis 


+ 

Expenditures of the 
P-L-H sector on a 
NA basis 

+ 

Benefits paid 
by CPP & QPP 

Expenditures of the 
CGS on a NA basis 


National Accounts basis. 


Federal revenues on a 
PA basis 

+ or - 

Adjustment to 
NA basis 

+ 

Net surpluses 
reported by 'specified 
purpose funds' 

Federal revenues 
on a NA basis 


+ 

Revenues of the 
P-L-H sector on a 
NA basis 

+ 

Revenues of 
CPP & QPP 

Revenues of the 
CGS on a NA basis 


From some perspectives - for example, with respect to the monetiz¬ 
ation of the deficit - only the federal deficit is important. However, a 
meaningful measure of government activity must cover all levels of govern¬ 
ment. Statistics Canada aggregates the revenues and expenditures of all 
governments in Canada, nets out intergovernmental transactions, and 


14 







reports the budget position of the consolidated government sector. This 
measure also includes the receipts and expenditures of the Canada and 
Quebec Pension Plans and of public hospitals. The progression from the 
federal deficit calculated on a Public Accounts (PA) basis to the deficit of 
the consolidated government sector (CGS) computed on a National Accounts 
(NA) basis is presented above. 1 

Table 1 shows the precise values recorded under the various account¬ 
ing definitions from 1970 to 1982. In recent years, federal financial and 
net cash requirements as well as federal deficits on a National Accounts 
basis have been smaller than federal deficits measured on a Public 
Accounts basis. This difference is due to the exclusion of public employee 
pension fund surpluses from the Public Accounts. Similarly, deficits of 
the consolidated government sector have been smaller than federal deficits 
on a National Accounts basis because surpluses recorded by the Western 
provincial governments and by the Canada and Quebec Pension Plans are 
included in the consolidated government sector accounts. 

No accounting framework can be relied upon to reveal the whole truth 
about government transactions with the rest of the economy. For example, 
receipts and outlays of the Canada and Quebec Pension Plans are recorded 
on a cash, rather than an accrual, basis in the consolidated government 
sector accounts. If the public pension funds were treated on an accrual 
basis, the change over the course of the year in the value of future 
pension obligations to current plan members would have to be included. 
The difficulties involved in attaching a precise value to the change in 
future obligations stand in the way of accrual accounting of the public 
pension funds. Nevertheless, it is clear that current contributions are 
substantially below the levels necessary to fund the future pensions of 
current contributors. If the public pensions could be treated on an ac¬ 
crual basis, the deficit position of the consolidated government sector 
would increase. 2 

It is also worth noting that 'deficits' of government-owned business 
enterprises do not show up in the consolidated government accounts unless 
they are financed by direct subsidies. It is assumed that Crown corpor¬ 
ations borrow only to finance the acquisition of revenue-generating capital 
assets. If this is the case, prospective purchasers should evaluate debt 
issued by public corporations according to the same rules used to evaluate 
private sector debt. In other words, borrowing by a publicly owned cor- 


15 


TABLE 1 

Budget Positions 1970-82 ($ millions) 


4_) /—v 
> + 

O 

00 W 














3 














"O rH 1 

0J 

H U P 

NO 

o 

H 

CM 

m 

O' 

CM 

m 


rH 

CO 

CO 

o 

CO 3 -H 

o 

CO 

CO 

in 

O'. 


CM 

o 

in 

O' 

00 

CO 

o 

33 W U 

CO 

rH 


CM 


o 

CM 

o 

O' 

NO 

O' 

CM 

NO 

•H *H 

rH U 4H 




rH 

CM 


CO 

m 

NO 

<r 

m 

CM 

03 

o o <u 

W +J T3 

3 U 

O <u u 

+ 

+ 

+ 

+ 

+ 

1 

1 

i 

i 

i 

1 

1 

rH 

1 

U w o 




























+ 














s—✓ 














1—1 w 

CO 3 

3 T-l 

O 

•H 5-1 1 

4_) 3 w 

CO W +J 

NO 

m 

NO 

o- 

O' 

m 

rH 

co 

m 

-3- 

CO 

O' 

CO 

2 -H 

NO 


nO 

oo 

o 

o 

O' 

o 

00 

NO 

UO 

r— 

00 

W CJ 

CM 

rH 

m 

CO 

rH 

00 

CO 

CO 

NO 

CM 

rH 

O' 

o 

r—1 +J *H 









•N 



r 


co a 4-i 





rH 

CO 

CO 


o 

O' 

o 


rH 

Feder 
Accou 
or de 

+ 

1 

i 

+ 

+ 

1 

1 

1 

rH 

1 

rH 

i 

CM 

1 

CO 














p 














3 














<U 














X! s 

O' 

no 

m 

<!■ 

O' 

o 

CO 

O' 

O' 

O' 

CO 

03 

rH 

.—f CO <u 

m 

lO 

rH 


NO 

m 

r- 

m 

CM 

<r 

<r 

00 

rH 

co 3 S-i 

r—1 

rH 

CM 


o 

o 

r— 

m 


o 

CO 

m 

CM 

k O 'H 


* 

rv 











0) 3 

CM 

CM 

rH 


CM 

m 


NO 

NO 

rH 

o 

OO 

03 

Fed 

net 

req 

1 

1 

1 

+ 

1 

1 

i 

' 

1 

rH 

1 

rH 

1 

1 

rH 

1 

CO 














p 














s 














r—1 0) 














CO s 

LD 

CO 

03 

CO 


O' 

CO 

00 


O' 

CO 

CO 

CO 

rH *H CJ 

CO 


O 

m 

CM 

CM 

CO 

o 

00 

rH 

m 


03 

3 U 5-1 

UO 


rH 


CC 

CO 

CO 

r^. 

o 

r- 

rH 

CM 

'S’ 

5-1 3 -H 



*>• 


•N 


r, 




r 



<L) 3 3 


rH 

rH 


rH 

m 

<3- 

o- 

CM 

o 

l-H 


o 

3 C O 1 

11 'H U 
fin <h S-i 

1 

1 

i 

i 

1 

1 

1 

i 

rH 

1 

rH 

1 

rH 

1 

i 

CM 

1 















+ 














’s—-' 














CO 














3 














Publie 
surpl 
it(-) 

CO 

CO 

rH 

NO 

-<r 

co 

00 

O' 



CM 

CO 

m 

w u 

NO 

CM 

CO 


CO 

CO 



<r 

o 

t-H 


03 

rH 4-J * H 

t-H 

r^. 




CO 

o 

rH 

CM 

m 

m 

CO 

CM 

3 S 

5-i 3 QJ 






<f 

m 

00 

CO 

rH 

CM 

i—H 

CM 

Fede 
Acco 
or d 

1 

i 

1 

1 

1 

1 

1 

1 

rH 

1 

rH 

1 

rH 

1 

t-H 

1 

CM 

1 

5-1 

o 

rH 

CM 

CO 


in 

NO 

r- 

03 

O' 

o 

rH 

CM 

3 


r— 


r- 

r^. 



r- 



00 

00 

03 

<U 

O' 

O'. 

ON 

ON 

O' 

O' 

O' 

O' 

O' 

O' 

O' 

O' 

O' 

>H 

r—1 

rH 

rH 

rH 

rH 

rH 

rH 

rH 

rH 

t-H 

rH 

l-H 

rH 


SOURCES: Bank of Canada, Monthly Review ; Canada, Department of Finance, (1982). 










poration should not have a qualitatively different impact on the economy 
than borrowing by a private firm. 

If one believes, on the other hand, that Crown corporations are like 
other branches of government and are not operated like private firms, 
then one will conclude that measurement of only direct government borrow¬ 
ing is inadequate. If capital market participants believe that there is no 
distinction between government and Crown corporations, the measure of 
government borrowing should include borrowing by publicly owned enter¬ 
prises. 'Net borrowing requirements' of Canadian governments and gov¬ 
ernment-owned business enterprises are reported by Statistics Canada in 
the Financial Flow Accounts. However, because of the difficulties involved 
in collecting and assessing data from both governments and their enter¬ 
prises, the Financial Flow Accounts generally lag at least one year behind 
the National and Public Accounts. 

Now that some of the basics of budgeting have been covered, it might 
be helpful to place Canada's current government deficit in perspective by 
attempting some comparisons with deficits in other countries and with past 
Canadian deficits. One method of making rough comparisons across coun¬ 
tries and across time is to calculate the ratio of the deficit relative to 
national income. A comparison of deficit/GNP ratios casts some light on 
the question of whether Canada's current government deficit is dispro¬ 
portionately large relative to foreign deficits and relative to our own past 
deficits. 

Accordingly, Table 2 presents data for the deficit/GNP ratios of 
seven major industrialized countries over the period 1978-82. Canada has 
the fourth highest ratio over this period. Many may be surprised to find 
that Canada's deficit/GNP ratio over this period has been less than those 
of Japan and West Germany, two countries with a reputation for pursuing 
'conservative' economic policies. 

It is also interesting to note that Canada's deficit/GNP ratio for 
1978-82 is less than the 3.6 per cent average prevailing during the 1930s 
and the 11.3 per cent average prevailing during the Second World War. 
The ratios for selected periods are presented in Table 3. This information 
concerning our own past deficits and deficits in other countries may be of 
some use in assessing Canada's current government deficit. 


17 


TABLE 2 

Deficit/GNP ratios of seven countries, 1978-82 


Country 

Deficit(-) or surplus 

1978 1979 1980 

(+) as % 

1981 

of GNP 

1982 a 

Average 

1978-82 

Canada 

-3.1 

-2.0 

-2.1 

-1.3 

-6.4 

-3.0 

France 

-1.9 

-0.7 

+0.3 

-1.6 

-2.9 

-1.4 

Italy 

-9.7 

-9.3 

-8.3 

-11.9 

-12.2 

-10.3 

Japan 

-5.5 

-4.8 

-4.2 

-3.9 

-3.3 

-4.3 

United Kingdom 

-4.2 

-3.1 

-3.2 

-2.0 

-2.0 

-2.9 

United States 

0 

+0.6 

-1.3 

-1.0 

-3.7 

-1.1 

West Germany 

-2.5 

-2.7 

-3.1 

-4.0 

-4.1 

-3.3 

NOTE: Data are 

for the 

consolidated government sector in each country, 

not just the central government. Figures for 

1982 are 

estimated. 

SOURCE: Economic 

Outlook. 

OECD (December 1982) 




TABLE 3 

Deficit/GNP ratios 

over selected periods in Canada 


Period 

Deficit as ° 

i of GNP 

1978-82 

-3.0 


1961-77 

-0.2 


1946-60 

+0.8 


1940-45 

-11.3 


1930-39 

-3.6 



SOURCES: Canada, Department of Finance (1982); Statistics Canada, National 

Income and Expenditure Accounts V. I (1975) Cat. 13-533 Occasional 


MEASUREMENT OF THE DEFICIT 


Once a suitable accounting framework has been chosen, computing a deficit 
might appear to be simply a matter of subtracting total revenues from total 
expenditures. However, a number of adjustments to this basic method 
have been proposed on the grounds that the reported deficit, even on a 
National Accounts basis, is not a reliable indicator of the eccnomic impact 
of the government budget. 


The full-employment budget 


Perhaps the best-known adjustment is calculation of the full-employment 
budget (sometimes referred to as the high-employment budget or cyclically- 


18 












adjusted budget). Full-employment budgeting involves estimating what the 
government's revenues and expenditures would have been had the economy 
been operating at full capacity throughout the year. 

Proponents of full-employment budgeting point out that the actual 
budget position is affected by cyclical fluctuations in economic activity. 
As personal and corporate incomes, sales receipts, and payrolls fall below 
anticipated levels during cyclical downturns, so do revenues flowing to the 
government from income, sales, and payroll taxes. On the expenditure 
side, unemployment insurance payments increase as unemployment rises 
during a recession. Consequently, a larger deficit may not be due to a 
shift in government policy or to profligacy on the part of the government. 
Instead, an increase in the deficit may be caused by the response of 
automatic stabilizers to a worsening of the recession. 

However, if calculation of a full-employment budget revealed a 'struc¬ 
tural' deficit even at full employment, then some commentators would ex¬ 
press concern over this remaining 'structural' deficit. Their concern 
would centre on the fact that, with recovery, tax and expenditure pro¬ 
grams would have to be changed in order to balance the budget. Recog¬ 
nizing this concern, other commentators might argue that full-employment 
deficits were warranted by the severity of the recession, and that tax and 
expenditure changes could be readily implemented after recovery arrived. 
The significance of a structural or full-employment budget deficit depends, 
then, on one's theoretical position and one's confidence in the ability of 
government to alter its programs as recovery arrives. 

The full-employment budget is also useful in that changes in it over 
time reflect deliberate shifts in fiscal policy. From this perspective, the 
full-employment budget position is a better measure for the purpose of 
analyzing fiscal policy than is the unadjusted deficit. 

A change in a government's budget position from one year to the next 
can be decomposed into a change due to the cyclical sensitivity of reve¬ 
nues and expenditures and a change resulting from adjustments in tax 
rates and spending plans. Figure 1 illustrates this decomposition. 3 The 
line BB 1 depicts the budget positions of a hypothetical government that 
would result at different levels of national income. A given set of tax 
rates and spending plans is assumed to be fixed initially. BB 1 crosses the 

P 

horizontal axis at Y , the full capacity level of national income, implying 
that the budget would be balanced at full employment. If income is ini¬ 
tially at Y 1 , below the full employment level, the government runs a deficit 


19 


Figure 1 


Budget 

surplus 



Budget 

deficit 


National 

income 


of D 1 , because tax revenues are lower and unemployment insurance pay¬ 
ments are higher than they would have been at full employment. 

Suppose that the government decides to balance the budget over the 
coming year and that income is expected to remain at Y 1 . A balanced 
budget at Y 1 can be achieved only by raising tax rates and/or reducing 
planned expenditures. A new line, BB 11 , depicts the budget positions at 
different income levels under the new tax and spending regime. Assume 
for simplicity that BB 11 is everywhere parallel to BB 1 . BB 11 crosses the 


20 









horizontal axis at Y 1 and shows a surplus of S (exactly equal to D 1 if BB 1 
and BB 11 are parallel) at Y^. 

Now suppose that national income falls unexpectedly to Y 11 . Despite 

the government's intention to balance the budget, the fall in national 

income is so large that the deficit actually increases to D 11 . D 11 is a 

measure of the effect of the unanticipated fall in national income on the 
government's budget position. S measures the effect of the government's 
efforts to reduce the deficit. The net change in the actual deficit, D 11 - 
D 1 , is equal to D 11 , the cyclical shift towards deficit, less S, the policy 
shift towards surplus. 

The change in the budget position calculated on a full-employment 
basis shows only S, the shift towards surplus resulting from higher tax 

rates and/or reduced spending plans. This is not to suggest that actual 

deficits caused by recessions can be ignored. What full-employment bud¬ 
geting can make clear is that increases in government deficits are some¬ 
times the result of unexpected cyclical fluctuations, rather than of deli¬ 
berate government policy. It cannot be concluded simply on the basis of a 
rise in the reported deficit that a government is pursuing a policy of 
deliberate fiscal expansion. Because unforeseen cyclical effects are re¬ 
moved, a change in the full-employment budget position can be interpreted 
as being indicative of a conscious shift in the fiscal stance of government. 
The full-employment budget position is put forward, not as a replacement 
for conventional measures, but as a supplementary indicator of the stance 
of discretionary fiscal policy. 4 

Calculation of this supplementary indicator can be a perilous exercise. 
The simple example illustrated in Figure 1 obscures the difficulties in¬ 
volved in actually computing the full-employment budget position of a 
government. In practice, only the actual budget position and the actual 
level of income in any particular year are observable. The position of a 
line such as BB 1 in Figure 1 can only be approximated by estimating what 
budget positions would have been realized at different levels of income 
given the underlying tax and spending regime. In addition, the precise 

P 

value of Y at any given time can never be established with complete 

P 

confidence. Y is, of course, a purely hypothetical concept. The best 

P 

that can be attempted is a reasonably precise estimate of Y . 

p 

Computing the full-employment budget position involves estimating Y , 
estimating the effect of a change in national income on the government's 
budget position (the slope of a line such as BB 1 ), and using these esti- 


21 


mates and knowledge of the actual budget position and actual income level 
to calculate what the budget position might have been at full employment. 
Methods of computation run the gamut from simple calculations designed to 
yield rough approximations to complex series of equations intended to 
generate more precise estimates. 5 

No method can be relied upon to produce absolutely reliable estimates 

P 

of Y during a period of structural readjustment. It is widely believed 
that the decline of what were once leading industries due to a loss of 
competitiveness relative to foreign producers has contributed to the sever¬ 
ity of the recession in North America. In other words, part of the fall in 
output and employment in some sectors may be a permanent, rather than a 
cyclical, phenomenon. As labour and capital can only be shifted gradually 
out of declining industries into expanding sectors, potential real growth 
must be relatively low during such a period. Past growth trends can no 
longer be relied upon to estimate the current full-capacity level of national 

p 

income. Estimates of Y and of the full-employment budget position must 
be treated with even more caution during a period of structural change. 

These considerations must be kept in mind when one examines Tables 
4 and 5. Table 4 presents estimates of the full-employment budget posi¬ 
tion of the federal government from 1970 to 1981 made separately by the 
federal Department of Finance, the Conference Board of Canada, and the 
Ontario Ministry of Treasury and Economics. Table 5 presents estimates of 
the full-employment budget position of the consolidated government sector 
over the same period made by the Department of Finance and the Confer¬ 
ence Board. 6 Not surprisingly, no two sets of estimates exactly coincide. 
Each group generated its own estimates of the full-employment level of 
national income and of the effect that a change in income would have had 
on revenues and expenditures. 

Inflation adjustment of the budget 


During a period of high inflation, great care must be exercised before 
conclusions are drawn from data reported in unadjusted dollar values. For 
example, the fact that Canada's GNP at the end of 1982 measured in end- 
of-1982 dollars was greater than GNP at the end of 1981 measured in 
end-of-1981 dollars says nothing, by itself, about whether the real level of 
production rose or fell during 1982. Rising prices during 1982 meant that 
a given number of dollars bought a smaller amount of goods at the end of 


22 



w 

d 

o 


•H 

E 

</> 


rH 

•o 

4 -> 







00 

d 

d 







| 

<0 

QJ 







o 

o 

E co 








00 

Jo <U 







o 


O 4-1 

00 

cn 

rH 


u 0 

o 

rH 

<u 

r—H CTJ 

LO 

O' 

CM 

o 


CM 


o 

Cl, E 

<N 

rH 

X 

rH 

O 

co 


d 

E •h 

+ 


— 



- 

4 -> 

<u 

<U +-> 


rH 

CM 

CN 

vo 

O' 

C 

d 

i co 


4 - 

1 

1 

1 

i 

aj 

0 ) 

x: i» 







g 

4-4 

oo 







2 

d 

•rH 







h 

o 

x: 







aj 

CJ 

- 








> 

o 

oo 


<0 

d 

<u 

-a 

aj 


QJ 

X! 

4 -> 

4-i 

O 

c 

o 

•H 

4-1 

•H 

w 

o 

C4 


<D 

00 

T3 

d 

rQ 


a 

<u 

E 

Jo 

o 

rH 

CL, 

E 


01 

x! 


4-1 

o 

w 

flj 

<t 4-> 
<0 

W E 

iX -H 
0 Q 4 -> 
C M 
E—• W 



+ 4 -> 

x 

lO 

X) 

i"- 

o 

LO 

rH 

cn 

< 3 - 

cn 


<r 

r —1 

v — / ’H 

XI 


X) 

00 

o 

o 

O' 

o 

LO 

rH 

o 

o 

CO 

CO U 

CM 

rH 

m 

cn 

rH 

CO 

cn 

cn 

vO 

CN 

VO 

lD 

d 

d -h 

+ 

| 

i 

+ 

*• 

- 


*- 



* 


4-1 

>—4 4-4 





rH 

cn 

cn 

r-* 

o 

O' 

o 


U 

a, <u 





+ 

i 

i 

i 

’-H 

1 

rH 

i 


d T 3 
d 


00 


w 

<u 

4-1 
W CO 

u E 

4-1 *r 4 *r 4 
O E 4-1 
O CO 
>> fl HI 

Cl o 

4-10 — 

CO w 4-1 

* r4 d 

a x) ai 

C E 

03 >i 
O 

Jo i—I 
d (X 
d E 
c 0 W Cl 
CO I 
<u ,-4 

H rH 

E-t d 

4-1 


•rH 

2 

o 

•H 

Cl 


4-1 

c 

o 


T3 

<d 

11 o 
u co 
d d 
CO -i - ) 

d to 


CO CO 

o a; 


4-1 
CL, ,—I 
CD O 

a Jo 
o 


X 

CM 

< 3 " 

O 

O' 

LO 

o 

oo 

cn 

CM 

<r 

<T 

r-. 

00 

rH 

CM 

CM 


•O 


LO 

+ 

' 

+ 

+ 

CM 

rH 

<r 


X) 

o 


I"- 

I 


aj 

•H 

<0 4-» 

rH 

CO 

LO 

00 

cn 

CO 

<r 

Ha- 

o 

CN 

rH 

rH 

i 

Ph 

1 CO 

<r 

+ 

LD 

c^ 

i 

rH 

co 

00 

X 


o 

O' 

rH 


>. E 

LD 


X 

U0 


CN 

CO 

o 

o 

o 

rH 

rH 

rH 

4H 

rH *H 

+ 


1 

1 



— 


— 



r 

d 

O 

rH 4-J 






CO 

CO 

vO 

O' 


r^. 

cn 


o 

r- 


r-'- 

o> 


CN 

cn 

<r 

LT) 

X 


00 

O' 

o 

rH 

r^. 



r^ 

I-'- 

r^- 

r-- 

r^ 

00 

00 

O' 

O' 

O' 

O' 

O' 

O' 

O' 

O' 


c^ 

rH 

rH 

rH 

rH 

rH 

rH 

rH 

rH 

rH 

rH 


SOURCES: Canada, Department of Finance (1982); Carmichael (1979); Ontario Budget 1979 . 






TABLE 5 

Estimates of the full employment budget position of the consolidated government sector, 1970-81 
($ millions) 


i 



+ 

4-> 

vO 

0 

r—H 

v—✓ 


0 

CO 

a 

CO 

u 

00 

T—H 

a 

d 


+ 

+ 


4-> i—I 4H 

u a d 

<5-1X3 

a 

W J-l 

o 


T—4 

Cv) 

lO 

O 

CM 

00 

uO 

ov 


CM 

+ 

Ovl 

r~- 

O 

04 



r. 




H 

CM 


CO 


+ 

+ 

1 

1 


LO 

< 

rH 

CO 

CO 

O 

LO 


CO 

CO 

O 


VsO 

os 

CM 

«»\ 

— 

rv 



UO 

VO 

<r 

m 

CM 


i i i i 


X3 4-1 

a a 
a tu 


0 

ca 

£ M 
>> <U 

O 4-1 

cm 

0 - 

O 

CM 

CO 

0 - 

<u 

r—1 CQ 

r— 

co 

Os] 

< 

< 

00 

u 

O, E 

0 

Os 

r- 

OS 

OS 

00 

a 

E *H 



rv 

1 


•S 

0) 

<U U 

r-H 

CM 

r-H 


CM 


a 

CO 

+ 

4- 

1 


1 

1 


0 ) xi tu 

MH 00 

a -h 
o si 
cj - 


T3 

<U 

<u +-> 
u w 
a a 

a •>—) w 

a -o a 


• H 

a 4-> 

OS 

CM 

CO 

CM 

0 

CO 

m 

m 

UO 

o- 

UO 

CM 

(JU| 

1 a 

os 

CM 

O- 

00 

CM 

vO 

CO 

uo 

uo 

00 

r-- 

vO 


>> E 

OS 

CO 

| 

1 

CO 

CM 

CM 



00 

uo 

rH 

4H 

r-H »H 

— 

+ 



— 

r> 

r. 

•V 



1 


O 

rH +J 

r—4 




1 — t 

00 

CO 

CO 


CM 


<r 


a w 

+ 




+ 

1 

1 

1 

1 

1 


4- 


• u <u 

4J -H 
P 4 1 —I 

tu u 
« >> 
u 


O 

rH 

CM 

CO 


uO 

vO 

r^ 

00 

ON 

0 

H 

r^- 


r^ 

r^ 


r~ 

o- 

r^- 

O' 


00 

00 

G> 

ON 




Os 

OS 


Os 


OS 

0 

r-H 

r-H 

r-H 

r-H 

rH 

r-H 

r-H 

r-H 

r-H 

r-H 

r-H 

T—! 


SOURCES: Canada, Department of Finance (1982); Carmichael (1979). 





1982 than the same number of dollars bought at the end of 1981. GNP 
figures must be expressed in terms of the same base-year dollars before 
production levels can be compared over time. 

It has been suggested that a similar adjustment should be made before 
a government deficit is measured in any particular year. A government 
deficit on a National Accounts basis measures the increase in the money 
value of the government's liabilities. In other words, an annual deficit of 
$20 billion means that the government owes $20 billion more at the end of 

the year than it owed at the start of the year. However, during a time of 

inflation, comparisons over time of total government debt measured in 
current dollars run into the same problems as comparisons over time of 
nominal GNP data. 

To take a simple example, suppose a government owes $100 billion at 

the start of a year, runs a deficit of $20 billion over the course of the 

year, and ends the year with a total debt of $120 billion. As in the case 
of nominal GNP figures, total debt at the end of the year measured in 
end-of-year dollars should not be compared directly with total debt at the 
beginning of the year measured in start-of-year dollars. In this hypo¬ 
thetical case, suppose prices rose by 10 per cent during the course of the 
year. A 10 per cent rate of inflation means that a $100 billion debt mea¬ 
sured in start-of-year dollars is equivalent to a $110 billion debt ($100 
billion x 1.1) measured in end-of-year dollars. Measurement in a common 
unit shows that total government debt has increased by only $10 billion in 
end-of-year dollars (or by $9.1 billion in start-of-year dollars if the 
alternative calculation is performed), rather than by the $20 billion nominal 
deficit. 7 

Several prominent economists have argued that the change in the real 
value of the government debt is a more meaningful measure than the nomi¬ 
nal deficit, which measures the change in the money value of the debt. 8 
At first glance, suspicions might arise that this procedure is merely a 
'sleight-of-hand' accounting technique concocted by government apologists 
to reassure the public that deficits are not as large as they seem. One 
commonly voiced objection is that inflation adjustment is irrelevant from the 
point of view of gauging the effect of a government deficit on the private 
sector. Some argue that whatever the importance of the change in the 
real value of the government debt, it is the nominal deficit that must be 
financed by government borrowing. 

However, the inflation-adjusted budget position can also be advocated 


25 


as the best measure of direct government activity in financial markets. 
This viewpoint rests on the assumption that asset-holders seek to maintain 
the real value of their holdings. They accomplish this by demanding that 
an inflation premium be built into nominal interest rates. An inflation 
premium compensates holders of government bonds (or any other financial 
asset) for the inflation-induced erosion of the purchasing power of the 
bond. Thus, it should not be expected that holders of government bonds 
treat the inflation premium component of nominal interest receipts as reg¬ 
ular income to be divided between consumption and saving. If real wealth 
is to be maintained, inflation premium payments must be entirely allocated 
to additional saving. Under this assumption, the accurate measure of the 
'net' demand by government for loanable funds is the reported deficit 
minus that portion of government interest expenditures channelled auto¬ 
matically into private saving. 

Two methods of inflation-adjusting the deficit have been established. 
One approach seeks to measure the change in the real value of total gov¬ 
ernment debt. This involves subtracting from the reported deficit the 
amount by which the actual inflation rate caused a real depreciation of the 
debt. The other method entails estimating the portion of interest payments 
that were actually inflation premiums paid to holders of government debt. 
These inflation premiums are based on expected inflation, which may not 
equal actual inflation. 

If the actual inflation rate were always perfectly anticipated, the two 
inflation adjustment procedures would exactly coincide. In other words, 
the deficit minus inflation premium payments would always equal the change 
in the real value of the debt measured in end-of-year dollars. However, 
when the actual rate of inflation diverges from the anticipated rate of 
inflation, the two inflation adjustment methods arrive at different results. 

Both approaches have been applied to Canadian data. The Depart¬ 
ment of Finance has produced a set of estimates of the inflation-adjusted 
budget positions of the federal and consolidated government sectors based 
on measuring the change in the real value of government debt. In a 
study commissioned by the Economic Council of Canada, Professor Gregory 
Jump of the University of Toronto published estimates of the inflation- 
adjusted budget position of the federal government based on his estimates 
of expected inflation. 9 A comparison of these two sets of estimates ap¬ 
pears in Table 6, which shows that the two methods produced slightly 
different results in every year. Both methods demonstrate that an increa- 


26 


TABLE 6 

Estimates of the inflation-adjusted budget positions of the federal and consolidated government sectors, 
1970-80 ($ millions) 


o 

o 

CD 

Su >4-i c /3 

O O 



/-N 4H 

G 

O 

r-H 

eg 

m 

G 

CM 

m 

<r 

I— H 

eo 

4 

1 > 

o 

03 

OC 

m 

G 

<r 

eg 

o 

m 


OC 


’ o 

OC 

r-H 

4 

eg 

r- 

o 

eg 

o 

G 

G 

G 

w 

4-1 oc 

+ 

4- 



•V 






** 

r-H 3 

•H 




t-H 

CM 


eo 

m 

G 


UO 

03 hH 

o • 




4 

4 

i 

1 

1 

1 

1 

1 

3 O- 

•H W 












*J S-i 

4h 3 












U 3 

CD O 












<fi c/3 

33 U 












33 

CA 












CD 

3 












CD *-> 

O 












U C/3 

o 












3 3 













33 •’—) 













3 33 

O !-4 

OO 

uO 

vO 

O 

r- 


o 

vO 


o 


•H CQ 

4h O 

<r 

<r 

OC 

eg 

eo 


i-H 

CM 

eo 

G 

eg 

1 

4—> 

T-H 

t-H 

r*- 

t-H 

g 

cn 

1- 

CO 

t-H 

r-H 


3 

C/3 O 

rv 





i 

4 





4-i O 

CD CD 

eg 

t-H 

r-H 

m 




CM 

eo 

<r 

r-H 

O -H 

4-) 00 

4 

4 

4 

4 

4 



1 

1 

4 

4 

4-> 

CO 












• 03 

5 












4-> i—1 

•H 4-> 












3- 4-( 

4-) > 












CD 3 

C/3 O 












G -H 

<D oc 


























/ — S W 













4 4-1 













— s *H 

G 

uo 

G 


O' 

in 

r-H 

eo 

<r 

eo 



o 

G 


vO 

OC 

o 

o 

a> 

G 

m 

r-H 

G 

r—i 

W -H 

eg 

r—i 

m 

eo 

r-H 

00 

eo 

eo 

G 

eg 

G 

i-H CO 

3 4h 

4 

| 

| 

4 

ri 




rv 


rs 

03 3-1 

rH CD 





t-H 

eo 

eo 

r*^ 

G 

G 

O 

3 CD 

G 33 





4 

1 

1 

i 

T-H 

1 

t-h 

4-> 33 

S-I 









1 


1 

U (D 

3 4 












c 4-c 

co o 













CO 












33 /-n 

CD 












CD 1 

4-> 

o 

o 

o 

O 

C 

o 

G 

G 




4-1 w 

G 03 

o 

o 

o 

o 

O 

o 

O 

G 




W 

E E 

OC 


OO 

vO 

eo 

CM 

eg 

uo 




3 4-> 

Z ' *H 

4 

4 

4 

r* 


rv 


•V 




•r—) *H 





r-H 

mi 

eg 

r-H 

m 




33 CJ 

CO 




4 

4 

1 

1 

1 




CO *H 

CD 












1 <4H 













3 CD 













O 33 

CD 












• rH 

CJ 












4-> S-I 

G 












03 O 

03 












r-H 

3 












4-4 /—\ 

• rH 












3 4 

Ph 

uO 

OO 

OC 



H 


vO 


00 

G 

*H ' s — 

CO 


VO 

vO 

r-H 

co 

eg 




G 

00 


<H-H dJ 

o 




t-h 

vO 

o 

G 

eo 

O 

uO 

rH W 

O 4-1 


4 

4 









03 3 

03 

t-H 



CM 


r-H 

r-H 

m 

cc 

eo 

<3- 

^H r—I 

• E 

4 



4 

4 

| 

1 

| 

1 

1 

1 

CD G 

4-> -H 












33 Su 

G 4-1 












CD 3 

CD C/3 












G w 

Q CD 













S-I 

o 

T-H 

CM 

eo 


m 

vO 

r^. 

O0 

G 

G 


03 




r- 

r^* 

r^» 



r- 

r*- 

00 


CD 

G 

G 

CO 

G 

G 

G 

G 

G 

G 

G 

G 



T-H 

f—< 

r-H 

t-H 

t-h 

t-h 

t-h 

r-H 

r-H 

I-H 

t-h 


SOURCES: Canada, Department of Finance (1981); Jump (1980a). 














sing proportion of recent deficits is attributable to inflation. 10 
The capital budget 


Division of the government budget into separate current and capital ac¬ 
counts is yet another suggested modification. Proponents argue that 
formation of a capital budget would bring the government accounts in line 
with established business accounting practices. In the private sector, a 
distinction is made between current operating expenses and expenditures 
on assets expected to yield future revenues. Debt financing of capital 
investment is considered sound business practice, so long as borrowing 
costs are less than the expected returns on capital. Since governments 
also purchase assets that generate revenue and/or public benefits in the 
future, it is argued that governments should also distinguish between 
current and capital expenditures. With the aid of a capital budget, a 
large deficit due to capital investment might be viewed with less concern 
than a smaller deficit caused by an excess of current spending over 
revenues. 11 

In fact, some have gone so far as to propose that current expendi¬ 
tures should always be covered by tax revenues, while capital spending 
should always be entirely debt financed. The rationale for such a policy 
is to ensure that the beneficiaries pay for the services flowing from public 
investment in a fair manner. Rather than obliging taxpayers to pay the 
full price for these services in the form of higher taxes in the years 
assets are purchased, debt financing allows them to pay gradually as the 
benefits from public capital accrue over time. 

A number of practical difficulties must be resolved before a capital 
budget can be drawn up. Foremost among these difficulties is the problem 
of deciding what constitutes public investment. In business accounting, 
any asset that is expected to generate revenues beyond the year of pur¬ 
chase is treated as capital. Unfortunately, this rule cannot easily be 
applied to the government sector. An additional problem is the calculation 
of depreciation on public sector capital. 

According to a narrow definition, only expenditures on assets that 
generate revenue in the form of direct user charges (such as toll bridges 
and roads) should be considered government capital spending. However, a 
case can be made that spending on what is known as social overhead 
capital (such as transportation and communication networks and even health 


28 



and education facilities) improves long-run economic efficiency, thereby 
producing higher tax revenue in the future. Use of this broader defini¬ 
tion can expand the size of the capital budget considerably. 

In Canada, some of the provincial governments currently provide a 
systematic accounting of capital spending. In background papers accom¬ 
panying the 1978, 1980, and 1982 budgets, the Ontario Ministry of Trea¬ 
sury and Economics has provided breakdowns of annual capital expend¬ 
itures extending as far back as the 1970-1 fiscal year. 12 The Ministry 
favours a broad definition of public capital comprising all assets 'that have 
a lifespan of more than one year and provide public benefits beyond the 
initial year.' 13 It should also be noted that the Ministry's figures include 
expenditures on existing assets acquired from the private sector. Whether 
or not a clear distinction should be made between spending to purchase 
existing financial assets and spending to create new capital is yet another 
unresolved issue in the mechanics of capital budgeting. 

The practical difficulties involved in capital budgeting stem from the 
basic fact that a government is not a business. The purpose of govern¬ 
ment is not to maximize its profits, but rather to provide a framework 
within which individuals can maximize their own welfare. Therefore, use 
of business accounting methods based on profit maximization criteria may 
not always be appropriate in the public sector. 

For instance, use of the capital budget to implement a rigid rule such 
as full debt financing of capital spending in all years would greatly reduce 
the flexibility thought to be necessary for the formulation of effective 
fiscal policy. Fiscal policy for the purpose of stabilizing short-run fluctu¬ 
ations in economic activity could only be accomplished by varying the level 
of capital spending from year to year. As the purpose of public capital 
formation is to improve long-run economic efficiency, a basic conflict 
between the short-run and long-run goals of government would arise. 

For this reason, most advocates of capital budgeting have shied away 
from urging the adoption of fixed rules linking government capital spend¬ 
ing and debt financing. Instead, the capital budget has been put forward 
as the best means of providing regular information concerning the amount 
and allocation of public investment. The case for the capital budget rests 
on its utility as a descriptive tool that sheds light on how public funds are 
spent. Before the capital budget can take its place as a useful source of 
public information, the difficult question of defining public investment must 
be resolved. 


29 


MACROECONOMIC EFFECTS OF DEFICITS 


The simple version of the National Income Model taught in most first-year 
Economics courses suggests that governments are responding correctly to 
the current recession by allowing their deficits to increase. This model 
explains changes in aggregate income and employment as being due to 
fluctuations in total demand. 14 The danger is said to be that an initial 
fall in demand, with its negative consequences for income and employment, 
can lead to increased caution on the part of consumers and investors, 
which generates further reductions in demand. The behaviour of the 
Canadian economy during the 1929-33 period, when production fell by 30 
per cent and the rate of unemployment rose from 3 per cent to 24 per 
cent, 15 is held up as a striking example of the consequences of this 
process. 

According to Lord Keynes, whose theories form the basis of the 
National Income Model, governments should have shored up demand during 
the Great Depression by increasing their deficits. Government deficits 
represent direct injections of additional spending into the economy. To 
the extent that income generated by a government deficit is then spent by 
recipients, a deficit can have a 'multiplier' effect on aggregate demand far 
greater than the size of the deficit itself. While government deficits did 
rise during the 1930s, this was not due to a deliberate policy of stimulat¬ 
ing demand. In fact, governments strove to contain deficits at what were 
viewed as manageable levels. This policy of restricting deficits has been 
blamed for helping to prolong the depression. 

Some observers have drawn parallels between the current situation 
and the early years of the Great Depression. Between the second quarter 
of 1981 and the final quarter of 1982, real income in Canada fell by 6.5 
per cent. The unemployment rate jumped from 6.9 per cent in August of 
1981 to 12.8 per cent by the end of 1982. Fears have been expressed that 
the current climate of uncertainty is not conducive to a revival of private 
sector spending. If this is the case, then it may be argued that a further 
expansion of government deficits is necessary to prevent the recession 
from turning into a depression. According to this view, governments that 
heed calls for reduced deficits at the present time will be repeating the 
mistakes of the past. 

More sophisticated modern versions of Keynes' original model also 
conclude that deficit spending can be an effective means of stabilizing 


30 


aggregate demand. It should be kept in mind that Keynesianism does not 
imply a policy of permanent deficit spending. Rather, Keynesians assign a 
useful role to government deficits only when the level of private spending 
falls temporarily. By minimizing the contraction in total demand, a deficit 
helps to restore the level of confidence necessary for an eventual upturn 
in private expenditure. As income and employment return to their normal 
levels, the need for a government deficit disappears. 

However, Keynesian economics has never commanded universal accep¬ 
tance and has come under increasing attack in recent years. Opponents of 
Keynes reject his view that the economy can become stuck at an 'under¬ 
employment equilibrium.' Non-Keynesians tend to have greater faith in the 
ability of the economy to bounce back from a cyclical downturn. 

This conflict is essentially over the flexibility of market prices. 
Keynesians believe that market rigidities may permit involuntary unemploy¬ 
ment of productive resources to persist over long periods of time. There¬ 
fore, government action may be the only means by which the economy can 
be returned to full employment. Non-Keynesians argue that temporary 
fluctuations in the levels of production and employment are corrected by 
the automatic adjustment of market prices. According to this view of the 
world, fiscal stabilization is unnecessary and may even be counter-produc¬ 
tive. 

This broader debate over the nature of the market economy cannot be 
treated adequately in a brief survey paper. 16 Instead, this section 
focuses on critiques of deficit spending that have been made within the 
context of the simple, fixed-price model of the economy. The possibility 
that government deficits may 'crowd out' private expenditures and have no 
effect on aggregate demand is discussed. Reference is made to the pos¬ 
sible effects of government deficits on the rate of interest, the exchange 
rate, and the consumption-savings decision. Additional topics covered are 
the link between deficits and inflation and the significance of forecasting 
for current policy choices. 

Crowding-out 

Opposition to government deficits is often based on fears that a higher 
deficit will drive up interest rates, thereby depressing interest-sensitive 
private expenditures. One of the components of private spending that is 
widely believed to be most negatively related to the rate of interest is 


31 



investment. Investment in productive capital is a necessary ingredient for 
future economic growth. If a higher government deficit forces interest 
rates up, the resulting decline in investment will reduce growth potential 
over the longer run as well as offset any positive impact of the deficit on 
demand in the short run. It is the possible consequences of government 
deficits for future economic growth that many of their opponents find most 
disturbing. 

This crowding-out argument can be demonstrated in the simple IS-LM 
model of the closed economy. The closed-economy model, which ignores 
imports, exports, and capital flows, is especially applicable to a country 
such as the United States, whose foreign trade sector is small relative to 
the domestic sector. Consequently, American economists have tended to 
debate the merits of deficit spending within the closed economy framework. 
Since this American debate has influenced Canadian attitudes toward 
deficits, it is worthwhile setting out the case against fiscal activism in a 
closed economy. The open economy case, which is more applicable to the 
Canadian situation, will be discussed in the next subsection. 

The simplest textbook versions of the Keynesian model assume a 
constant rate of interest. A fixed interest rate implies that the LM curve 
depicting points of money market equilibrium is horizontal. Given a hori¬ 
zontal LM curve, an increase in the government deficit, which shifts the 
goods market IS curve outwards, moves the economy to a new equilibrium 
at a higher level of national income (see Figure 2). 

An early criticism of deficit spending was that the LM curve, far from 
being horizontal, is, in reality, vertical (or nearly so), because the de¬ 
mand for money is independent (or nearly so) of the rate of interest. A 
vertical LM curve means that an increased deficit merely pushes the eco¬ 
nomy to a new IS-LM intersection at a higher rate of interest, while hav¬ 
ing no effect on national income. 17 In other words, private spending in 
the form of interest-sensitive investment expenditure falls to offset exactly 
the increase in the deficit. This form of crowding-out, via the impact of a 
deficit on the rate of interest, is the version that should be most familiar 
to regular readers of business articles in today's newspapers. 

While the slope of the LM curve was a controversial topic during the 
1960s, substantial evidence indicating that money demand is interest-sensi¬ 
tive has effectively put an end to this debate. In any event, the 
crowding-out argument does not rest solely on the possibility of a vertical 
LM curve. 


32 


Figure 2 

Crowding-out in a simple IS-LM framework 


2a. Zero crowding-out and fully effective fiscal policy 
(horizontal LM curve at constant?) 


interest 

rate 



2b. Full crowding-out and completely ineffective fiscal 
policy (vertical LM curve) 


interest 

rate 



ISi 

(after rise 
in deficit) 


33 














2c. Partial crowding-out and partially effective fiscal policy 
(Positively-sloped LM curve) 


interest 

rate 



LM 


IS, 

(after rise 
in deficit 


An increase in the deficit financed by the sale of government bonds 
to the public increases the stock of privately held wealth under the as¬ 
sumption that government debt is not treated as a liability by taxpayers 
(see page 39). Empirical evidence indicates that the demand for money is 
positively related to net wealth. Eventually, a deficit-induced rise in 
wealth causes the LM curve to shift upwards. One contention is that the 
end result of shifts in both the IS and LM curves is a new long-run 
equilibrium at a higher interest rate and the same level of national 
income. 18 Once again, the effect of an increased deficit on aggregate 
demand is neutralized by a fall in interest-sensitive private expenditures 
(see Figure 3). 

Whether or not long-run wealth effects ultimately neutralize the initial 
stimulative impact of a higher deficit is an empirical question. During the 
late 1960s, economists at the Federal Reserve Bank of St. Louis sought to 
test the impact of fiscal changes on U.S. national income. Econometric 
results derived from the St. Louis equations indicated that fiscal changes 
had no effect on income beyond the very short run. However, the statis¬ 
tical methodology underlying the St. Louis equations has been called into 
question. From the econometric point of view, the jury is still out on the 


34 








Figure 3 

Long-Run crowding-out in the IS-LM framework 
3a. Full crowding-out and ineffective fiscal policy 


interest 

rate 



3b. Partial crowding-out and partially effective fiscal policy 

interest 

rate 



35 














crowding-out issue in the United States. 19 


International capital flows: implications for crowding out 

The previous subsection considered the crowding-out argument exclusively 
within the framework of a closed economy. Aggregate demand for Cana¬ 
dian goods is, of course, heavily influenced by foreign demand for our 
exports and by the division of domestic demand between Canadian products 
and foreign imports. Even if it could be proved conclusively that fiscal 
policy is effective in an economy closed to foreign trade, Canadian econo¬ 
mists would still be faced with the task of gauging the impact of fiscal 
policy in an open economy. 

If the result derived in the widely taught Mundell-Fleming model is to 
be believed, fiscal policy is impotent in a small, open economy under a 
flexible exchange rate system. 26 The key to the model lies in the defini¬ 
tions of the terms 'small' and 'open.' An open economy is simply one in 
which domestic residents are free to purchase foreign goods and assets, 
while foreigners are able to buy domestic goods and assets. Smallness 
refers to the fact that the economy is too insignificant a demander or 
supplier of internationally traded goods and assets to be able to affect 
their prices. 

In other words, exports from a small, open economy such as Canada's 
must compete with products from other countries and cannot be priced 
arbitrarily without regard to world market conditions. Similarly, Canadian 
consumers must accept import prices determined on the world market. In 
addition, Canadian interest rates cannot deviate from interest rates pre¬ 
vailing in the rest of the world. If Canadian bonds are priced too high, 
both domestic and foreign holders will exchange their Canadian bonds for 
cheaper foreign assets. This will drive prices of Canadian bonds down to 
prevailing world levels. Conversely, if Canadian bond prices fall below 
world levels, increased demand for cheap Canadian bonds will push prices 
back to world levels. As long as no restrictions are placed on capital 
flows across international boundaries, interest rates in Canada are deter¬ 
mined, not by domestic supply of and demand for bonds, but by world 
supply and demand. 21 It goes without saying that world market conditions 
are beyond the control of the domestic authorities. 

The ramifications of smallness and openness can be illustrated in the 
IS-LM framework of Figure 4. Assume the economy is initially in equili- 


36 



Figure 4 

Crowding-out in a small, open economy 

interest 

rate 



brium, with the IS and LM curves intersecting at the world rate of inter- 
* 

est, r . An increase in the government deficit tends to shift the IS curve 
outwards to IS', thereby putting upward pressure on the domestic interest 
rate. However, any upward movement in the domestic rate induces asset- 
holders to switch out of foreign bonds into domestic bonds. The resulting 
capital inflow places upward pressure on the exchange value of the 
domestic currency in terms of foreign currencies. As the domestic 
currency appreciates, foreign demand for exports falls and domestic de¬ 
mand shifts towards imports at the expense of home goods. This shifts 
the IS curve back towards the initial equilibrium position. 

The end result of the rise in the deficit is currency appreciation and 
deterioration of the trade balance, with no change in the level of national 
income. The speed at which all this takes place depends on how quickly 
bondholders react to deviations in interest rates. If capital is highly 
mobile, currency appreciation follows almost immediately after a rise in the 


37 







deficit, with not even a temporary rise in income. Crowding-out occurs 
with the interest rate unchanged. In this model, a deficit affects the 
exchange rate, rather than the interest rate, and crowds out spending 
that is affected by changes in the exchange rate. 

The obvious question is the extent to which the Mundell-Fleming 
model is applicable to the Canadian economy. Does Canada fit the small, 
open economy definition? Is capital so mobile between countries that no 
scope exists for changes in Canadian interest rates independent of move¬ 
ments in world rates? 

In addition, it can be argued that the exchange value of the Canadian 
dollar is not fully flexible at the present time. Many observers accuse the 
Canadian government of attempting to restrict exchange rate fluctuations 
within a narrow range of its own choosing. It could be argued that fiscal 
policy can play an effective role under a managed floating system. 

Finally, open economy macroeconomics has changed considerably since 
the Mundell-Fleming model was introduced in the early 1960s. The import¬ 
ance of expectations in the determination of exchange rates has been re¬ 
peatedly emphasized in recent work. The simple version of the Mundell- 
Fleming model outlined above contains the implicit assumption that asset- 
holders do not expect the fully flexible exchange rate to change over time. 
A more realistic view of exchange rate expectations could alter the con¬ 
clusion that fiscal policy is ineffective under flexible rates. 22 

These considerations should not obscure the main lesson of the 
Mundell-Fleming analysis. No model can be expected to deliver absolute 
truths about such a complex organism as the economy. At most, models 
can be used to make broad generalizations about what seems to be true. 
What can be learned from the Mundell-Fleming model is that, in a flexible 
exchange rate system, fiscal policy is likely to have much less effect on 
income and employment the smaller the size and the greater the openness 
of the economy. In Canada, this is a lesson that cannot be ignored. 

Bond financing vs. tax financing 

Robert Barro, a leading American economist, has put forward yet another 
argument against fiscal activism. Barro's case against deficit spending 
rests on the assumed ability of taxpayers to forecast their expected life¬ 
time tax liabilities. It has long been recognized that the stimulative impact 
of deficit spending depends upon the extent to which taxpayers realize 


38 



that government debt issued to finance today's deficit must be serviced 
with higher taxes tomorrow. If taxpayers expect that their future tax 
liabilities will be increased as a result of bond financing of a current 
deficit, they will tend to increase current saving in order to prepare for 
higher taxes in the future. Thus, the stimulative impact on aggregate 
demand of a higher deficit will be offset by a shift towards saving and 
away from consumption expenditure. 

At one time, the general consensus was that, while some public aware¬ 
ness of the link between future taxes and current deficits must be allowed 
for, full discounting of the future tax liability attendant upon bond finan¬ 
cing is highly improbable. It was held to be unrealistic to assume that all 
taxpayers act on the basis of a fully worked-out lifetime plan. Further¬ 
more, even if full rationality could be assumed, the ability of governments 
to postpone repayment by rolling debt over would mean that taxpayers 
should not expect to be liable for debt not likely to be retired until long 
after their deaths. 

In his 1974 article entitled 'Are government bonds net wealth?', 
Barro argued that economic behaviour is affected by consideration of 
events beyond the expected lifetime of the individual, so long as concern 
exists for the welfare of descendants. If this is the case, rational indivi¬ 
duals may be expected to set aside savings to pass on as bequests in 
order that their beneficiaries might meet tax liabilities resulting from past 
deficit financing. 23 

If the Barro model is a close approximation of reality, the effects of 
bond financing of government spending do not differ from those of tax 
financing. By levying taxes, governments force a reduction in private 
expenditures. With full discounting of future taxes, a substitution of 
bond financing for taxation reduces private spending by an equivalent 
amount as income is set aside to cover future tax obligations. Conse¬ 
quently, a change in the government budget position has no effect on 
aggregate demand. This proposition was dubbed the 'Ricardian equiva¬ 
lence theorem' because of its similarity to comments on the same subject 
made by David Ricardo, the great classical economist of the 19th 
century. 24 

Barro's achievement lies in his derivation of theoretical conditions 
under which full discounting is possible. Critics have not found flaws in 
the analytical structure of Barro's model. Rather, they have cast doubt 
on the practical relevance of the Ricardian equivalence theorem. The 


39 


uncertainties facing individuals in the real world, however interested they 
may be in the well-being of their descendants, are said to be so great as 
to ensure that planning for future taxes is of only minor importance. 25 

Since criticism of Barro's work has focused on its practical plausi¬ 
bility, empirical testing would seem to offer the only basis for resolving 
this issue. Much of the statistical analysis of the Ricardian equivalence 
theorem has involved estimation of the aggregate consumption function. 
Unfortunately, a clear-cut answer has not yet emerged from the econo¬ 
metric literature. Results of some studies indicate that consumer spending 
does fall on approximately a one-to-one basis as government debt issue 
rises. This is the expected result if Barro's proposition holds in the real 
world. In other work, however, no consistent relationship between aggre¬ 
gate consumption and government borrowing is found. 26 

Before moving on, it might be instructive to examine within the IS-LM 
framework the 'Ricardian' argument that deficit financing is equivalent to 
tax financing. Supporters of the Ricardian equivalence theorem believe 
that a rise in the deficit will not affect national income. However, their 
argument is that any increase in the deficit is immediately offset by an 
equivalent rise in private saving, so that the position of the IS curve 
never changes. Full discounting of future taxes implies that government 
bonds are not treated as net wealth by the private sector, so that a 
change in the deficit does not affect the position of the LM curve either. 
The level of national income, the rate of interest, and the exchange rate 
are unaffected by a change in the deficit. Crowding-out occurs, not 
through the effect of a higher deficit on the interest rate or exchange 
rate, but via the direct impact of deficit financing on the consumption- 
savings decisions of rational individuals. 

Deficits, monetary policy, and inflation 


To this point, the possible effects of fiscal policy on income and employ¬ 
ment alone have been discussed. As is evident from recent experience, 
macroeconomic policy can be set with another target in mind - the rate of 
inflation. Some critics of fiscal activism have taken the position that 
deficit spending, in addition to being an ineffective means of stabilizing 
aggregate demand, can do real harm by adding to inflationary pressures. 

The most direct link between fiscal policy and the rate of inflation 
runs through the effect deficit spending can have on the rate of growth of 


40 



the money supply. Up to now, analysis has been restricted to the effects 
of a government deficit financed by bond sales to the private sector. 
However, central government bonds can also be purchased by the central 
bank. 27 The end result of a series of transactions arising from govern¬ 
ment spending financed by bond sales to the central bank is an increase in 
the reserve holdings of the private banks. Increased bank reserves can 
be used to back an expanded volume of demand deposits, which are a 
principal component of the money stock. A greater supply of money, all 
other things being equal, puts upward pressure on the general level of 
prices. 

It is important to recognize that there is no necessary connection 
between government deficits and monetary policy. Central banks in most 
countries enjoy some degree of independence. The Bank of Canada, for 
instance, is under no legal obligation to purchase bonds from the federal 
government. Moreover, when the Bank does buy bonds, it remains free to 
take steps that neutralize the effect of its purchases on the money supply. 

Nevertheless, many observers attribute the upsurge in inflation rates 
in Canada and elsewhere during the 1970s to accommodation by the central 
banks of large government deficits. It is argued that the nominal freedom 
granted to central banks is, in actual practice, a facade. Government 
officials, motivated by political concerns, are said to exert considerable 
pressure on central banks to finance budget deficits. 28 

Since the alleged link between deficits and inflation is not a theore¬ 
tical necessity, but rather an alleged probability, empirical evidence of 
such a relationship would seem to be required. In Canada, attention has 
been focused on the early 1970s, when double-digit inflation made its first 
appearance. This period was also marked by a drift towards larger fed¬ 
eral deficits, large purchases of government bonds by the Bank of 
Canada, and high annual rates of money growth. By late 1975, concern 
about the inflationary effects of federal deficits had became so great that a 
group of prominent Canadian economists sent a letter to the prime minister 
urging a reduction in the federal deficit. However, research done at the 
Conference Board of Canada by Robert Crozier indicated that the Bank of 
Canada had retained control over monetary policy and that there had been 
no causal connection between deficits and the rise in inflation during this 
period. 29 The subsequent ability of the Bank to implement its policy of 
gradually reducing the rate of monetary growth, in spite of continued 
large deficits, tends to bear out the conclusions of the Crozier Report. 


41 


Another theoretical basis for linking deficits and inflation has emer¬ 
ged. Thomas Sargent and Neil Wallace, two prominent American econo¬ 
mists, have recently developed a model in which continued deficits even¬ 
tually force an inflationary expansion of the money supply. Their result 
hinges on the assumption that there exists 'an upper limit on the stock of 
bonds relative to the size of the economy .' 30 If large deficits continue for 
so long that this limit is reached, the government has no recourse but to 
finance continued deficits by money creation. Furthermore, if expectations 
of future inflation play an important role in current price formation, 
inflation may remain high, despite an anti-inflationary monetary policy, 
because current high deficits lead rational agents to believe that an infla¬ 
tionary policy will be adopted in the future. 

Forecasts: implications for policy prescriptions 

This section is more speculative than would normally be appropriate in a 
survey paper. Nevertheless, we believe that, at the present time, differ¬ 
ent interpretations of forecasts have important implications for policy 
prescriptions. 

This paper has described several theoretical structures, each of 
which is built upon a particular view of the way the economy works. 
Much of the deficit debate is a dispute as to which theoretical structure 
best represents today's economy. Apart from this theoretical dispute, 
some of the policy controversy rests upon a disagreement concerning 
forecasts of economic trends. 

Indications are growing that the economy may be pulling out of the 
recession. Given these indications, one may favour deficit reductions; and 
one may predict that recovery will automatically reduce the deficits as 
revenues increase and recession-related expenditures fall. Recent move¬ 
ments in the Consumer Price Index suggest that inflation has fallen drama¬ 
tically. Given these movements, one may forecast lower interest rates and 
lower deficits as a result of smaller interest payments on outstanding 
government debt. People who believe these optimistic forecasts will not 
advocate larger budget deficits. In fact, such observers may argue for 
tax increases if they believe that structural deficits currently exist. 

On the other hand, one may point out that current indications of 
recovery are not widespread. For example, business investment in plant 
and equipment has not yet shown signs of recovery. Throughout the 


42 



depression of the 1930s, noticeable increases and decreases occurred in 
various economic indicators. It may be that current positive developments 
will be overcome by relapses, as happened throughout the 1930s. It can 
be argued that forecasters use models in which reaction patterns and 
relationships are those of the 1970s, and that we cannot expect these 
traditional patterns and relationships to hold in the current recession. 

In particular, it may be that businessmen have been so buffeted by 
their 1979-83 experiences that they will not increase their investments in 
either inventories or fixed capital as confidently as they did in the 1970s. 
Small current signs of recovery may fail to stimulate investment. Con¬ 
sumers have been shocked by the recent inflation, interest rates, and 
unemployment to the extent that their confidence and purchasing patterns 
may not be easily restored. Hence, it can be argued that governments 
should now expand their deficits as the best path to recovery. 

One's prediction of the future course of the economy will no doubt 
affect one's judgment with respect to the 'crowding-out' controversy. If 
one expects that business investment and consumer spending will be de¬ 
pressed for the next couple of years, then one may believe that larger 
deficits can be financed domestically with little 'crowding-out' of any type. 
On the other hand, if one expects that recovery is imminent, then one may 
fear that larger deficits could have negative 'crowding-out' effects at a 
critical turning point. 

In fact, one's forecasts may affect one's attitude towards the various 
theoretical structures presented above. It is quite possible that a partic¬ 
ular theoretical structure will describe the economy well at one stage and 
yet be less appropriate when basic underlying trends have changed. It is 
even possible that public knowledge and understanding of economic issues 
are increasing to such a degree that business and consumer reaction 
patterns are changing. In particular, reaction times may be shorter than 
in earlier days. Hence, economic policy may have to be based on expec¬ 
tations of a higher level of public rationality and a greater awareness of 
probable public response to that economic policy. 

Several other extremely important elements of the debate depend on 
one's forecasts. To argue for a change in the deficit is not particularly 
helpful unless one attaches some approximate dollar figure to one's recom¬ 
mendation, unless one presents a time profile for this action, and unless 
one suggests some guidelines for the composition of fiscal changes. With 
respect to each of these elements, one's view will be seriously affected by 


43 


one's forecasts as well as one's theoretical position. These elements prob¬ 
ably deserve more attention than they have so far received in the current 
deficit debate. 

This uncertainty surrounding forecasts of future economic trends 
forms the basis of an additional argument against deliberately increasing 
the deficit to stimulate the economy. It may be argued that the very fact 
that forecasts can vary so widely demonstrates that existing knowledge of 
how the economy operates is far from complete. Fiscal 'fine-tuning' can 
only succeed if policymakers can be sure of both the state of the economy 
when policy changes are made and the impact of these changes over time. 
If the future course of the economy is uncertain, it may be that by the 

time a policy intended to deal with a particular situation takes effect, 

conditions will have changed so much that the policy will have a perverse 
impact. This point of view is summed up by Martin Feldstein, Chairman of 
the Council of Economic Advisers in the United States, who wrote: 

We therefore do not have and may never be able to have enough precise 

information to be confident that discretionary fiscal changes can reduce 
the average amplitude of the business cycle...The extent of our ignorance 
and the potentially powerful effects of fiscal changes imply that the 
magnitude of discretionary fiscal changes should be very limited. 31 

CONCLUSION 


It should be clear by now that the two broad aspects of the deficit issue 
cannot be discussed separately. Attitudes toward the proposed adjust¬ 
ments to conventional deficit measures are based upon particular views of 
the macroeconomic effects of deficits. 

For example, those who believe that fiscal policy cannot affect income 
and employment are not much interested in the full-employment budget 
position. Obviously, there is no need to gauge the stance of a policy that 
is believed to be entirely ineffective. For those who believe that deficits 
do real damage by adding to inflation and/or by reducing long-run growth 
potential, it is the actual deficit to be financed, rather than the hypo¬ 
thetical full-employment deficit, that is the variable of interest. 

At the opposite pole, those with an abiding faith in the possibility of 
fiscal stabilization attach a great deal of weight to the full-employment 
budget. Moreover, fiscal activists are not overly concerned by the compu¬ 
tational difficulties involved in full-employment budgeting. In the Keyne¬ 
sian framework, the primary purpose of the full-employment budget is to 


44 


serve as an indicator of deliberate shifts in the fiscal stance of govern¬ 
ment. It is the change in the full-employment budget position from one 
year to the next, rather than the precise measurement in a single year, 
that is of greatest significance. 

As long as a reasonably reliable method of calculation is applied 
consistently to all years, it should be possible to follow the course of 
discretionary fiscal policy over time. In Tables 4 and 5, for example, 
different groups report slightly different numbers for the full-employment 
budget positions in each year. Yet disagreement concerning the direction 
of change from year to year is rare. The course of federal fiscal policy 
during the 1970s can be discerned by looking at any one of the three sets 
of estimates reproduced in Table 4. 

Between these two camps are those who believe that deficit spending 
can soften the impact of a recession, but that it should be avoided during 
a period of recovery. This view of fiscal policy allows a place for the 
full-employment budget. The key question is whether there exists a 
'structural' deficit that may cause crowding-out when income and employ¬ 
ment begin rising. In this framework, an accurate estimate of the full- 
employment budget position at a particular point in time becomes much 
more important. 

Divisions over inflation adjustment of the budget are not so sharp. 
Inflation adjustment is compatible both with a belief in full crowding-out 
and with faith in effective fiscal policy. The assumption that assetholders 
seek to maintain real wealth by channelling inflation premium payments 
directly into additional saving is not exclusively identified with any partic¬ 
ular school of thought. If this assumption were accepted, Keynesians 
would expect only the inflation-adjusted deficit to affect aggregate demand. 
Those who put forward a version of the crowding-out hypothesis would 
expect the inflation-adjusted deficit, not the entire unadjusted deficit, to 
have crowding-out effects. 

Fiscal activists might combine the full-employment and inflation 
adjustments. It is possible for a change in the full-employment budget 
position to be entirely due to a change in inflation premium payments. If 
these payments flow directly into private saving, such a change would not 
affect aggregate demand. In addition, since the size of the inflation 
premium is determined by market expectations, expenditures on inflation 
premiums cannot be quickly altered at the discretion of the government. 
Hence, a change in the full-employment budget position adjusted for in- 


45 


flation would seem to be a more accurate indicator of a shift in discretion¬ 
ary fiscal policy than a change in the unadjusted full-employment budget 
position. 

Of equal importance are considerations stemming from the open nature 
of the Canadian economy. In fact, the choice among open-economy macro- 
economic viewpoints determines which deficit is considered important. 
Acceptance of the Mundell-Fleming model as a reasonable framework for 
analysis of the Canadian economy implies that Canadian deficits, however 
measured, are unimportant in terms of their effects on income and employ¬ 
ment. According to this view, the government deficits that affect Canada 
for good or ill are those of countries large enough to influence world 
market conditions. Since the world’s largest economy, the United States, 
is also our largest trading partner, it is the American deficit that should 
be of most concern to Canadians. 

Attitudes toward the American deficit would then depend upon closed- 
economy macroeconomic analysis. Those who believe that deficits can 
stimulate demand in a large economy might look toward a higher U.S. 
deficit to raise American demand for Canadian goods. On the other hand, 
if expectations are that deficits of large countries push up world interest 
rates, then high American deficits should be viewed as a threat to Can¬ 
adian investment spending. Alternatively, those who adopt Barro's 
'Ricardian' view believe that deficits alter the composition of aggregate 
demand between private and public expenditure, but have no effect on 
interest rates. Thus, the U.S. deficit would be viewed as a threat only 
insofar as American governments probably have a lower demand for Cana¬ 
dian goods than does the American private sector. As Canada has little or 
no influence on American fiscal policy, the Mundell-Fleming framework 
implies that Canadians cannot avoid the consequences of American deficits, 
whatever they may be. 

Recommendations concerning current Canadian deficit policy are thus 
based on views of how the economy works and of where deficits are at the 
present time. For example, Keynesians interpret projections showing that 
the inflation-adjusted, full-employment budget position of the consolidated 
government sector will be in surplus in 1983 as evidence that Canadian 
governments are not doing enough to end the recession. 32 Keynesians 
argue that the quickest route to recovery and lower deficits in the future 
is via deliberate expansion of deficits now, when demand is weak and 
future prospects are uncertain. 


46 


This argument can be illustrated by a graph similar to the one intro¬ 
duced in Figure 1. Current projections are for national income in 1983 to 
be considerably below potential GNP at full capacity. As a result, the 
consolidated government sector will likely run up a considerable deficit 
even on an inflation-adjusted basis. In Figure 5, suppose that D 1 g 3 and 
Y X g 3 on budget line BB X gg are the actual deficit and national income levels 
realized in 1983, if current tax rates and government spending plans 
remain in place. S 1 g 3 is the inflation-adjusted, full employment surplus 
implied by this fiscal regime. 

Keynesians propose that the federal and provincial governments 
co-ordinate a general reduction in tax rates and/or an increase in spend¬ 
ing plans. In other words, Keynesians would like to see Canadian govern¬ 
ments shift to a new budget line, BB xl gg. They contend that such a 


li 


program of fiscal stimulation would increase national income to Y : gg, 
above the level that would be achieved if no action were taken. While the 
deficit might increase slightly to D xl gg, such an increase would not be as 
large as might be feared, because the fiscally-induced rise in national 
income would generate extra tax revenues and lower unemployment insur¬ 
ance payments. Furthermore, the increase in the deficit would lead to 
lower deficits in the future. The increase in income from Y 1 ^ to Y 11 


would restore consumer and business confidence 

rF 


83 A 83 
National income would 


continue rising towards Y 1 . Deficits would fall continuously as Canadian 
governments moved up along BB 11 . 

Keynesians fear that if no action is taken, the economy will become 
trapped at a level of output far below Y . The current recession is much 
deeper than any cyclical downturn experienced in the recent past. Some 
Keynesians argue that optimistic forecasts of an imminent upturn in private 
sector spending are based on the pattern of past business cycles. In the 
context of Figure 5, Keynesians are concerned that if the economy is 
allowed to remain at Y X gg for long, private sector confidence may deteri¬ 
orate further. The result would be a further fall in national income and a 
further increase in the deficit. 

Those who believe that fiscal policy in Canada cannot affect income 
and employment vigorously oppose the Keynesian prescription. According 
to these critics, economic recovery cannot be artificially induced by gov¬ 
ernments. In other words, a shift from BB^g to BB 1:l gg would leave 
national income unchanged at Y X gg. The result would be a much larger 
deficit of D*gg and consequently a greater government debt total and 


47 


Figure 5 

Inflation-adjusted budget position of the consolidated government sector 


Surplus 



increased debt servicing costs in the future. 

Non-Keynesians have much more faith in self-correcting mechanisms 
built into the economy. Recent reports that economic recovery is already 
underway are said to be fully consistent with historical experience of 
business cycles. According to those who oppose deliberate expansion of 
deficits, the proper government strategy is to allow the private sector to 
recover on its own. National income will gradually rise and government 
deficits will slowly fall. Opponents of fiscal activism believe that deliber¬ 
ately increasing deficits now could even slow the economy's return to 
by delaying recovery in the private sector. 

Of course, the deficit debate cannot be resolved in this paper. Our 


48 









purpose here has been to set out in simple terms the arguments made by 
both sides. We hope it has been made clear that both sides can call upon 
economic theory to support their positions, and that the appropriateness of 
various arguments may change over time as shifts occur in the state of the 
economy and the nature of public perceptions. 

NOTES 

1 See Bird (1979, 45-55), Canada (1983, 23-5), and Statistics Canada 

(1975, 163-93) for thorough explanations of the differences between 

the various accounting frameworks. 

2 See Bossons and Dungan (1983, 14-15) for a discussion of this issue. 
Of course, it can be argued that changes in the amount of deferred 
taxes that will be paid in the future when RRSPs are cashed in also 
affect the net liability position of the government. It may be that 
trying to find an all-encompassing measure acceptable to everyone is 
impossible. 

3 This figure has been borrowed from Ott and Ott (1978, 90). 

4 See Okun and Teeters (1970, 77-116), Solomon (1962, 105-11), and 
Tarshis (1979) for arguments in support of full-employment budgeting. 

5 See Canada, Economic Review (1978-82, various issues) for explanations 
of the methodology used by the Department of Finance in calculating 
what are called the 'cyclically adjusted' budget positions of the 
federal and consolidated government sectors. For a full description 
of a more complex methodology, see de Leeuw, et al. (1980). The 
methodology discussed in that article is used by the Council of Econo¬ 
mic Advisers to make official estimates of the full-employment budget 
position of the United States federal government. 

6 The 'cyclically adjusted' budget concept employed by the Department of 
Finance is not exactly equivalent to the conventional full-employment 
budget concept. The Department decided that the difficulties involved 
in estimating the full-capacity level of national income were too 
great. Instead, the Department chose to calculate the 'cyclically 
adjusted' level of national income, which is what GNP in any given 
year would have been had the economy followed a trend growth path. 
However, this technical difference is not of great import. The moti¬ 
vation for computing the 'cyclically adjusted' budget position coin¬ 
cides with the rationale underlying full-employment budgeting. 

7 This type of explanation of the meaning of inflation adjustment can be 
found in Bossons and Dungan (1983, 16). 

8 See Tobin (1980, 56), Barber (1979, 44-7), and Barro (1981, Ch. 6) for 
arguments in favour of correcting conventional budget measures to 
remove the effect of inflation. 

9. The brief explanation in this paper of why the expected rate of infla¬ 
tion, rather than the actual rate, might be used to perform the infla¬ 
tion adjustment is a simplification of the reasoning developed by Jump 


49 



and others. For a full exposition of the Jump position see Jump 

(1980b, 990-1004). 

10. It should be made clear that the inflation-adjusted budget position 
presented over a number of years is not a measure of government sur¬ 
pluses and deficits in constant or base-year dollars. The inflation- 
adjusted budget position in any given year is calculated in the dol¬ 
lars of that year. It can be broadly interpreted as what the budget 
position would have been had there been no inflation during the year 
in question. Thus, inflation adjustment can transform unadjusted 
deficits into inflation-adjusted surpluses. (See the 1971 and 1972 
figures for the federal government and the 1976, 1979, and 1980 
figures for the consolidated government sector reported in Table 6.) 
In such a case, the unadjusted deficit is less than the amount by 
which inflation depreciated the real value of the debt. Or, from the 
alternative point of view, the reported deficit is less than govern¬ 
ment interest expenditures stemming from inflation premiums. 

11 See Comiez (1966) and articles by Eckstein, Musgrave, and Shoup in 
Review of Economics and Statistics (May 1963). 

12 See Ontario (1978, 1980, 1982). 

13 Ontario (1982, 4). 

14 See Lipsey, et al., (1982, ch. 28-34) or any other standard introduc¬ 
tory text for a simple presentation of the National Income Model. 

15 Figures are taken from Urquhart, ed.(1965, 61, 132). 

16 For a modern 'Keynesian' view of price adjustment, see Lipsey, (1981, 

545-76); For a modern 'non-Keynesian' view of price adiustment, see 
Lucas (1980). 

17 The IS curve shows the combinations of interest rates and income 

levels at which withdrawals from aggregate demand (savings, taxes, and 
imports) are equal to injections (investment, government spending, and 
exports). The IS curve is negatively sloped because investment is 

negatively related and savings are positively related to the rate of 
interest, while savings rise due to an increase in income by more than 
does investment. The LM curve shows the combinations of interest 
rates and incomes at which money demand is equated to a given money 
supply. It is positively sloped because money demand falls as the 

rate of interest rises, but increases as income rises. The inter¬ 

section of the IS and LM curves is the only interest-income combin¬ 
ation at which both the money market and the goods market are in 

equilibrium. For a full exposition of the IS-LM framework, see Parkin 
(1982, ch. 18-20). Throughout our analysis, the extreme assumption 
that the price level is fixed will be maintained. No distinction 

between nominal and real income will be made. This is obviously a 

highly unrealistic assumption; however, as will be seen, the various 
arguments against fiscal activism can all be made within the framework 
of the fixed-price IS-LM model. 

18 Of course, an increase in net wealth also shifts the IS curve out¬ 
wards. See Infante and Stein (1976) for a full presentation of the 
argument that the end result of all shifts in both the IS and LM 


50 



curves due to an increased deficit is likely to be no change in nat¬ 
ional income. See Blinder and Solow (1973) for an opposing argument. 

19 The original presentation of the St. Louis equations appears in 
Andersen and Jordan (1968, 11-24). Questions concerning the method¬ 
ological soundness of the St. Louis tests are discussed by Modigliani 
and Ando (1976, 30-42). 

20 The original presentation of this model is to be found in Mundell 

(1963, 475-85). A clear, concise exposition of the model appears in 

Dornbusch (1980, 193-9). 

21 Casual observation would seem to contradict the basic premise of the 
Mundell-Fleming model, namely that international capital mobility 
forces the equalization of interest rates across countries. For 
example, the November 1982 issue of OECD Main Economic Indicators 
reports that comparable short-term interest rates during September 
1982 were 12.73 per cent in Canada, 7.85 per cent in the United 
States, 3.25 per cent in Switzerland, and 17.72 per cent in Italy. 

However, these interest rate deviations can be reconciled within 
the Mundell-Fleming framework as being due to expectations concerning 
future exchange rate movements. What must be remembered is that rates 
of return on various assets are evaluated by a prospective purchaser 
in terms of some common currency, normally the domestic currency of 
the country of residence. For a Canadian, the rate of return on a 
foreign-denominated asset is equal to the rate of return in terms of 
the foreign currency plus the rate of change of the value of the 
foreign currency in terms of Canadian dollars expected over the hol¬ 
ding period. Thus, Canadians holding U.S. dollar-denominated bonds 
paying 8 per cent, even when Canadian bonds are available at 13 per 
cent, are not acting irrationally so long as their expectation is that 
the value of the U.S. dollar in terms of Canadian dollars will rise by 
5 per cent. Perfect capital mobility does not imply that r_, = 

r„ g at all times, but that r^^ ~ r u S_ + ex P ecte< ^ rate of change 

of the value of the U.S. dollar in‘terms or Canadian dollars. 

22 See Frenkel and Mussa (1981) for a brief survey of modern theories of 
open macroeconomics. 

23 See Barro (1974, 1095-1117; and 1981, ch. 6). 

24 See Ricardo (1951). 

25 See Tobin (1980, 49-72) for a critique of Barro's model. 

26 The econometric evidence in favour of the Ricardian equivalence 
theorem is summarized in Tanner (1979, 317-21). Evidence against full 
discounting is presented in Tobin and Buiter (1979). 

27 Since the Bank of Canada does not deal in provincial government secur¬ 
ities, there is no direct link between provincial deficits, provincial 
borrowing, monetary policy, and inflation. The analysis in this 
section refers exclusively to federal deficits. 

28 The most forceful exposition of this view can be found in Buchanan and 
Wagner (1977). These two American economists believe that democratic 
societies are plagued by a built-in tendency towards money-financed 


51 



deficits. The work of Buchanan and Wagner has been cited prominently 
by supporters of the balanced-budget amendment that recently came 
before the U.S. Congress. 

29 See Crozier (1976) for a summary of the Crozier Report. Also see 
Canadian Public Policy (1977) for articles presented at a symposium on 
deficits and inflation. 

30 Sargent and Wallace (1981, 4). 

31 Feldstein (1982, 17-8). 

32 For example, see Bossons and Dungan (1983). 

REFERENCES 

Andersen, L. and J. Jordan (1968) 'Monetary and fiscal actions: a test of 
their relative importance in economic stabilization.' Federal Reserve 
Bank of St. Louis Review 51, November 11-24 

Barber, C. (1979) 'Inflation distortion and the balanced budget.' Chal ¬ 
lenge 22, September-October 44-7 

Barro, R. (1974) 'Are government bonds net wealth?' Journal of Political 
Economy 82, 1095-1117 

- (1981) Macroeconomic Analysis (New York: Academic Press) 

Bird, R. (1979) Financing Canadian Government: A Quantitative Overview 
(Toronto: Canadian Tax Foundation) 

Blinder, A. and R. Solow (1973) 'Does fiscal policy matter?' Journal of 
Public Economics 2, 319-37 

Bossons, J. and P. Dungan (1983) 'The government deficit: too high or 
too low?' Canadian Tax Journal 31, 1-29 

Buchanan, J. and R. Wagner (1977) Democracy in Deficit (New York: 
Academic Press) 

Canada, Department of Finance (1978-82, annual issues) Economic Review 
(Ottawa) 

- (1983) The Federal Deficit in Perspective (Ottawa) 

Canadian Public Policy (1977) 'Inflation and the budget balance.' 3, 268-98 

Carmichael, E. (1979) 'The budgetary impact of high employment.' Canadian 
Business Review 6, Summer 27-32 

Comiez, M. (1966) A Capital Budget Statement for the U.S. Government 
(Washington: Brookings Institution) 

Crozier, R. (1976) 'Deficit financing and inflation: facts and fictions.' 
Canadian Business Review 3, Spring 10-13 


52 





















de Leeuw, F. et al. (1980) 'The high-employment budget: new estimates 
1955-80.' Survey of Current Business 60, November 13-43 
Dornbusch, R. (1980) Open Economy Macroeconomics (New York: Basic 
Books) 

Feldstein, M. (1982) 'Government deficits and aggregate demand.' Journal 
of Monetary Economics 9, 1-21 

Fleming, J. (1962) 'Domestic financial policies under fixed and floating 

exchange rates.' International Monetary Fund Staff Papers 9, 369-79 
Frenkel, J. and M. Mussa (1981) 'Monetary and fiscal policies in an open 
economy.' American Economic Review 71, 253-8 
Infante, E. and J. Stein (1976) 'Does fiscal policy matter?' Journal of 
Monetary Economics 2, 473-500 

Jump, G. (1980a) 'Inflation-related spurious elements in measured savings 
of various sectors of the economy: the Canadian experience, 

1962-77.' Economic Council of Canada Discussion Paper 151 
- (1980b) 'Interest rates, inflation expectations, and spurious elements in 
measured real income and saving.', American Economic Review 70, 
990-1004 

Lipsey, R. (1981) 'The understanding and control of inflation: Is there a 
crisis in macroeconomics?' Canadian Journal of Economics 14, 545-76 
Lipsey, R. et al. (1982) Economics, 4th edition (New York: Harper & 
Row) 

Lucas, R. (1980) 'Methods and problems in business cycle theory' Journal 
of Money, Credit and Banking 12, 696-715 
Modigliani, F. and A. Ando (1976) 'Impacts of fiscal actions on aggregate 
income and the monetarist controversy: theory and evidence.' In J. 
Stein, ed.. Monetarism (Amsterdam: North-Holland) 30-42 
Mundell, R. (1963) 'Capital mobility and stabilization policy under fixed 
and flexible exchange rates.' Canadian Journal of Economics and 
Political Science 29, 475-85 

Mussa, M. (1976) 'The exchange rate, the balance of payments and mone¬ 
tary and fiscal policy under a regime of controlled floating.' 
Scandinavian Journal of Economics 78, 229-48 
Okun, A. and N. Teeters (1970) 'The full employment surplus revisited.' 

In Brookings Papers on Economic Activity no. 1, 77-116 
Ontario, Ministry of Treasury, Economics and Intergovernmental Affairs 
(1977) 'Reassessing the scope for fiscal policy in Canada.' Ontario 
Tax Study No. 15 


53 




















- (1978) 'Ontario's borrowing and public capital creation.' Ontario 

Budget 1978 , Budget Paper A 

Ontario, Ministry of Treasury and Economics (1980) 'A solid fiscal founda¬ 
tion for the 1980s.' Ontario Budget 1980 , Budget Paper C 
Ontario, Ministry of Treasury and Economics (1982) 'Public investment and 
responsible financial management.' Ontario Budget 1982 , Budget 
Paper C 

Ott, D. and A. Ott (1978) Federal Budget Policy , 3rd edition (Washington: 
Brookings Institution) 

Parkin, M. (1982) Modern Macroeconomics (Scarborough: Prentice-Hall of 
Canada) 

Review of Economics and Statistics (1963) 'Budgetary concepts: a 
symposium.' 

Ricardo, D. (1951) 'Funding system.' In P. Sraffa, ed.. The Works and 
Correspondence of David Ricardo , vol. 4 (Cambridge: Cambridge 
University Press) 

Sargent, T. and N. Wallace (1981) 'Some unpleasant monetarist arithmetic.' 

Federal Reserve Bank of Minneapolis Quarterly Review 5, Fall 1-17 
Solomon, R. (1962) 'The full employment budget surplus as an analytical 
concept.' American Statistical Association Proceedings of the Business 
and Economics Section 1962 , 105-11 

Statistics Canada (1975) The National Income and Expenditure Accounts 
vol. 3, Catalogue 13-549E, Occasional (Ottawa) 

Tanner, J. (1979) 'Fiscal policy and consumer behaviour.' Review of 
Economics and Statistics 61, 317-21 

Tarshis, L. (1979) 'Full-employment budgeting.' Ontario Economic Council 
Discussion Paper 

Tobin, J. (1980) 'Stabilization policy: ten years after.' Brookings Papers 
on Economic Activity No. 1, 19-71 

- (1980) Asset Accumulation and Economic Activity (Chicago: University 

of Chicago Press) 

Tobin, J. and W. Buiter (1979) 'Debt neutrality: a brief review of doc¬ 
trine and evidence.' In G. von Furstenberg, ed., Social Security 
vs. Private Saving (New York: Ballinger Publishing) 

Urquhart, M., (ed.) (1965) Historical Statistics of Canada (Toronto: 
MacMillan) 


54 























Government deficits, inflation, and 
future generations 

Franco Modigliani* 


I have three main points to make in this paper. The first is that current 
deficits are not a major cause of inflation; on the contrary, I shall argue 
that inflation is a major cause of deficits. The causation goes from stag¬ 
flation to deficit, not the other way around. 

Next, I want to argue that deficits are, nonetheless, a bad thing. 
They may not be bad in the short run, under some circumstances, but if 
they continue in the long run, they have serious conseguences. 

Finally, I propose to establish that deficits are not the cause of our 
current trouble, despite the fact that they are bad, for the simple reason 
that there are scarcely any deficits. That is, in terms of that deficit 
which is harmful, there are very few deficits of any significance at this 
time. 

Let me begin by making it clear why we are so concerned with defi¬ 
cits. Figure 1 shows the behaviour of deficits in the United States. The 
top line shows the U.S. national debt for 1946 alone, and then the rapid 
and finally precipitous rise in that debt from 1969 on. It shows how, as 
Mr. Reagan was taking office, the national debt was just crossing the $1 
trillion line. Mr. Reagan told us that if the debt were stacked in dollar 
bills it would reach the moon, or some such gruesome thing. Of course, 
since then, his administration has proceeded to bigger and bigger deficits, 
and one can see from the projections for the future, shown by the broken 
lines, that the stack is really going beyond the moon, towards Mars. The 
top line of Figure 2 (I will come back to the rest of the figure later) 
shows roughly the same thing, but in terms of the flow of deficits rather 
than the stock of debt. One can see how the current deficit took off 


Professor, Department of Economics, Sloan School of Industrial 
Management, Massachusetts Institute of Technology. 


55 


Figure 1 

U.S. federal government debt and net worth 

/ 

/ 

$ billions / 



somewhere in the late seventies and, like the national debt, is also going 
through the ceiling. These huge deficits have been connected, in many 
minds, with our current deep economic problems - with the problem of 
stagflation. Figure 3 may help to explain why. The top panel again 
shows the deficits, but this time by five year periods. It shows that we 
had small deficits right after the Second World War, but that beginning in 
the late sixties deficits got bigger and bigger. As this happened, the 
next panel shows, we went from little inflation to more and more inflation. 
At the same time, as the last panel shows, productivity growth, which had 
been high, became small, zero, and finally negative. 

Figure 4 draws a similar picture from the experience of nine major 
OECD countries. The height of the first column in the upper panel shows 
the annual average deficit (if above the horizontal line) or surplus (if 
below) for each country during the 1960s. The second column relates to 
the period 1975-80. It is apparent from the columns that in the sixties 
there were surpluses almost everywhere. Then came the seventies, and 
almost everywhere deficits replaced surpluses. The columns in the bottom 
panel represent inflation in the same periods. It is apparent that, as 


56 





Figure 2 

U.S. Government deficits with adjustments 


$ billions 



deficits grew everywhere, so did inflation. Furthermore, the growth of 
inflation was greatest in Italy, where the deficit went way up, and in the 
U.K., where the budget went from a large surplus to a substantial deficit. 
Again, it seems as though deficit goes with inflation. The connection 
looks impressive, and it may have inspired the story that Reagan told the 
American people when he took over. The problem we have, according to 
Reagan, is that the government spends too much. As it spends more and 
more, it finds that it cannot increase taxes any further, because people 
cannot take it. Consequently, it starts running a deficit - it resorts to 
the printing press. The printing press creates inflation, and inflation 
creates disruption, depression, and low productivity. That is the mech¬ 
anism. In my view, this argument is almost entirely fallacious. 

The first thing that I want to examine is the relation between deficit 
and inflation. There is absolutely no reason to think there is association 
between the two, at least in the sense that deficit implies recourse to the 
printing press which, in turn, means inflation. We have to remember that 
in all reasonably developed countries the government finances its deficits 
by selling bonds, not by printing money it cannot print. Money is printed 


57 




Figure 3 

Federal deficits, inflation, and productivity growth in the United States 
(5-year intervals) 





by the central bank, which decides how much government debt and other 
assets it wants to buy with it. There is no mechanical connection between 
running a deficit and creating money supply, and we have countless illus¬ 
trations of the lack of such a connection. Just think, for example, of 
what happened in the United States in 1982. The deficit was enormous; it 
was well over $100 billion. Yet the central bank expanded its monetary 
base by only about $10 billion. Furthermore, this increase was less than 
it had been in years when the government deficit was smaller. Of course, 
in an underdeveloped country, one that has no financial markets, the 
government may sometimes have no way to sell bonds except to the central 
bank. Even when this does occur, one suspects that there is a will to 


58 





















































Figure 4 


Deficit 

Income 

(%) 

10 1 
8 - 

6 - 

4 - 

2 - 


-2 - 
-4 - 


Inflation 

(%) 



finance the deficit by inflation. But we are not concerned with primitive 
countries. In general, in developed countries there is no simple connec¬ 
tion between deficits and money creation. 

If the deficit is not a cause of inflation, is the deficit otherwise so 
bad? The issue of whether the deficit is bad, the issue sometimes refer¬ 
red to as crowding-out, is one of the classic issues that economists have 
enjoyed debating for several centuries. Many reputations have been made 
and destroyed in the course of that debate. There are essentially four 
views relating to the question of whether a government deficit is a burden 
on anybody, particularly whether it is a burden on future generations. 
These can be labelled (i) the 'naive' view that there is a burden, (ii) the 
'naive' view that there is no burden, (iii) the 'sophisticated' view that 
there is a burden, and (iv) the 'super-sophisticated' view that there is no 
burden after all. Let me go over these four views quickly. 

The naive view that there is a burden is the view of the man in the 


59 




















































































































CO 

ON 

CO 


NO 

rH 



UO 

CO 

00 

'O 

CO 


rH 

NT 

00 

00 





0)0 

o N 

CO 

O' 

m 

H 


CN 

ON 

u0 

rH 

ON 

ON 

U0 

CO 

o 

NO 

CO 


rH 

ON 




X3 +-) 

1 

O' 

o 

m 


CO 

OO 


Nf 

NO 

o 

CN 

00 

o 

H 

CO 

CO 

oo 

CO 





4-1 -H 

u0 


CO 

<r 

NO 


CN 

ON 

o 


oo 

CO 

rH 


00 


CN 

o 

CN 





O 


CN) 

o 

CO 

H 

rH 

O 

o 

o 

o 

NO 


o 

rH 

CN 

CN 

o 

CO 

O 

NO 

rH 



O -H 

on 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

O 

o 

o 

o 

O 

o 

o 

o 

o 

o 



4-S N-i 

rH 























0) 


o 

o 

o 

o 

c 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 



G "O 
o 

•H X! 
4-> 0) 

U +J 


ON 

00 

LO 

'O 

1 

00 

NO 




UO 

ON 

r— 


OO 

rH 

1 

CO 

NO 


CN 



0) O 

ON 

CO 

O' 

rH 

CO 

LO 

OO 

CN 

ON 

uO 

ON 

NO 

rH 

NO 

NO 

ON 

CN 

00 


ON 

ON 



O O 

1 

o 

<r 

m 

CN 

rH 

00 

CN 

o 


NO 

CO 

UO 

uo 

ON 

CN 

CO 

CO 

00 

CO 

CN 



o 0*0 

rH 

CO 

rH 

H 

ON 


O 

CO 

rH 

NO 


CO 

OO 

o 

CO 


CN 


CN 

CN 



O 0) 

NO 

rH 

o 

CN 

H 

o 

o 

H 

c 

o 

CO 

rH 

o 

H 

rH 

o 

rH 

rH 

o 


rH 



CJ S-i 

ON 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 




1—1 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 





u0 

'O 

rH 

UO 

ON 

ON 


oo 


CN 

UO 

uo 


CO 

ON 

CO 


r- 


ON 




ON 

ON 

uo 

CN 

00 

CN 

CN 



rH 

CN 

ON 

CN 

OO 

UO 

o 

o 

CN 

r^- 


NO 




| 

00 



CO 

00 

NO 


NO 

o 

CO 

00 

o 

rH 

CO 

00 

CO 

NO 

O' 

CO 

CO 




LD 


r- 

o 

'O 

CN 

m 


H 

NO 

OO 

CO 

H 

ON 

uo 

00 


rH 

ON 

CN 

o 




r- 

rH 

uo 


oo 

ON 

rH 

O' 

<3- 

CO 

<t 

NO 


NO 

00 

rH 

00 

o 

CN 

U0 

r*- 




ON 

H 

o 

o 

o 

o 

rH 

o 

o 

H 

H 

rH 

o 

o 

o 

CN 

rH 

rH 

O 

rH 

o 



rO 

rH 























a 

o 


o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

O 

o 

o 

o 

o 

o 



■H 

4-1 
























TO 
























rH 


m 

NO 

NO 

r— 

NO 

CN 

m 


CO 


'O 

UO 

r— 


CO 


O' 

CN 

NO 

rH 



U-i 

ON 

'O 


m 

o 

CO 

H 

uo 

'O 



NO 

O' 

r— 

CO 

CO 


00 

UO 

<3- 

CN 



C 

1 

o 

ON 

m 

<r 

CO 

CO 

CO 


r— 

n- 


OO 

CO 

CO 

O' 

rH 

uo 

00 

rH 

NO 



1—1 

o 

<3- 

CO 

H 

CN 


NO 

rH 


CO 

CO 


uO 

o 

r— 

uo 

00 

O' 


O' 

CN 




NO 

CN 

CO 

CO 

CN 

m 



CN 

CN 

'O 

CO 

UO 

<3- 

CO 

CN 

u0 

CO 

CO 

CO 

CN 




ON 

O 

o 

o 

O 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 





o 

o 

o 

O 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 



00 
























C 
























•H 


CO 

m 

CO 

CO 


CN 

m 

uO 

rH 



00 

uO 

NO 


NO 

m 

CO 

CN 

<r 



> 

O'- 

LO 

Nj 

ON 

<T 


NO 

UO 


CO 

NO 


r^. 

<r 

o 

NO 


CN 

CN 

o 

ON 



TO 

1 

CO 

CO 

NO 

ON 


o 

o 

o 

r^. 


m 

o 

u0 

O' 

u0 

O' 

00 

UO 

O' 

r— 



w 

U0 

CO 

o 

ON 



<r 

<3- 

uo 

rH 


uO 

CN 

CO 

CO 

CO 

rH 

NO 

NO 

CO 

CO 





H 

rH 

CN 

CO 

rH 

m 

O 

O 

CN 


'O 

CN 

H 

NO 

CN 

CN 

o 

rH 

CN 

CN 



+J 

ON 

o 

o 

o 

o 

o 

o 

o 

O 

O 

o 

o 

o 

o 

o 

o 

O 

o 

o 

o 

o 



G 

0)TO 

e o 

G -H 

S-i 4-> 

rH 

o 

o 

1 

o 

1 

o 

o 

1 

o 

o 

O 

o 

1 

o 

1 

o 

o 

1 

o 

o 

o 

1 

o 

o 

o 

o 

1 

o 

1 



0) TO 
> S-i 


LO 

ON 

rH 

NO 

CO 

rH 

NO 


ON 

rH 

UO 

NO 



00 

r*-. 

rH 

rH 

ON 

'J- 



o 

ON 

NO 

o 

ON 

NO 

rH 

r- 

CN 

CN 

CO 

o 

<3- 

r^- 

ON 

CO 

CTs 


00 

CN 

UO 

CO 



oo 

1 

<3- 

NO 

CO 

r^. 

ON 


uO 

u0 

ON 

H 

O 

rH 

NO 



NO 

UO 

O' 


ON 




o 

o 


H 

CO 

CO 

OO 

H 

rH 

NO 

ON 

'O 

o 

O 


00 

NO 

rH 

o 

uO 

c 



i—1 

NO 

LO 

CO 

o 

o 




CO 

H 

o 

CN 

ON 

UO 


o 

CO 

m 

CN 

CO 

rH 



TO 

ON 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

c 

o 



S-i 

rH 























4-» 

G 


o 

o 

o 

1 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

O 

o 



0) 

c_> 















































re 
























3 
























0c 

•H 
























Ph 





















E 
























o 

w 


g 





















XJ 

0) 


•H 















C/D 





-a 

oc 

4-1 

















X! 





c 

G 

CQ 


X) 



cQ 












c 





TO 

•rH 

4-1 


0) 



•H 












TO 


rH 



rH 

US 

CO 

rH 

00 



rH 

TO 

p 


-* 

03 


in 


TOO 



rH 


TO 



s-l 




G 


S-i 

CQ 

•H 


TO 

S-i 

C 

<u 

c 

<u 

c 



S-i 

in 

OO 


G 

ID 

XJ 

X! 

w 



4-» 

S-I 

S-i 

•H 

X) 

TO 

TO 

u 

TO 

u 

TO 

in 

G 

0) 

TO 

3 

G 

<U 

N 

0) 

<D 


TO 


G 

4-S 

4-S 

CO 

TO 

s 

rH 

c 

E 

<u 

rH 

r-H 

CQ 

X! 

3 

4-S 


XJ 

4-1 

4-) 

4-1 

ec 

+J 


3 

00 

C/D 

r—l 

C 

c 

G 

TO 

S-i 

0) 

0) 

TO 

Cu 

4J 

S-i 

S-l 

TO 

<U 

•H 

•H 

•H 

< 

TO 


O 

3 

3 

<u 

TO 

0) 

•H 

S-i 

<u 

S-I 

S-i 

+j 

03 

0) 

o 

o 

cu 

3 

3 

c 

c 

H 

Q 


CJ) 

< 

< 

CQ 

C_> 

Q 

pH 

pH 

o 

o 

1—1 

1—1 


sc 

Si 

0-1 

CO 

CO 

CO 


G3 


a Shown in upper graph of Figure 4; b Shown in lower graph of Figure 4; c Shown by distance between 
the upper and lower limits of the shaded bars in the upper graph of Figure 4. 













street. It rests on the notion that the debt of a government is just like 
the debt of a family - in the future you will have to pay the interest on it 
and to repay the principal, and that will curtail your future ability to 
consume. The same must be true of the government too: a government 
debt is, therefore, bad in the sense that it 'mortgages' the resources of 
future generations. 

The naive no-burden view claims that the first view is all nonsense 
because the debt of the family is a debt due outside the family, but the 
debt of the government is due inside the country - we owe it to ourselves. 
Certainly, future generations will have to pay interest on the debt, but 
future generations will also receive that interest. All that takes place is a 
transfer from one group to another. According to the eighteenth-century 
French economist, le Chevalier de Melun, the public debt is like the debt 
of the left hand to the right hand, which in no way weakens the body. 
In other words, what matters is disposable income, or income produced 
minus taxes paid, plus interest received. When you raise the taxes and 
pay the interest, these two flows cancel out, and disposable income will 
not be affected. 

One other way of supporting the no-burden argument is to say that it 
is impossible for this generation to eat the wheat of future generations. 
We can only eat the wheat we produce. Thus, we cannot shift the burden 
from the present to the future. 

Next comes the sophisticated view that there is a burden. In es¬ 
sence, this view says that we can, indeed, affect future output. We can 
affect it by the amount of capital we bequeath to the future. Whenever 
the government runs a deficit, it finances that deficit by tapping current 
saving that would otherwise have gone into investment. Therefore, in¬ 
vestment is reduced. Investment is a source of income for the future, and 
that income has been lost. The disposable income argument is false', for 
while it is true that when we have a deficit one hand pays the other, if 
we did not have the deficit one hand would receive not from the other but 
from the capital that produces income. Therefore we have lost the income 
that the capital would have produced. 

This argument is sound, except that one may wonder whether the 
deficit's effect on investment might not be made up later. How do we 
know that the loss is permanent? It is useful at this point, I think, to 
fall back on another way of looking at the issue, which I refer to as a 
life-cycle perspective. 


61 


The life-cycle hypothesis (LCH) is, essentially, a model of the deter¬ 
minants of aggregate wealth. There used to be a view that the wealth we 
see around us was something we had inherited from the past - that was 
how the wealth came to be here. About thirty or forty years ago, the 
work I did on the life cycle pointed out that there was no need of be¬ 
quests to explain all the wealth that is in existence, for it could easily be 
generated by 'hump saving' - that is, people saving currently for later 
expenditure and for retirement. All that happens is that the same phy¬ 
sical assets change hands from the old who are selling them to consume to 
the young who are accumulating; you do not need transfers by bequest at 
all. 

In a more general formulation, one may expect that wealth is gener¬ 
ated in part by bequests and in part by life-cycle saving. In any event, 

this model tells us that the amount of wealth that exists in society is 

determined basically by society's income. That income determines both how 
much we allocate to the future and how much may be bequeathed. Wealth 
may also be influenced by forces such as income growth and population 
structure, but what is important is that one can rule out the proposition 

that the amount of wealth that society wishes to hold is systematically 

related to the size of government deficit. This means that a debt occupies 
a portion of the wealth that people would wish to hold, and thereby re¬ 
duces the portion of wealth available to finance capital. So if people were 
to hold in wealth, say, four times income, and if there were no debt, then 
capital would be four times income. But if there were debt equal to one 
times income, then the available capital would be only three times income. 
Essentially, then, what we have here is another way of looking at the 
crowding-out phenomenon. The government deficit crowds out capital 
because the desire for wealth is satisfied by the government debt instead 
of being satisfied by productive physical capital. 

This way of looking at things has many convenient aspects - for 
instance, it provides a measure of the overall burden of the national debt. 
Suppose a country has a deficit of $1 trillion. What is the burden, in 
terms of diminished capital, that this deficit imposes on future generations? 
It is equal to the return on the amount of capital it displaces. Now, 
according to LCH, $1 trillion of government debt should tend to displace 
$1 trillion of capital. If the interest rate is, say, 5 per cent per year, 
then the loss is essentially 5 per cent of $1 trillion, or $50 billion a year. 
In other words, to a good approximation, the flow of the burden can be 


62 


measured by the annual interest paid on the national debt. 

The basic argument also suggests its limitations and qualifications. 
First of all, the interest may not be a good measure of the burden, be¬ 
cause the capital may produce more than the interest rate on account of 
the tax wedge. With a corporate tax of 50 per cent, if the return to 
capital were 10 per cent, it would reduce to 5 per cent after tax, which 
also equals the interest rate. That is, when the government displaces $1 
billion of capital, there is not only a loss of $50 million in income to the 
holders, but also a loss of $50 million in tax revenue to the government. 
So the loss is not just the interest bill, but the interest bill plus the tax 
bill thereon. 

Again, the burden exists because the deficit displaces capital; but if 
the debt is used to finance government capital, then there will be a return 
from it that makes up for the private capital displaced. In principle, what 
matters is not the deficit but the deficit minus the value of income-earning 
assets, financial or physical, that the government may hold, with the 
return being calculated not necessarily in terms of cash return but in 
terms of social return. 

Two other points are, I think, relevant to the current discussion. 
One is that the argument I have developed about the burden of the debt is 
clearly valid when resources are fully utilized, so that the government 
deficit does, in fact, displace investment. If there is slack in the econ¬ 
omy, then the government deficit need not displace investment (and may 
actually increase it). Therefore, there need not be any burden, at least 
in the short run. One has to be careful here, because whether or not the 
deficit does displace other things is not just a function of the availability 
of resources, but also a function of the monetary policy being pursued at 
the same time. If we have slack resources and would like to use them, 
then there should be a monetary policy that permits investment and gov¬ 
ernment deficit to go on together, preventing the crowding-out. 

On the other hand, it may be that we have slack but do not want to 
use it. In many places, such as the United States, policymakers feel that 
the slack is needed at present to reduce inflation. In situations where the 
slack is wanted, the government deficit may again displace investment. 
But this applies only in situations like the current one where the slack 
must be maintained because of other services it performs. Let me remark 
in this connection that when I am asked by people whether we are going 
through a repetition of the Great Depression, I say that there is a great 


63 


difference between then and now: then we did not know how to control 
unemployment; now we do, reasonably well. We have been using the 
knowledge, particularly in the United States, to create a lot of unemploy¬ 
ment as a way to fight inflation. Whether that is a good way or not is a 
separate issue. At least it has worked, though with a lot of pain. 

Finally, there is the question of debt in the family - or outside the 
family. Suppose that instead of borrowing domestically, we borrow 
abroad. Does that make any difference to the burden argument? Even 
those who maintain that there is no burden from an internal debt must 
acknowledge that an external debt does cause a burden, since we have to 
pay the interest outside. My argument implies that, if we are using 
resources fully, it is immaterial, at least to a first approximation, whether 
the borrowing is inside or outside. If we borrow inside, we displace 
investment and lose the income from those investments. If we borrow 
outside, we continue to have those investments, but we have to pay the 
income to outsiders and, therefore, our society loses the same amount. 
The net result is the same, except for second-order refinements that 
depend on taxation and so on. 

It remains to consider the 'super-sophisticated' no-burden argument. 
This is an argument that probably has not gone much beyond sophisticated 
economists. It is sometimes traced way back to Ricardo, although I do not 
think this is a correct way of putting it. My opinion is that Ricardo did 
suggest the possibility but thought it was not practically relevant. The 
basis for the argument is as follows. Suppose we assume that life is 
infinite, as economists not infrequently do, and that all taxes are strictly 
proportional to income. Suppose further that the government intends to 
spend $100 and has a choice of either collecting a tax for $100 now or 
floating a loan for $100 and then raising taxes to pay the interest from 
now on. Whatever choice the government makes, everybody will be in 
exactly the same position. Paying $100 now and paying $5 forever have 
exactly the same present value, and the taxpayer really shouldn't care a 
bit which is chosen. So, tax financing and debt financing are equivalent. 
The burden is in the use of resources, not in the way their acquisition is 
financed. 

This argument, resting on infinite life, has recently been extended 
by Barro (1974) and others to the case in which life is finite, through the 
mechanism of inheritance. Suppose that you, at present, are making 
decisions for yourself and your heirs as to how you and your heirs will 


64 


allocate your resources to consumption over time - so much to you, so 
much to the heirs. Now along comes the government and says, 'We are 
going to reduce your taxes by $100 by raising a loan whose interest will 
be paid by future generations.' If you are a rational man you will say, 
'Why should I let the government decide that I should be better off - and 
my heirs worse off - than I originally decided? I've made a certain deci¬ 
sion and I am going to stick by it. To do this, I have only to use the 
tax rebate to buy the $100 worth of bonds that the government offers and 
pass them on to my heirs, so that when the collector comes around for $5 
a year they can simply pay him with the $5 they receive from the bonds. 
They will be as well off as before, and I will be as well off as before.' 
Again, in this kind of world, it is immaterial whether the government taxes 
or uses deficit financing. 

The only trouble with this argument is that it has nothing - or almost 
nothing - to do with the real world. Why is that so? There are many 
reasons, and let me quickly point out some. To begin with, just consider 
the simple case in which people are not identical: I happen to have no 
children and you happen to have twenty. In this case, the choice between 
taxing and borrowing can make an enormous difference. I, who have no 
children, would much rather see the tax put on children; you, who have 
twenty, would much rather have the tax paid now. It is clear that alter¬ 
native methods of financing are not neutral and that the differential effects 
on current consumption are not necessarily going to cancel out. 

A more serious shortcoming of the argument results from 'asymmetry'. 
Suppose I decide that the optimum allocation calls for my children to con¬ 
sume less and for me to consume more than our respective life incomes. If 
I could, I would elect for a negative bequest, or a transfer from my 
children to me. But I can't; the closest to optimum is to plan no be¬ 
quests. If an additional tax is put on me to retire debt, I would like to 
shift it to the future, which I could do only by reducing bequests. But 
with zero initial bequests, all I can do is stay where I am. So, in 
general, whenever the actual planned bequest is zero and the optimal is 
negative - calling for a transfer from children to parents - the choice of 
financing is relevant. Replacing taxes with deficits will transfer from the 
future to the current generation. In fact, it is precisely for this reason 
that the choice of financing may be a way for the current generation to 
achieve a desired outcome not achievable through private transfers. For 
example, if we finance a major war by debt, then the generation of the 


65 



war will come out of the war with debt that it can sell to support a higher 
consumption. It will, thereby, reduce the net saving of society in the 
future, since the new generation's saving will go into buying existing 
bonds instead of adding to the stock of capital. Thus, we can, by our 
choice of financing, control how the burden is distributed between the 
present and the future. 

In general, the empirical relevance of Barro's paradigm depends on 
the quantitative role of bequests as an intergenerational bridge. This is a 
very big issue, and I do not have time to discuss it fully here. It is 
closely related to the importance of bequests as a determinant of national 
wealth, which is currently hotly debated. Some people have recently 
claimed that most wealth is, in fact, the result of bequests. I have done 
some fresh work in this area, and my results suggest very strongly that, 
at least in the United States in this century, bequests are a modest por¬ 
tion of wealth between 15 and 20 per cent of the total. This estimate is 
derived from many different sources, from inquiries made through surveys 
to analysis of wills. They all point clearly in the direction of a relatively 
modest amount of transfers, it being understood that the transfers rele¬ 
vant here exclude those between husband and wife. These appear in wills 
as bequests but, of course, really occur inside the household. If they are 
taken out, and they are very substantial, the remaining bequests are 
relatively unimportant. 

Consequently, I would not expect bequests to significantly offset 
public financing decisions, and accordingly I see the national debt having 
an unfavourable effect on the stock of capital in terms of crowding out 
tangible capital. 

This conclusion is supported by some empirical work I have done in 
this area. We have estimates for the United States of private national 
wealth, or the wealth of households, going back to the beginning of the 
century, mostly due to the work of Goldsmith (1963). Since the beginning 
of the post-war period, we have official estimates from the Federal Reserve 
Flow of Funds series. We also have data on the amount of government 
debt - federal, state, and local - so we can ask the following question: Is 
there any evidence to support Barro's hypothesis, as against the LCH? 
How would we tell the two apart? 

If Barro is right, whenever the government borrows, people will 
decide to accumulate more so that they can offset the debt burden on their 
children. In other words, they want to pass on to their children addi- 


66 


tional wealth equal to the debt the government has floated, so that the 
future generations will be no worse off. Hence, one would expect wealth 
to rise, dollar for dollar, as the debt rises. This is the same as saying 
that the amount of wealth minus debt - call it private tangible capital - is 
independent of the amount of debt. That would be evidence that there is 
no burden, no crowding-out. On the other hand, there is evidence of 
crowding-out if wealth does not rise with debt and, hence, tangible capital 
declines with debt. 

An earlier study of mine (1966) examined the United States' evidence 
for the period 1900 to 1958, and found that the private wealth/income ratio 
is absolutely independent of the government debt/income ratio. In fact, 
the coefficient of the debt/income ratio, instead of being 1, as implied by 
Barro, is -.04, and statistically totally insignificant. I have recently 
rerun this test with alternative formulations and found approximately the 
same results. I have also extended the tests to the period 1900-80, and 
again the general result is that there is no evidence of the debt having an 
effect on wealth. Or, equivalently, private tangible wealth decreases 
roughly dollar for dollar with the debt. 

Barro has suggested to me that the test was unsatisfactory because, 
if people are truly rational, they will look not at the debt, but at the debt 
minus the assets the government holds. They know that whenever the 
government has one more dollar of debt they need to leave one dollar more 
of bequest, while when the government increases its assets by a dollar 
they need, on balance, to leave one dollar less. What really matters, 
then, is the net worth of the government. I have, therefore, run the 
same regressions using not the debt but the net worth of the government, 
and the results are exactly the same. There is no evidence that the net 
worth of the government has any effect. Interestingly enough, if you look 
separately at the effect of assets and liabilities, the assets have a slightly 
negative effect, consistent with Barro, but the liabilities also have a 
slightly negative effect - and both effects are very small and very insigni¬ 
ficant. On the whole, then, the data for the U.S. appear to reject 
Barro's hypothesis and to support the view that government debt crowds 
out private tangible wealth. 

Tests have also been carried out relying on a cross-section of coun¬ 
tries and looking for evidence that, where the deficit is large, saving is 
large. For, if Barro is right, where the government deficit on current 
account is large, savings should be large. Preliminary results show no 


67 


positive association between the private saving rate and the current 
account deficit. Indeed, the association appears slightly negative - where 
the deficit is large, saving is small and vice versa - but not significant. 
Again, there is absolutely no evidence in favour of the Barro hypothesis, 
and a strong presumption in favour of the hypothesis that debt crowds out 
capital. 

The conclusion that deficits have an unfavourable effect on the econ¬ 
omy, in particular through crowding-out of investment, raises the question 
of whether the recent exceptionally large, and widespread, cash deficits, 
of which we saw evidence at the start, could be a source of our present 
difficulties - high inflation, high unemployment, low productivity growth. 

I am going to suggest that the answer is very clearly negative, and 
the reason - somewhat paradoxically - is that there really are no signifi¬ 
cant 'crowding-out' deficits anywhere. This proposition might seem totally 
inconsistent with the figures discussed earlier. But there is no contra¬ 
diction, because the figures show the cash deficit, whereas the deficit that 
creates crowding-out is the rise in the real debt . The change in the real 
debt differs from the change in nominal debt, or cash deficit, because of 
the loss of real value of the debt due to inflation. When an adjustment is 
made for the depreciation of the debt, which is the same as computing the 
interest bill using the real, rather than the nominal, interest rate, one 
finds that this has a large impact on the deficit. 

Take as an example the United States, where the debt is around $1.2 
trillion. If inflation is, say, 10 per cent, it implies a $120 billion correc¬ 
tion. In 1982, the U.S. deficit, though exceptionally huge, is, in fact, of 
a very similar magnitude. Thus, after subtracting a correction for $120 
billion, one is left with roughly zero. This somewhat exaggerates the 
correction, because it should be applied not to the gross, but to the net, 
debt (i.e., net of financial claims). Still, if one does the correct calcula¬ 
tion for past years, one will find that, in fact, the correction exceeds the 
cash deficit itself. This is shown in Figure 2. The top curve is the 
unadjusted federal deficit, and the second, dotted line is the consolidated 
deficit of all governments - federal, state, and local. The consolidation 
already makes quite a bit of difference, because it happened that, while 
the federal government was running a deficit, the state and local govern¬ 
ments were running surpluses. The lower solid curve is corrected for 
inflation, and the broken line that follows it is corrected for inflation and 
for investment so that it represents deficit on current account. 


68 




One can see that by the time all the adjustments are made, there are 
essentially no deficits up to 1982. The modest exception occurs in 1975, 
but this deficit is largely of a cyclical nature. I pointed out earlier that 
it is all' right to have a deficit when there is a depression and that, in 
fact, a sound fiscal structure should lead to counter-cyclical deficits, as 
tax receipts should vary cyclically and expenditure should move counter- 
cyclically. If Figure 2 were adjusted for the cyclical deficit, as John 
Bossons and Peter Dungan have done for Canada in a paper (1983) that I 
understand is being debated these days, it would be clear that there 
really is no significant deficit anywhere in the United States. 

What we have just found for the United States holds roughly for 
other OECD countries and, in particular, for the nine shown in Figure 4. 
This exhibits a correction of the cash deficit for the depreciation of the 
debt due to inflation. The correction is represented by the shaded area 
subtracted from the column representing the cash deficit. Thus, the lower 
boundary of the shaded area represents the inflation-corrected deficit (if 
above the horizontal line) or surplus (if below). One can see, for in¬ 
stance, that for Austria, in the first period, there was very little inflation 
and hence a slight correction for inflation. In the second period, there 
was a larger cash deficit; corrected for inflation, however, the deficit 
becomes zero. Similar results hold for most other countries: correction 
for the effect of inflation generally wipes out the deficit and even turns it 
into a surplus. Indeed, it is apparent that, after correction for inflation, 
there were hardly any deficits on current accounts in the late seventies, 
except in Japan and, I am afraid, in Canada. Note that Japan, which has 
a slight deficit, is one of the countries with the lowest inflation rate. One 
can clearly see, from Figure 4, that there are no grounds for holding that 
the stagflation was produced by an outburst of uncontrolled deficits. 

On the contrary, one can argue just the opposite: that stagflation is 
the cause of deficits. Why? The main link comes from the fact that 
inflation means high interest rates, and high interest rates mean a large 
interest payment. It would be rational and consistent with sound fiscal 
policies to meet this enhanced interest bill by borrowing rather than by 
raising tax rates - unless one is bent on a policy of actually reducing the 
real national debt. The reason, of course, is that the higher interest 
reflects the inflation premium, which itself is equivalent to a repayment of 
the (real) debt. Thus, unless one intends to reduce the real debt, funds 
required to pay the inflation premium should come from borrowing, not 


69 


from taxes. As a result, a high rate of inflation could be expected to 
produce deficits of a magnitude related to the inflation correction. Simi¬ 
larly, high unemployment tends to cause high cyclical deficits. 

An examination of Figure 4 suggests that, in most countries, policy¬ 
makers have been willing to accept, in part, the higher cash deficits 
implied by stagflation under constant tax rates. In part, however, they 
have tended to contain the higher cash deficits through higher tax rates 
(partly the result of the so called 'bracket creep' effect of progressive 
taxation), with the outcome that, on balance, correctly measured deficits 
have been lower, not higher, in the more recent period. It is these 
considerations that lead me to reject the popular view that stagflation has 
been the result of irresponsible fiscal policies and to suggest, instead, 
that the observed association of deficits and stagflation reflects the causal 
link from stagflation to deficits, when tax rates and real expenditures on 
goods and services are held constant. 

That seems to be the essential story for the past. There has been 
no problem coming from deficits, though deficits can pose a serious 
problem, because there have been no significant deficits. What about the 
future? Here it may be useful to go back to the United States' chart 
(Figure 1) for it suggests that Mr. Reagan is trying very hard to make a 
non-problem into a problem. For a man who came to power telling us that 
deficit was the worst thing in the world, the huge deficits he is proposing 
for the coming years are quite a change and quite a puzzle. If they come 
at a time when inflation has largely abated and unemployment has largely 
disappeared, they will be truly of the crowding-out variety. There is no 
warranted correction that can make them shrink. Therefore, they will 
represent a serious problem. 

Of course, the implications of the large deficit for the U.S. are 
serious in many ways for the rest of the world as well. The displacement 
of investment will, presumably, mean a continuation of high interest rates, 
which, given the size of the country, will tend to mean damagingly high 
interest rates for the rest of the world. However, one should separate 
the past from the future. In the past, deficits have not represented a 
major problem. In countries such as Canada, which are not facing a huge 
armaments build-up, I suspect that the future does not look too bleak. 
For the United States, the deficits do look quite disturbing, and they have 
ominous implications for the rest of the world. But one must hope that 
something will be done before things get serious. 


70 


REFERENCES 


Barro, R.J. (1974) 'Are government bonds net wealth?' Journal of 
Poltical Economy 82:6, 1095-1117. 

Bossons, J. and P. Dungan (1983) 'Government deficits: too high or too 
low?' Canadian Tax Journal 31:1, 1-29. 

Goldsmith, R. and R.E. Lipsey (1963) Studies in the National Balance 
Sheet of the United States ' (Princeton, N.J.: Princeton University 
Press, National Bureau of Economic Research. Studies in Capital 
Formation and Financing, 11). 

Modigliani, F. (1966) 'The life cycle hypothesis of savings, the demand 
for wealth and the supply of capital,' Social Research 33:2, 160-217. 


Discussion 


QUESTION: I will start by asking Franco to cast his mind back to his 
birthplace, namely Italy, and consider the differences between a small, 
open economy such as Italy - or Canada - and a relatively closed economy 
such as the U.S. Do those differences affect your view of whether or not 
the deficit is a problem? 

FRANCO MODIGLIANI: You mean whether the deficit creates a different 
kind of problem? 

QUESTION: Whether or not crowding-out is a different issue in an 

economy such as the U.S. than it is in an economy such as Canada. 

FRANCO MODIGLIANI: I think that in the case of Canada, an open 
economy like the U.S. but much more open, the effects of a large deficit 
would be somewhat different, and, in particular, would result in less 
crowding-out of investment. This is because the rise in interest rates 
would be kept down by an inflow of foreign capital, which, in turn, could 
finance the government deficit without diverting domestic saving away from 
financing investment. Thus, to some extent the deficit, instead of 
crowding out investment, would 'crowd in' foreign capital or net imports. 
This would be true even if there were slack in the economy, as long as 


71 









the money supply were given and did not allow a significant expansion in 
domestic output. 

But even in this case the debt would impose a burden on future 
generations, which would have to pay taxes in order to pay the interest 
on the foreign debt. The amount of that interest would be commensurate 
with the return on capital that would be crowded out by debt in the closed 
economy - though probably somewhat smaller. 

QUESTION: It occurs to me that you look at each country somewhat in 

isolation - that the transfer from current net wealth to future net wealth 
tends to be parallel through all the countries. I am interested in what 
happens when the awful happens, when there is a financial breakdown and 
a particular country can't service its debt and becomes bankrupt. In 
other words, what happens when you put what you have told us into an 
international context? 

FRANCO MODIGLIANI: First, I have suggested that foreign debts can 
create a lot of problems, in spite of the argument that even if you run a 
deficit you don't crowd out investment because you can attract capital. 
This argument has been made in the United States - in fact, it appeared 
in last year's Economic Report of the President . Well, I say that even if 
you attract capital the deficit produces a burden, because you have to 
service the debt. On top of that, we are all aware that it is less prob¬ 
lematic to service domestic debt than it is to service foreign debt, because 
domestic debt can always be financed by inflation. I didn't really dwell on 
this, but there is a way of placing the burden on the current generation 
even with deficit financing, namely by covering the deficit by money 
creation. In that case, the burden falls on the current generation because 
money creation amounts to a tax. This tax is just like any other tax, and 
any expenditure that is financed by tax is a burden on the current 
generation. 

In the case of domestic financing, you always have that possibility. 
It is on this ground that Mr. Wriston of Citibank suggested at one point 
that there is never a risk in lending to a sovereign government. But 
really he went too far. With loans denominated in a foreign currency, 
there is a risk - since the sovereign government can't print that cur¬ 
rency. There may be no risk in lending to a state in cruzeiros if that is 
its currency and it can print cruzeiros. But when it comes to servicing 


72 



foreign debt, the borrower has the problem of finding the resources in the 
foreign currency. There would certainly be extra problems if the borrow¬ 
ing were on a large scale. 

I think the problems facing some of the countries that are in debt are 
not primarily problems of their having borrowed excessively in absolute 
terms. Their problems are really related to the fact that their debt is in 
foreign currencies. Problems have arisen from the combination of higher 
interest rates and reduced world activity, which has reduced exports. 

On the whole, I would say that there are special problems involved in 
foreign borrowing, and that we see some of these problems in the current 
situation. That situation is not, however, quite as bad as it may seem. I 
expect that some of the major problems will resolve themselves as inflation 
abates, interest rates are reduced, and there is a pick-up in international 
trade. 

QUESTION: There are a lot of people here who are practical investors, 

i.e., they put up their money. If a government continues to run deficits 
for a long period of time, year after year, that deficit goes up, it has to 
be serviced. The people who invest their money have a gut feeling that 
they are going to have to pay more and more taxes to service that debt. 
It gets to a point where they say, 'I am not going to invest any more 
money to make profits that are going to be continually taxed by the state 
to pay for the deficit.' Why should people continue to invest their capital 
when they see a percentage of the money that is being taxed going up and 

pushing their taxes up to pay the debt and their profits subsequently 

being more highly taxed? 

FRANCO MODIGLIANI: You said people should be concerned if the 
amount of interest governments have to pay is rising faster than their tax 
bases. I quite agree with you. If there isn't a good reason for this to 
happen, such as a very temporary bulge in the deficit, due to a depres¬ 
sion, or an effort to build capital that will produce the income that will 
pay the interest, then I would certainly say there were grounds for 

concern. But I have been trying to distinguish between the fact that the 

nominal debt and the nominal interest bill rise, and the fact that they rise 
faster than income. Now, in the current situation one finds that the 
deficit, corrected so as to measure the rise in the real debt, is not rising 
relative to real income - in fact, it may be declining. 


73 



Please refer again to Figure 2. It shows that for the United States 
there was no problem, even though there were cash deficits throughout 
the seventies and the debt was rising, as is shown in Figure 1. The 
lower line in Figure 1 measures the U.S. federal government's net worth, 
i.e., the difference between the government deficit - the upper line - and 
government assets, consisting of claims on the private sector and of phys¬ 
ical assets. One can see that, though the debt and the interest on it 
have been rising, nonetheless, in real terms, there has been an increase 
in the net worth of the government. The point is that the rise in debt 
and interest largely reflects inflation; inflation has caused income to rise 
even faster than debt, so that the ratio of debt to GNP has fallen steadily 
since the war, when it reached almost 2.0, to where it is now, approxi¬ 
mately 0.3 (including the state, local, and federal governments). In 
short, a rise in real debt, especially if faster than real income, is a valid 
reason for concern. But one must be careful to distinguish what is really 
going on from what seems to be going on, because of illusion due to in¬ 
flation. 

I do not have the figures for Canada on hand, but I understand from 
what John Bossons has been telling me that the ratio of debt to income has 
declined in Canada also until very, very recently and has just now been 
rising under very depressed conditions. If the depressed conditions go 
away and you find that the ratio continues to rise, then, I think, you will 
have a very good case for saying: 'Hey, what are we doing here?' 

COMMENT: I am getting away from the topic, but it strikes me that 

what you are proposing assumes great economic literacy on the part of the 
people of the United States and Canada. I submit that there isn't that 
economic literacy, that people in fact don't differentiate between real and 
inflationary debt, and that those who are making the investments, on 
either an individual or a corporate basis, look only at the raw figures and 
say, 'Reagan has gotten us into a trillion dollars of debt, Trudeau has 
gotten us into a $28 billion deficit, and I ain't going to invest.' 

FRANCO MODIGLIANI: I do not join some of my colleagues, such as 
Barro and other followers of the so-called rational expectations hypothesis, 
in believing that people are very rational. I don't believe they are, and I 
don't think we should expect them to be when they vote. But what I am 
asking the public to understand is something simple, something I hope 


74 


anyone can understand. My point is that one should never look at just 
the debt: instead, one should look at the ratio of debt to income. To 
compute the debt/income ratio, one doesn't have to deflate by any price 
index. The debt/income ratio, in real or nominal terms, is the same. 
Thus, one only needs to know long division, or to own a small calculator! 
What I am saying is that the debt can be growing quite fast, say at 10 per 
cent per year, and the debt/income ratio not at all, because income is 
growing at 12 per cent. 

It seems to me that this is a minimum that anybody can understand. 
Don't look at the interest, because the interest is inflated, and the reason 
why is a hard concept to understand. But just stress that, despite every¬ 
thing, the debt/income ratio is not rising. 

QUESTION: I am interested in the discount factors that you use in 

inflation-adjusting the deficit. You assume that there is some kind of 
constant relationship between the rise in incomes and the capacity to pay 
the deficit. In Canada we have an indexed tax system, and I wonder if 
that would alter your view of the assessment of the valuation of the defi¬ 
cit, given that the capacity of the government to repay the debt is af¬ 
fected by taxes. 

FRANCO MODIGLIANI: Let me answer your question in two steps. One 
is easy and the other more complicated. The easy one is this: indexation 
in no way changes what I have said, because indexation means that, with 
respect to the effect of inflation, taxes remain proportionate to income. 
Indexation is meant to avoid the bracket creep which causes taxes to rise 
faster than income. With indexation, you should find that government 
income rises in proportion with income, so there is no special problem 
there. 

By the way, it should be realized that, where there is no indexation 
of the brackets, the change in brackets generally tends to be done 'by 
hand'. In the United States we have had repeated reduction in tax rates 
to offset the inflation. Now indexation is much better, because when you 
have to do it by hand you always get into a fight about who is going to 
get the benefit. There really ought to be no question; inflation should 
not change the position of taxpayers. But in reality, arguments develop: 
for instance, that the rich can stand inflation and the poor can't, there¬ 
fore we have to change taxes, rather than the tax rate, equally. 


75 


There is another aspect I must touch upon. When I say that one 
should correct the government deficit for inflation, one may object that the 
government still has to finance the full interest. The point is that if 
people behave rationally and reasonably, they will be increasing their 
saving roughly to the extent to which the government needs to borrow 
more. Why? Because when they get a 10 per cent higher interest rate 
because of 10 per cent inflation, they should realize that the extra 10 per 
cent is not income but a compensation for loss of principal, and that they 
have to save it in order to maintain their principal. There should, thus, 
be a matching relation between the increased cash saving measured in the 
conventional way and the inflation-correction of the debt. In other words, 
there is no crowding-out from the borrowing to pay the inflation premium 
because there is additional saving corresponding to the additional interest 
paid by the government. 

There is, however, a question of whether people behave this way. 
Here the evidence is complicated - it's not an easy thing to establish. In 
the case of Canada, the evidence seems to suggest that they do - that 
people have increased their savings by at least this amount. I have 
looked at many other countries and the evidence is not quite as clear, but 
by and large there seems to be evidence that with inflation there has been 
a rise in savings sufficient to compensate for additional government inter¬ 
est. In the U.S., the evidence is mixed - perhaps it hasn't happened, 
perhaps it does not occur fully in the short run. Yet I think that, given 
enough time, people will learn to adjust. But in the short run, one could 
have some crowding-out effect, despite the fact that this additional in¬ 
terest only goes to compensate for loss of principal. 

QUESTION: You seem to indicate. Professor Modigliani, some divergence 

of opinion about the past and present as far as the United States' situation 
is concerned. For example, you suggest that the ratio of debt to GNP has 
been declining in the United States, but if I understood you correctly you 
also suggested that there is a structural deficit in the United States. 
That is, if inflation rates dropped to zero and if the U.S. went back to 
some kind of full employment of resources, there still would be a deficit. 
Financial markets in the United States are still telling us that there are 
very high real interest rates, especially in the long run if you subtract 
from the nominal rate an inflation rate of something like 2 or 3 per cent. 
Does your analysis suggest whether that high real rate is warranted or 
not? 


76 


FRANCO MODIGLIANI: First of all, let me repeat for the sake of clarity 
that in the U.S. there are some substantial deficits in sight, whereas 
there were none in the recent past. The prospective deficits are perfectly 
understandable, given that we have a huge build-up in our defense ex¬ 
penditure and a very large cut in taxes. It is that huge build-up, the 
fastest ever as I understand it, that causes the real problem, combined 
with the fact that taxes have been and are still being cut. It is a com¬ 
bination of lower receipts and higher government expenditure that will 
cause the future 'structural' deficit where there was none before. 

Next, I certainly do believe that concern about the future deficit is a 
major cause of what appear to be unusually high long-term rates compared 
with inflation, though I think perhaps this concern is not the only reason. 
I think there is probably, on balance, some rise in the spread between 
shorts and longs due to the fact that the behaviour of the long-term rates 
has been so volatile that, in fact, long-term bond issues have become 
nearly as risky as stocks - the change in volatility of long-term issues has 
really been enormous. The latest game in town among the managers of 
portfolios is to play the bond market. 

In addition, warranted levels of real rates have been raised signifi¬ 
cantly, in my view, by the very important investment incentives provided 
in recent legislation. But, on the whole, I think the spreads are also 
quite high. Are they excessive? Yes, in the sense that, come two or 
three years from now interest rates will not, I would bet, be as high as 
those long-term rates imply. I do feel that they are exaggerated. For 
several years I have been advising people to stay short, but now I would 
not object to lengthening the maturity. 

In that sense, I do believe interest rates are somewhat on the high 
side. I think they will be stuck there for a while, and until there is some 
solution to this deficit they may create a problem. 


77 


The Ontario budget deficit: 
a cause for concern? 

D. A. L. Auld* 


The moral stigma attached to budget deficits is a strange phenomenon, 
since it is applied to what is nothing more than an accounting term devoid 
of any normative overtones. It is derived in part from an analogy with 
the family, where 'living within one's means' is viewed as a standard of 
acceptable behaviour. However, few people would consider it immoral to 
assume a mortgage for the purpose of 'owning' a home. On the contrary, 
home ownership is often seen as an important part of a proper family 
environment. It is true that corporations showing losses must restore 
profit to the income and expense statement lest they eventually go bank¬ 
rupt, but few would refuse a corporation the right to borrow money for 
certain purposes if it had the potential to repay the debt. 

The present attitude toward budget deficits may reflect the frustra¬ 
tion by a society that has watched deficits and rising unemployment occur 
simultaneously. As long as deficits appeared to stimulate growth and 
maintain ever-increasing standards of living, philosophical or moral oppos¬ 
ition to deficits was never taken seriously. That has obviously changed, 
not only in Ontario but in Canada as a whole and throughout the devel¬ 
oped industrial nations. In a relative sense, concern over Ontario's defi¬ 
cits of 1 to 2 per cent of gross provincial product seems unfounded, given 
recent deficits in France (4 per cent), Italy (15 per cent), Japan (5 per 
cent), the U.S. (5 per cent), and the U.K. (7.5 per cent). 1 

The concern over public sector budget deficits focuses on four major 
issues: the impact of deficits on the intertemporal distribution of the 

* Chairman, Department of Economics, University of Guelph. The author 
gratefully acknowledges the research support of Douglas Crocker and Aly Sy 
of the Ontario Economic Council. Meetings with Bob Christie and Cathy 
Dowling of the Ontario Ministry of Treasury and Economics contributed to 
the preparation of this paper. A lengthy and stimulating discussion with 
David Conklin of the Council helped to shape some of the research agendas. 


78 




costs of public goods; the potential effect that deficits, financed by the 
issue of high-powered money, have on inflation; the impact of foreign 
borrowing on exchange rates; and the impact that deficit financing has, at 
least potentially, on interest rates and the ability of governments to repay 
debt on time. With the exception of the relationship between deficits and 
the money supply, all of these concerns apply to the provincial as well as 
the federal budget deficit. Financial analysts have concluded that 'The 
fear of huge government deficits ...[has] plunged the capital markets into 
a state of disarray...' ( The Globe and Mail , 25 January 1983, B3.) There 
is also considerable concern that Canada's public sector borrowing overseas 
($8.7 billion in 1982) will, through its future drain on the current account, 
pull the Canadian dollar down ( The Globe and Mail , 24 January 1983). 
The Ontario situation is no exception. The announcement in May 1982 that 
Ontario was planning a deficit of $2.23 billion elicited the comment from 
one broker that such a deficit was bound '...to be an obstacle to falling 
interest rates.' ( The Globe and Mail , 17 May 1983, Bl.) 

While we recognize that a provincial budget deficit may have undesir¬ 
able consequences, it is our contention that current concerns spring from 
a less than thorough analysis of the province's deficit. An adequate 
analysis would have to take into full account each of the following factors: 

the impact of inflation on the provincial budget, ^ 
the cyclical nature of the budget result, and 
the nature of public capital expenditures. 

All of these issues can be summarized by a single algebraic expression 
of the provincial budget identity, shown in Table 1. First, the impact of 
inflation, particularly accelerating inflation, affects the left-hand side of 
this identity by pushing up interest payments, contributing to the need to 
find additional revenue. As we shall see later, the effect is dramatic. 
Second, a recession characterized by reduced income growth causes trans¬ 
fers, G(Y), to rise and tax revenues, T(Y), to slow or fall. Third, the 
k 

inclusion of G on the left-hand side of the identity suggests that, net of 
depreciation, capital spending at least ought to be balanced on the right- 
hand side by the issue of debt. 

The purpose of this paper is to address each of these issues in light 
of the past and current deficits of the Ontario government. The paper 
concludes with a short note on the manageability of the provincial public 


79 






TABLE 1 

The budget identity 


[G° + G(Y t ) + G* + R t ] = [T(Y t ) + D t ] 


G° = public expenditure on goods and services 


G(Y) = transfer payments, sensitive to level of income, 


= gross fixed public capital formation. 


T(Y) = tax revenue 


D = new bonds, borrowing, or reductions in reserves. 

R = interest payments on past bond issues and equal to Z^ (r+7t) 

where 71 is the actual or expected rate of inflation, r is the 
real rate of interest, and Z represents the stock of public 
debt. 


debt. However, before we proceed to an analysis of the interaction 
between inflation and the budget balance, a brief summary of the historical 
development of the budget balance seems in order. 

A BRIEF HISTORICAL REVIEW 

From an historical perspective, the concern over budget deficits that has 
arisen in the past six or seven years is somewhat of a puzzle. In 1960, 
for example, the Ontario deficit was 10.4 per cent of public expenditures, 
and in 1962 it rose to 12.1 per cent. Going back even further, the deficit 
was 9.6 per cent of expenditures in 1950 and 7.2 per cent in 1955. By 
comparison, the deficit was 13.0 per cent of expenditures in 1977-8 and 
4.6 in 1980-1, and is expected to be 9.8 per cent in 1982-3. 2 Over the 
past twenty-four years, there have been deficits in all years but two, 
ranging from 1.0 to 15.9 per cent of provincial expenditure (see Table 2). 
The average deficit over this period has been 7.2 per cent of provincial 
government expenditure, and the figure has risen only slightly since 
1969-70 (see Figure 1). 


80 




TABLE 2 

Ontario budget result as a percentage of total provincial expenditure 


1959-60 

-10.4 

1971-2 

-10.4 

1961 

-11.5 

1973 

- 6.7 

1962 

-15.1 

1974 

- 8.9 

1963 

- 6.7 

1975 

- 9.9 

1964 

- 5.1 

1976 

-15.9 

1965 

- 2.1 

1977 

-10.6 

1966 

- 1.0 

1978 

-13.0 

1967 

+ 1.1 

1979 

- 8.2 

1968 

- 4.7 

1980 

- 3.7 

1969 

- 9.5 

1981 

- 4.6 

1970 

+ 3.5 

1982 

- 7.6 

1971 

- 2.6 

1983 

- 9.8 


SOURCE: See Figure 1. 


In absolute terms, however, deficits have grown considerably in 
recent years, reflecting the growth of the public sector. The emergence 
i in 1975-6 of a deficit of almost $2 billion was the 'break' with historical 
standards that precipitated the call for lower deficits. The deficit re¬ 
mained above $1 billion for the next three years and then fell below that 

I 

figure for two years. With the doubling of the deficit in 1981-2 over the 
1980-1 figure and the estimated $2.2 billion deficit in 1982-3, there have 
been renewed demands by the private sector and the opposition that the 
budget be steered back to a balance between revenues and expenditures. 

Figures 2 and 3 summarize the changes in the annual budget result 
since 1970-1. These changes suggest that the Ontario budget is highly 
responsive to changes in the level of activity in the economy, changes in 
fiscal instruments, or both. There seems to be no tendency for budget 
results to change more widely over time. The swings in budget results in 
the 1975-82 cycle of economic activity are no larger in relative terms than 
those during the 1961-7 cycle. As a percentage of Gross Provincial 
Product (Figure 4), the deficit rose significantly between 1969 and 1976, 
but the trend to 1980-1 was downward. It has risen again over the last 
three years. 


81 











Figure 1 

Ontario budget result as a percentage of provincial government expenditure 

Per cent 



Year 

SOURCE: Ontario budgets, 1965,1969, 1977, and 1982. 


THE IMPACT OF INFLATION ON PUBLIC DEBT CHARGES AND THE 
DEFICIT 


We now turn to the first of the three issues outlined earlier and attempt to 
answer the question: how has inflation, and in particular accelerating 
inflation, affected the budget balance of the Province of Ontario? 

Inflation grossly distorts the real effects of debt financing and the 
cost of carrying that debt in both businesses and government. Consider a 
government that wishes to finance $100 million worth of public capital 
investment in period 1 and again in period 2 by the issue of long-term 
bonds. We assume a real rate of interest of 2 per cent and, in Case I, no 
inflation. In Case II, we assume a 10 per cent rate of inflation. In Case 


82 




Figure 2 

Budget deficit ($ millions) 


$ Millions 



Figure 3 

Change in deficit ($ millions) 

S Millions 



Year 

SOURCE: See Figure 1. 


83 






















































































Figure 4 

Ontario deficit as a percentage of gpp 


Per cent 



SOURCE: See Figure 1. 


Year 


I, the public debt has doubled by the end of the second period, as have 
total debt charges (see Table 3). In Case II, it is assumed that to main¬ 
tain the real value of the investments in a world of 10 per cent inflation, 
it will be necessary to issue $110 million worth of bonds in period 2. This 
results in an interest cost in period 2 of $13.2 million on the $110 million 
and cumulative debt charges of $15.2 million associated with a total debt of 
$210 million. Thus, while the total debt has increased a little more than 
two-fold, interest charges have accelerated more than sevenfold! These 
nominal interest charges, which are carried as current expenditures on the 
expenditure side of the budget, can very quickly contribute to a substan¬ 
tial increase in government expenditure. If the rate of inflation continues 
to accelerate, then interest payments will rise at a much faster rate than 
the increase in public debt. 

In the context of the previous example, suppose that the inflation 
rate remained at 10 per cent in the third period. To maintain the constant 
dollar value of public investment of $100 million per year, $115.5 million 
would need to be raised. The nominal interest on the total debt of $331 


84 




















































TABLE 3 

The effect of inflation on debt costs ($ millions) 



Case I (no 

inflation) 

Case II 

(10% inflation) 


Period 1 

Period 2 

Period 1 

Period 2 

Value of bonds 

issued 

100 

100 

100 

110 

Real interest 

2 

2 

2 

2.2 

Inflation 

compensation 

0 

0 

— 

11 

Nominal interest 

2 

2 

2 

13.2 

Cumulative debt 

charges 

2 

4 

2 

15.2 

Total debt 

100 

200 

100 

210 


million would be $27.72 million. Thus, while total nominal debt has in¬ 
creased 3.31 times since the first period, debt charges have increased 
14.86 times. Even compared to period 2, the debt charges in period 3 are 
increasing faster than the debt itself. If the inflation rate is unchanged 

over all periods, then the rate of increase in debt will parallel the rate of 

increase in debt charges. 

Figure 5 provides a summary illustration of how increasing inflation 
affects the relationship between debt charges and total public debt. The 
figure illustrates the annual growth of interest paid and total provincial 
funded debt as an index using 1970-1 as the base year (1970-1 = 1.00). 
Until 1975-6, the two series moved closely, then diverged as interest costs 
jumped from a range of 8 to almost 10 per cent between 1972-3 and 1975-6 

to 15 per cent by 1980. By 1982-3, debt charges were 6.98 times their 

1970-1 level, while total public debt was 3.58 times its 1970-1 level. If 
inflation had remained constant , at its 1968-71 average, public debt 
charges would have been approximately $1 billion lower in 1982-3 3 (Table 
4). 

While the impact of inflation on total interest charges may not be as 
acute if the rate of inflation is constant, it nevertheless distorts the 
current budget of the government. The portion of the interest paid on 
the provincial debt that represents the inflation premium required to 
induce people to hold bonds is in a fundamental sense a capital transaction 


85 








Figure 5 

Index of public debt and debt charges 1970-71 to 1982-83 


Index 



SOURCE: See Table 4. 


that retires part of the debt before the actual redemption date. In other 
words, after a year of 10 per cent inflation, the real debt of the govern¬ 
ment has declined by 10 per cent. If governments had well-developed 
capital budgets, the capital transfer associated with the inflation premium 
would be included in the capital, not the current, budget. 4 

Although there is general agreement that one should consider the 
impact of inflation on the government's real debt position, there is no 
agreement as to just how the adjustment should be carried out. Concep¬ 
tually, it would make sense to use what had been the expected rate of 
inflation when each issue of debt was placed on the market. Adjustments 
would be slow, particularly if the term to maturity was long, and it would 
require agreement on how to measure expected inflation. A more 'crude' 
method is to base the inflation adjustment on the extent to which real debt 
is affected each year by the current rate of inflation. This method may 
'overstate' the adjustment in the short run, when inflation rates are high; 
nevertheless, it is satisfactory for our purposes here. Thus, the actual 
nominal deficit, D, is adjusted to obtain an inflation-adjusted deficit D* by 


86 






II 


TABLE 4 

Ontario deficit and funded public debt ($ million) 


Year 

Public 

debt 3 

Index of 
growth of 
debt 

Interest 

paid 

Index of 
interest 
paid 

Actual 

deficit 

1970-1 

4,237 

1.00 

311 

1.00 

569 

1971-2 

5,270 

1.24 

380 

1.22 

1,021 

1972-3 

6,300 

1.49 

408 

1.31 

744 

1973-4 

7,008 

1.65 

525 

1.69 

708 

1974-5 

7,844 

1.85 

589 

1.89 

977 

1975-6 

9,818 

2.31 

725 

2.33 

1,799 

1976-7 

10,895 

2.57 

890 

2.86 

1,319 

1977-8 

12,364 

2.96 

1,033 

3.32 

1,762 

1978-9 

14,039 

3.31 

1,233 

3.96 

1,180 

1979-80 

15,196 

3.58 

1,388 

4.46 

584 

1980-1 

16,214 

3.82 

1,595 

5.12 

803 

1981-2 

17,592 

4.15 

1,838 

5.90 

1,560 

1982-3 

19,409 

4.58 

2,172 

6.98 

2,232 


a Excluding Ontario Hydro 

SOURCES: Miller (1979, Table C13; 1982a, Table CIO). 


D t = D t' ¥' 

where is the stock of outstanding debt in year t-1 and is the rate 

of inflation measured by the implicit GNE deflator. 5 

The impact of making such adjustments is shown in Figure 6. The 
actual and inflation-adjusted budget balances parallel each other until 
1975-76, when, with increasing rates of inflation and higher nominal inter¬ 
est, the series diverge. By 1982-3, the inflation premium associated with 
a funded debt of $19.4 billion is in the neighbourhood of $1.4 to $1.8 
billion. A very large portion of the 1982-3 deficit is therefore necessary 
to finance this capital transaction, due to the impact of inflation on the 
real value of the debt. The cash requirement associated with the deficit 
is, of course, very real, but it is not due to 'uncontrolled' or 'wasteful' 
public expenditure. 


87 







Figure 6 

Impact of inflation on budget deficit 

$ Billion 



THE CYCLICAL SENSITIVITY OF THE ONTARIO BUDGET 

Fluctuations in nominal interest rates have not been the only cyclical 
phenomenon affecting the budget result. The past twelve years have also 
seen cyclical fluctuation in output and employment and fluctuation in 
either factor can cause the budget result to swing sharply toward or away 
from a surplus. The present section attempts to estimate the significance 
of these forces. 

To some observers, the sudden development of a budget deficit or an 
increase in the budget deficit is an indication of uncontrolled spending or 
the introduction of new programs without the appropriate level of tax 
financing. While the scenario may be true, it is necessary to separate 


88 











rhetoric from fact before such a conclusion can be advanced. One of the 
steps that must be taken is to distinguish between a passive deficit, 
induced by a change in the level of economic activity, and a deliberate 
deficit, brought about through tax reductions and/or increases in public 
expenditure. Furthermore, a deliberate increase in the deficit (or reduc¬ 
tion in surplus) may be the result of conscious efforts to stimulate eco¬ 
nomic activity. The use of deficits for this purpose has been the explicit 
policy of the Ontario government for more than a decade. 

For the past three years, this Province has run large cash deficits in 
order to create new jobs and incomes. Ontario's fiscal policy has sub¬ 
stantially reinforced federal action in the national economy. (White 
1973, 5.) 

...the Budget I place before you...provides for an expansionary deficit to 
stimulate the economy... (Miller 1982, 21.) 

This section of the paper addresses two questions. First, to what 
extent is the Ontario budget result a response to changes in the level of 
economic activity? Second, have deliberate deficits been counter-cyclical? 

In order to measure just how sensitive the budget result is to 
changes in the level of economic activity, it is necessary to 'purge' the 
budget result of the effect of discretionary changes in fiscal instruments. 
Such an exercise also allows us to estimate the relative demand-creating or 
demand-destroying impact of deliberate fiscal policy. To carry out these 
estimates for the period 1969-78, we have used the technique of the full- 
employment budget surplus (FEBS). 6 There are no recent estimates of 
full-employment GPP, so for subsequent years we have used the alternative 
technique described in the appendix to this paper. 

The FEBS in a given year is the budget result that would occur if 
the economy were operating at its potential and the actual fiscal parameters 
(tax rates, benefit rates, and expenditure on goods and services) were 
fixed. Suppose that the FEBS in one period was $+200 million. If in the 
next period the FEBS was again $+200 million, we can conclude that no 
discretionary fiscal policy was undertaken in this period or that action on 
the tax side was offset by action on the expenditure side. 7 Had the FEBS 
declined to $+100 million, this decline would be a clear indication that 
stimulating policies were applied. The difference in the actual budget 
result between one period and the next would indicate not only discretion¬ 
ary policy effects (if there were any), but also the effect of changes in 


89 













TABLE 5 

Explanation of deviation of full-employment budget surplus 


Time 

period 

A 

GPP 

Case I 

F 

GPP 

r a 

FEBS 


1 

100 

110 

+50 

+ 100 


2 

90 

110 

-75 

+ 100 

Time 

period 

A 

GPP 

Case II 

F 

GPP 

r a 

FEBS 


1 

100 

110 

+50 

+ 100 


2 

90 

110 

-100 

+ 70 


actual GPP on the budget result. 

The example given in Table 5 may help to explain this approach. 

A 

Consider Case I. When actual gross provincial product (GPP ) declines, 

A 

the actual budget result (R ) moves from a surplus (+50) to a deficit 
(-75) as tax revenues fall and transfer payments rise. None of this is 
due to any discretionary action, since the full employment budget surplus 
(FEBS) remains unchanged at 100. If, in the face of mounting unemploy¬ 
ment, the provincial government reduced taxes, forgoing some of its reve¬ 
nue, the scenario might be as shown in Case II. In this example, there is 
a greater decline in the actual budget result and a decline in the FEBS, a 
clear indication of discretionary policy, the effect of which is included with 
the induced budget reponse in the change in the actual budget result. 
Attempts to maintain the $50 million surplus by higher taxes or lower 
spending in the face of a recession would only deepen the recession and 
possibly contribute to a higher deficit in the short run. 

Having briefly reviewed the concept of the FEBS, we turn to its 
application to Ontario in the 1969-78 period. Table 6 shows actual and 
full-employment budget results for 1969-78. The actual deficit, D, in any 
year when there was a change in discretionary policy would have been 
different in the absence of any policy change. For example, an actual 
deficit of $500 million would have been larger if a tax reduction had not 
been implemented in that same year. Thus, the existence of discretionary 


90 














TABLE 6 

Actual budget results and FEBS in Ontario (National Accounts basis) and 
output gap, 1970-8 


Year 

FEBS 3 

Change 

Actual 

budget result 

Change 

Output 

gap 

Change 




$ millions 


% change 

1969 

+272 



+216 


+2.5 


1970 

+ 117 

-155 


+ 52 

-164 

+ 1.2 

-1.3 

1971 

-229 -342 

-346 


-380 -362 

-432 

-0.5 

-1.7 

1972 

-311 -408 

- 82 

- 66 

-386 

- 24 

+0.9 

+0.4 

1973 

-411 


- 3 

-282 

+ 104 

+2.9 

+2.0 

1974 

-359 


+ 52 

-219 

+ 163 

+ 1.8 

-1.1 

1975 

-1,301 


-942 

-1,464 

-1,145 

-3.5 

-5.3 

1976 

-1,241 


+ 87 

-1,392 

+ 72 

-1.9 

+ 1.6 

1977 

-886 


+328 

-1,279 

+ 113 

-3.8 

-1.9 

1978 

-1,084 


-198 

-1,602 

-323 

-4.1 

-0.3 

a The 

1969-72 and 

1971-8 

FEBS 

estimates were , 

arrived at 

in different ways, 

causing 

a split in 

the series. 






SOURCES: Jones, Bardecki, and Hull (1977); White (1973); Miller (1979 and 

1982a); Ontario (1977). 


policy, if it is counter-cyclical, dampens the rise in the deficit or fall in 
the surplus. To obtain a measure of the budget result's automatic re¬ 
sponse to the cycle, we must account for the 'feedback' effects of discre¬ 
tionary policy as indicated by the change in the FEBS. The change in the 
FEBS is the net, unweighted injection of aggregate demand into the pro¬ 
vincial economy. 8 If this is multiplied by the disposable income multiplier, 
we have a rough measure of the effect of the change in discretionary 
policy on GPP. If this, in turn, is multiplied by the marginal response of 
provincial revenue to a change in GPP, we have an estimate of the feed¬ 
back effect on the budget deficit or surplus. Therefore, the hypothetical 
deficit in the absence of policy (D ) is 

D H = -(AFEBS) (8) [ *T pp ] + D 

where 6 is the multiplier, T represents tax revenue, and D is the actual 
deficit. 


91 












Figure 7 

Actual deficit and deficit in the absence of discretionary policy 

$ Millions 



Actual and hypothetical deficits since 1970 are shown in Figure 7 (the 
'split' in the series at 1978 is due to the lack of consistent data for the 
entire period on a calendar year basis). The pattern of hypothetical and 
actual deficits show clearly that the Ontario budget is very responsive to 
changes in economic conditions, such as the recessions of 1975 and 1981-3. 
In short, deficits are to a large degree created by the general economic 
environment. The area between the lines is what the budget result would 
likely have been if full employment had been maintained in the 1975-82 
period. 

The second question we set out to answer was the following. If 

changes in the deficit were deliberately engineered for counter-cyclical 
reasons, have these changes been at least qualitatively successful? To 

answer this question, we compared the change in the FEBS with the 
change in the official output gap for the period 1970-8. (It was necessary 
to make an estimate of the output gap for the fiscal years 1978-9 to 

1981-2. These estimates are explained in detail in the appendix.) Effi¬ 

cient counter-cyclical fiscal policy suggests that as the economy moves 
away from full employment, discretionary policies would be expansionary, 


92 













as indicated by a negative change in the FEBS or a negative change in 
policy-induced budget result. The reverse would be expected as the gap 
narrowed. Observations based on these data should give us a trend if 
fiscal discretionary action has been stabilizing. 

Figure 8 plots actual observations of the change in FEBS or policy- 
induced deficit and change in the GNP gap for the 1969-70 to 1981-2 
period. There appears to be one major instance, in 1977, when discre¬ 
tionary policy was not counter-cyclical. The increase in the FEBS in 1977, 
when the economy moved farther away from potential GNP, was, we 
suspect, largely due to public pressure on the government to reduce its 
deficit, which had been expanded in 1974-6 in response to the widening 
gap between actual and full-employment GPP. 

In three other years, the direction of policy appears to have been 
pro-cyclical. This tendency was pronounced only in 1979-80 (see Table 
7), when the deficit was deliberately increased even though unemployment 
was declining. However, although the output gap was closing, it was still 
high by historical standards, and it may have been felt that in this 
context, additional deficit-producing policies were justified. On balance, 
discretionary policy over this period appears to have been counter¬ 
cyclical. 

PUBLIC CAPITAL SPENDING AND THE DEFICIT 

While inflation and cyclical adjustments to government budgets have been 
debated and calculated for several years, it is only in the last year or two 
that there has been a renewed interest in the role of capital expenditures 
within the framework of the conventional government budget. It is this 
topic that we turn to now. 

If we accept the notion that economic downswings, particularly if 
accompanied by high interest rates, lead to temporary deficits or increased 
deficits that should not be offset by tax increases, should a balanced 
budget be the norm when the economy is close to potential? For several 
years now the Ontario government has championed both the principle of a 
balanced budget and borrowing. In his 1978 budget speech the treasurer 
said, 'We are targeting for a balanced budget to make room in the economy 
for the private sector to grow and to flourish.' (McKeough 1978a, 10.) 
Later, in one of the Budget Papers, the statement is made that: 


93 



Figure 8 

Measuring the impact of discretionary policy 


- +600 


Pro-cyclical 


- +300 


I I I I I I 1 L 


-5.0 -2.0 


A 

A 


Counter-cyclical 

A 


A In FEBS ($ millions) or 
policy-induced budget result 


Counter-cyclical 


i i r I i I i I l L 


+ 1.0 +4.0 


Change in 
T GPPa 


LGPPf 


- 1.00 100 


- -300 


Pro-cyclical 


- -600 


- -900 


Sound public finance suggests that government should finance its operating 
expenditures from current taxation and its capital expenditures from debt 
issues... long-term financing of capital expenditures ensures that the 
future work force shares in the costs ...(McKeough 1978b, 3.) 

This latter view was reiterated in the 1982 budget speech, in which the 
treasurer suggested that 'debt financing is a.. .preferable means of paying 
for capital investments...' (Miller 1982.) Both the 1978 and 1982 Budget 
Papers went into some detail to describe the composition of expenditures on 
fixed assets, pointing out (in 1978) that investment expenditure since 
1972-3 had exceeded borrowing. In the 1982 Budget Paper, it was stated 
that since 1972 '...the overall level of borrowing has been prudent and 
has virtually all been dedicated to Ontario's capital investment.' 
(Miller 1982, 8.) 


94 









TABLE 7 

Unemployment, the output gap, and the impact of discretionary changes on the 
budget, 1978-9 to 1981-2 


Year 

Unemployment 3 

rate 

(Ontario) 

Unemployment^ 3 

gap 

c 

Output 

gap 

Change 

Change in 
deficit due 
to discretion¬ 
ary policy 
($ millions) 

1978-9 

7.0 

1.7 

5.9 



1979-80 

6.6 

1.3 

4.5 

-1.4 

+ 195 

1980-1 

6.8 

1.5 

4.9 

+ .5 

- 46 

1981-2 

7.9 

2.6 

7.9 

+3.1 

+547 


a Statistics Canada The Labour Force . 

b Using 5.3 as the full employment norm. See Ontario (1977). 
c Estimate on the basis of the relationship between the unemployment gap 
and output gap, 1971 to 1977 as per note b. 
d See appendix 

There is no question that a strong case can be made for public fixed 
capital formation and that at least some portion of this ought to be fi¬ 
nanced by long-term debt (see Say, 1853, 481). It follows, then, that as 
long as these capital expenditures are included in the annual budget, 
expenditure budget planning should be geared toward a deficit as the norm 
and the concept of a balanced budget, at full employment , should be 
abandoned. Herein lies the dilemma of fiscal conservatism: advocacy of 
balanced public budgets and, at the same time, acceptance of the principle 
of debt-financing expenditure on capital assets. 

To a large extent, the problem stems from not separating explicitly 
the two types of expenditures (capital and current) and establishing a 
capital budget for the province. While the concept of a capital budget may 
seem straightforward, there are certain fundamental issues that must be 
decided: 

What constitutes public investment, expenditures on human capital, 
such as expenditures for education, or financial investments? 

Should gross or net investment be the basis for long-term debt finan¬ 
cing? 

How would the inflation premium on debt charges be incorporated in 
such a budget? 


95 














TABLE 8 

Public sector current and capital budget accounts 



Current budget 




Expenditures 


Revenues 

1 . 

Goods and services 
(exhaustible). 

1 . 

Taxation revenue. 

2. 

Transfers to persons and 
and businesses 

(excluding capital assistance). 

2. 

Fees, charges, etc. 

3. 

Interest on public debt. 

3. 

Transfers from private sector. 

4. 

Depreciation on existing 
assets. 




Capital budget 




Expenditures 


Revenues 

1 . 

Investment in capital 
assets. 

1 . 

Proceeds of debt issues 

for net investment expenditure 

2. 

Interest payments 

2. 

Transfers from current 
budget (depreciation and 
interest). 


Table 8 outlines the structure of a capital budget and illustrates how this 
budget might be related to the current or 'exhaustible' expenditure 
budget, assuming that only net investment is debt financed. A strict 
interpretation and application of this framework (requiring a balanced 
current budget) would apply only in an environment in which the economy 
was operating close to its potential. Deficits caused by the response of 
automatic stabilizers and as a consequence of discretionary policies should 
not be 'balanced off' against public capital expenditures. In this sense, 
the Ontario government's claim that its borrowings roughly approximate its 
capital spending and therefore constitute sound finance is not entirely 
correct. In a full-employment environment, its borrowings would have 
been considerably less than total investment, implying that a considerable 
portion of capital expenditure may have been financed from current tax 
revenue. 9 

This itself may be closer to 'sound financing' if we accept the notion 
that replacement investment ought to be financed from current revenue and 
only net additions to the capital stock should be debt financed. It has 


96 







been suggested by the Ontario Treasury that 1973-4 was a period of full 
employment. If we adjust for discretionary fiscal policy and the effect of 
the business cycle on the budget, a budget balance of approximately -$300 
million to -$400 million would appear to be the 'passive' outcome of the 
budget structure. 10 This is approximately 4 to 5 per cent of total expendi¬ 
ture. Assuming that a full-employment deficit of this relative magnitude 
has occurred over the 1972-82 period, and given our estimate of net invest¬ 
ment, Table 9 provides the pattern of net investment expenditure and 'full 
employment' deficits for the 1973-82 period. It must be emphasized again 
that it is possible that expenditure on capital assets might well have been 
less if the economy had operated at close to full employment throughout 
this period. Consequently, the net investment figure may be an overesti¬ 
mate. In addition, our estimate of the full-employment deficit may be less 
than what might have occurred given the significant change in the elasti¬ 
city of the personal income tax in the latter half of the 1970s and into the 
1980s. 11 What these rough data do suggest is that the budget structure 
does produce a deficit at full employment. Given that there has been net 
investment in this period, and given the present accounting practices of 
the government, such a deficit is quite compatible with this investment. 

IS ONTARIO'S TOTAL DEBT MANAGEABLE? 

While the extraordinary events of the past five years may go some distance 
to explain why deficits have increased and become so persistent, there 
remains the nagging question: Has the public debt, nourished by these 
deficits, become unmanageable? 

Statistics on government debt, like statistics on prices, productivity, 
and profits, are subject to broad interpretations (and misuse) unless one 
is precise about what is allegedly being measured. Since we shall be 
referring to a variety of concepts of public debt, it is important that the 
terminology be clarified at the outset. 

Public debt, in its broadest interpretatiopn, refers to a future finan¬ 
cial obligation on the part of the public sector. In the Province of 
Ontario, this obligation encompasses the debt incurred by the decisions of 
the legislature, by the crown corporations of the government, and by local 
governments, which are incorporated by the province. 

Most of Ontario's public debt is funded debt, meaning that the finan¬ 
cial obligation is backed by notes and debentures. Debt is unfunded if 


97 



TABLE 9 

Net investment and structural budget result ($ millions) 


Year 

Structural 
hypothetical 
budget deficit at 
full employment 

Net investment m 

a 

capital assets 

Total 

investment 

1973 

240 

654 

935 

1974 

287 

607 

920 

1975 

352 

638 

1,029 

1976 

380 

788 

1,272 

1977 

444 

619 

1,126 

1978 

472 

709 

1,313 

1979 

528 

672 

1,246 

1980 

608 

722 

1,416 

1981 

660 

665 

1,447 

1982 

752 

697 

1,621 


a Gross figures were adjusted by using the ratio of net provincial invest¬ 
ment to total provincial investment in Statistics Canada (1981). 


there is no specific term to maturity and no security instrument. For 
example, there are certain obligations to the Public Service Superannuation 
Fund but no provincial debt instrument backing these up. Funded debt is 
usually direct or guaranteed. Direct debt involves the issue of a bond, 
note, or security in the name of the Province of Ontario, while guaranteed 
debt is issued in the name of a crown corporation (such as Ontario 
Hydro). While some funded debt is not guaranteed (but considered part 
of the consolidated public sector debt), the distinction may be nominal. 
The unguaranteed public debt in the province is largely limited to the local 
public sector, and it is difficult to imagine a situation in which the pro¬ 
vince would permit a local government to default on its debentures. As of 
1982, approximately 85 per cent of total Ontario debt was funded; of that 
funded portion, 85 per cent was direct or guaranteed. 

What do the data related to Ontario's public debt tell us about its 
manageability? One's ability to manage debt depends on the interpretation 
of the word 'manage.' Income flow, the certainty of that flow, and market- 


98 





able assets are crucial factors in determining how a household can manage 
existing or additional debt. Given an uninterrupted flow of expected 
increasing income and/or highly liquid assets, manageability will be ranked 
high, with a credit rating involving a relatively low interest rate. In the 
corporate world, a similar analysis prevails, with cash flow and expected 
profits replacing wage income as criteria. 

Can the same criteria be used for governments? Obviously, parallel 
comparisons can be made, but they are not strong. The 'capacity to 
service debt' is often measured by ratios such as debt per capita, debt as 
a percentage of personal income, and debt as a percentage of gross pro¬ 
vincial product. They are not, strictly speaking, indicators of service 
capability unless nominal interest rates are constant over an extended 
period of time. A household mortgage of $50,000 over three or four years 
may, as a percentage of income, slightly decline, but if the interest rate 
on the mortgage goes from 11 to 20 per cent, debt to income is not a good 
measure of capacity to service debt in the short run . The standard 
measures noted above better describe capacity to redeem debt. 

Since this paper is primarily concerned with the budget deficit (or 
surplus), which, as a result of its year-to-year changes, changes the total 
outstanding direct funded debt, we turn first to an examination of direct 
funded debt. In Ontario, direct funded debt on a per capita constant 
dollar basis increased at an average rate of 6.5 per cent per year from 
1971 to 1978 and then began a slow decline. As a percentage of GPP, 
debt rose from 15.4 to 18.8 per cent between 1971 and 1978 and has 
declined to a little over 17 per cent. 

The increases in real per capita direct funded debt and direct debt 
as a percentage of GPP are, in our opinion, very modest in view of 

the need to maintain capital expenditure increases to ensure that the 
public capital stock does not rapidly deteriorate, 

the slowdown in economic activity, which has induced a decline in 
revenue growth, and 

the unprecedented increases in interest rates and the effect that 
these rates have had on the current expenditure side of the budget. 

As was suggested earlier, it would be unwise when speaking of debt 
management to ignore guaranteed debt and even unguaranteed debt that is 
placed under the umbrella of consolidated public debt. A small increase in 


99 



direct debt due to a budget deficit might be cause for concern if other 
forms of debt have been accelerating rapidly. In Table 10 we have includ¬ 
ed guaranteed debt in columns 4 and 5 and added the figures to direct 
debt giving totals in columns 6 and 7. The relative increase in guaranteed 
debt on a real per capita basis or share of GPP has been moderate, and 
the total funded debt figures show an increase up to 1978 and slight 
decline since then. 

Finally, Table 11 records and Figures 9 and 10 illustrate the growth 
of total public debt; that is, direct, guaranteed, and unguaranteed debt, 
the last encompassing local government debt. These data, like those in 
Table 9, reveal that total debt as a percentage of GPP or in real per 
capita terms grew until the late 1970s and has since declined. While it is 
not guaranteed in the formal sense, we have suggested that local govern¬ 
ment debt is, for all intents and purposes, guaranteed debt. The consoli¬ 
dated, aggregate data suggest there is very little reason to be concerned 
about Ontario's debt at this time. The precipitously high interest rates of 
1979-82 have created problems of debt management, but these problems 
should disappear with a return to more moderate interest rate levels. 

SUMMARY AND CONCLUSION 

The Ontario budget, like other public sector budgets, has been buffeted 
by both the high interest rates of the late 1970s and early 1980s, and the 
severity of the current recession. These events, cyclical in nature, have 
caused interest expenses and welfare payments to rise and induced slow 
revenue growth, producing deficits that in absolute terms are large by 
historical standards. Judged over a longer time horizon and in relation¬ 
ship to total public expenditure or gross provincial product, the deficits 
are not alarmingly high. 

If adjustments are made for the cyclical effects, both recession and 
high nominal interest rates, and if one takes into account the role of the 
province in providing for essential public investment, the budget structure 
is, at the present time, sound. The only caution we would raise at this 
time is with respect to the full-employment norm in the longer run. If the 
'working definition' of full employment over the remainder of the 1980s is 
going to be significantly beyond the 6.5 per cent unemployment figure, 
then it may be necessary to sustain a small structural deficit or realign 
expenditures and revenues. The choice will obviously depend on a variety 


100 


TABLE 10 


Direct 
basis) 

and guaranteed funded debt: Ontario 

1971-82 (Nat 

ional Accounts 





Guaranteed 



Year 

Direct funded debt 

funded 

debt 

Total 



Debt 


Debt 


Debt 


Real debt 3 

as % 

Real debt 

as % 

Real debt 

as % 


per capita 

of GPP 

per capita 

of GPP 

per capita 

of GPP 


$ 


$ 


$ 


1971 

793 

15.4 

345 

6.7 

1,139 

22.0 

1972 

900 

16.5 

350 

6.4 

1,249 

22.9 

1973 

928 

16.1 

357 

6.2 

1,285 

22.3 

1974 

930 

15.5 

347 

5.8 

1,278 

21.3 

1975 

1,028 

17.6 

364 

6.2 

1,392 

23.8 

1976 

1,038 

16.8 

438 

7.1 

1,476 

23.9 

1977 

1,087 

17.6 

445 

7.2 

1,531 

24.8 

1978 

1,153 

18.8 

448 

7.3 

1,600 

26.1 

1979 

1,142 

18.2 

430 

6.9 

1,572 

25.1 

1980 

1,109 

17.9 

431 

7.0 

1,540 

24.9 

1981 

1,127 

17.3(e) 

436 

7.1 

1,563 

24.4 

a Nominal debt 

deflated by Consumer Price 

Index; 1971 

= 100.0 



TABLE 11 
Total funded 

debt Ontario 1973-4 to 1981-2 

(includes local government) 


Funded per capita debt 
in constant dollars 

Funded debt as a 

Year 

($ millions) 

percentage of GPP 

1973-4 

1,710 

30.7 

1974-5 

1,695 

29.2 

1975-6 

1,855 

32.7 

1976-7 

1,896 

31.9 

1977-8 

1,933 

32.2 

1978-9 

1,975 

33.0 

1979-80 

1,918 

31.4 

1980-1 

1,851 

30.7 

1981-2 

1,807 

30.0 


a Fiscal year figures deflated by Consumer Price Index, 1971 = 100.0. 


SOURCE: Miller (1982b, 17). 


of circumstances, and, to ensure that any adjustment will minimize trans¬ 
ition costs, the question should receive serious attention in the near 
future. 


101 










Figure 9 

Debt management capability, Ontario 1973-4 to 1981-2 

Per cent 



102 




Figure 10 

Debt management capability, Ontario 1973-4 to 1981-2 

$ Millions 



103 






APPENDIX 

CALCULATION OF THE IMPACT ON THE BUDGET BALANCE OF 
DISCRETIONARY FISCAL POLICY, 1978-9 TO 1981-2 

Expenditures 


The change in expenditure due to discretionary policy was estimated as 
AG d = + [G® - (G®_j)(l+6)] 

where 

AG U is the change in expenditure due to discretionary policy; 

G^ is government expenditure on goods and services; 

G is government expenditure on transfer payments; 

6 is the coefficient associated with the 'automatic' growth in 
various transfers. For example, in the case of income support 
transfers, the factor was the the growth rate from one year to 
the next in the number of citizens over 65 years of age. 

Taxes 


The Treasury provides an estimate of the effect of all important tax 
changes on revenue in the coming year. It also estimates the total change 
in tax revenue for each revenue source. Ex post , there is only the actual 
change in total tax revenue. 

We have therefore computed an adjusted effect of discretionary tax 
* d 

changes, AT , in the following manner: 

AT d = (AT d )(AT aCt /AT) 
where 

is the estimated revenue effect of the tax change; 
is the actual change in tax revenue; 
is the estimated total change in tax revenue. 


AT 


AT 

act 


AT 


104 





NOTES 


1 See Organization for Economic Co-operation and Development (1980), 
volume 2, and Council of Economic Advisers (1981). 

2 Due to changes in reporting procedures and accounting practices, data 
from the 1950s and early 1960s are not strictly comparable with recent 
expenditure figures. In earlier years, certain data on expenditures 
were reported as 'net' figures (after subtracting revenues directly 
applicable to the expenditure function). 

3 This is an underestimate, since it assumes that nominal public debt 
would be the same as it actually was. If inflation had averaged 3-4 
per cent, the need to raise nominal debt to achieve some target real 
debt would have been reduced and public debt would have been less than 
actual. 

4 As an aside, the inclusion of total interest in the current budget is 

also misleading in terms of compensatory fiscal policy and the impact 
of the deficit on the economy in a period of stagflation. Again, let 
us consider a simple model for the purpose of exposition. If the 
government sells $100 million worth of bonds in order to expand cur¬ 
rent spending and stimulate the economy, it must (if the inflation 
rate is 1 \ to 8 per cent) use $10 million of the proceeds for interest 
payments. Bondholders are not likely to treat the $10 million as an 
addition to current income: they will need to add $7.5 to $8.0 

million to their investments to ensure the constant real value of 
their bonds. 

5 An alternative adjustment is to estimate the real rate of return (r) 
each year by calculating (r-p), where r is the nominal interest rate 
and _p is the rate of inflation, and multiplying net outstanding debt 
by r to obtain real interest payments. This is labelled 'real 
interest rate adjustment' in Figure 6. 

6 The FEBS technique removes, in part, the effect that changes in eco¬ 
nomy activity have on the budget result. This permits the analyst to 
make a relative measure of the impact of discretionary policies on the 
economy. In a very simple model, the budget balance (B) is the dif¬ 
ference between revenues and expenditures or 

B = T - G 

where T = T + t.Y 

G = G - g.Y 

where T and G are autonomous taxes and expenditures, t is the tax 
rate, g is the rate of transfer payments associated with the level of 
economic activity, and Y is GNP. 

If the government does nothing, the budget balance will rise 
(greater deficit) as Y falls and vice versa . The FEBS measures 
changes in B when Y is fixed at potential GNP. 

7 This is not entirely true in a growing economy because the full 
employment level of GNP would rise from period to period, inducing 
higher and higher revenues. 


105 













8 'Unweighted' refers to the absence of a coefficient on any tax and 
expenditure variables that would represent the 'first round' impact of 
change in that variable on aggregate demand. 

9 This would not be the case if a significant share of capital spending 
during periods of economic slowdown had been legislated as a short¬ 
term measure to combat unemployment. 

10 These estimates accompany McKeough (1977). 

11 In the 1979 Ontario budget, it was reported that the elasticity of the 
provincial personal income tax with respect to GPP had declined from 
1.32 in the early 1970s to 0.4 in the 1977-9 period. While the de¬ 
cline was clearly due to relatively slower growth in personal income, 
the introduction of indexation in 1974 introduced a structural change 
to the tax system that reduced elasticity (see Miller 1979, C—7). 


REFERENCES 


Jones, B., N. Bardeckl, and B. Hull (1977) Regional Stabilization in 
Canada: the Ontario Record (Toronto: Ministry of Treasury and 
Economics) 

McKeough, D., (1971) Ontario Budget , (Toronto: Ministry of Treasury 
and Economics) 

- (1977) Ontario Budget (Toronto: Ministry of Treasury and Economics) 

- (1978a) Ontario Budget (Toronto: Ministry of Treasury and Economics) 

- (1978b) Ontario's Borrowing and Public Capital Creation Budget 

Paper C (Toronto: Ministry of Treasury and Economics) 

McNaughton, Charles (1969) Ontario Budget (Toronto: Ministry of 
Treasury and Economics) 

Miller, F. (1979) Ontario Budget (Toronto: Ministry of Treasury and 
Economics) 

- (1982a) Ontario Budget (Toronto: Ministry of Treasury and Economics) 

- (1982b) Public Investment and Responsible Financial Management , 

Budget Paper C (Toronto: Ministry of Treasury and Economics) 
Ontario, Ministry of Treasury and Economics (1977) Reassessing the Scope 
for Fiscal Policy in Canada , (Toronto) 

Organization for Economic Co-operation and Development (1980) National 
Accounts of OECD Countries 

Say, J.B. (1853) A Treatise on Political Economy , D.R. Princep, trans. 

(Philadelphia: Lippincott, Grambo and Co.) 

Statistics Canada (various issues) The Labour Force 71-001 monthly 
(Ottawa: Statistics Canada) 


106 


















- (1981) Fixed Capital Flows and Stocks 13-211 annual (Ottawa: 

Statistics Canada) 

United States (1981) Economic Report of the President (Washington: U.S. 
Government Printing Office) 

White, J. (1973) Ontario Budget , (Toronto: Ministry of Treasury and 
Economics) 


W. R. White* 

Professor Auld's paper is an interesting and useful one, particularly for 
anyone who, like myself, knows little about the fiscal position of the 
provinces. Yet it must be said that it is also a very 'traditional' paper in 
terms of the methodology it employs. Is the deficit too high? A large 
number of economists, including Professor Auld, contend that the answer 
can be found by taking the actual deficit number and adjusting for (a) 
cyclical effects on revenues and expenditures, (b) inflation effects on the 
cost of debt service, (c) capital expenditures, etc. If the resulting 
number is positive, the conclusion frequently drawn is that there is no 
'structural' problem, and for some it seems but a small step to go even 
further and conclude that in such circumstances 'discretionary' fiscal 
stimulus should be applied. 

This whole approach to analyzing deficits makes me uneasy. It seems 
to me to give enormous importance to what is, after all, only a single 
number describing the state of a very complex economy. Furthermore, the 
adjustments carried out do not always seem to me to be based on any 
explicit statement of what it is that is worrisome about big deficits. If we 
are after better measures of the supposed inimical effects of government 
deficits, then surely we must begin by specifying exactly what these 
effects are. Indeed, if there are numerous sources of concern (say, 
inflation, crowding-out, and debt-servicing capacity), then the specific 
form of the adjustments must depend on the source of concern identified. 
Professor Auld does indeed mention some reasons why people are concerned 
about deficits, but such considerations should be at the very heart of his 
methodology - not just mentioned in passing. 

At the risk of some oversimplification, let me now distinguish two 

* Chief, Research Department, Bank of Canada. 


107 











kinds of concerns about deficits, try to describe why economists make the 
adjustments that they do make in each case, and finally point out some of 
the difficulties inherent in making these adjustments. I will refer to 
Professor Auld's paper throughout, since his concerns about deficits seem 
to fall into both camps. 

DEFICITS: CONCERNS ABOUT INFLATION AND CROWDING-OUT 

Most people's worries about deficits seem to arise from a simple proposition. 
If the economy were at full employment and if the government increased its 
deficit and if the resulting private income were not wholly saved, then 
unpleasant things would happen. If the deficit were monetized, then 
inflation would increase. If the deficit were not monetized, then real 
interest rates would tend to increase and public expenditures would tend 
to crowd out private expenditures for either fixed capital formation or (in 
small, open economies) the consumption of domestically produced tradeable 
goods. 

The various adjustments made to the deficit numbers by Professor 
Auld and others constitute an attempt to modify reality to see what the 
potential for inflation and/or crowding-out would be if any or all of these 
hypothetical conditions really existed. The number of ifs that I have just 
enumerated gives some idea of how robust I believe this methodology to 
be; the answer is, not very. Having said this, I must add that even so I 
prefer adjusted deficit numbers to raw deficit numbers, which are also 
subject to misinterpretation. Nevertheless, concerns can be raised about 
each of the adjustments Professor Auld suggests. 

Cyclical adjustments to deficits 

Professor Auld uses cyclical adjustments of government revenues and 
expenditures to ascertain what the deficit would look like in a world of full 
employment. This is now standard procedure and well worth doing, but it 
bears repeating that it is an exercise fraught with hazard. 

First, there is the issue of measuring potential . Most of these 
measures tend to be little more than extrapolations of recent trends. 
Recent experience shows that such ex post measures are a risky basis for 
policy prescriptions. Indeed, at the national level, many economists have 
been revising their estimates of potential downwards almost continuously 


108 





since 1976. Thus, estimates of full-employment surpluses made in 1976 
would in retrospect now have to be considered full-employment deficits. 
Second, this approach really assumes an instantaneous return to full 
employment, with the result that the issue of debt accumulation during the 
recession can safely be ignored. Recent experience and future prospects 
imply that this approach may be overly optimistic. Ontario could have a 
full-employment surplus today, but if its actual deficit is expected to stay 
very large for a long time, then the associated implications for debt 
service should to some degree condition current policies. Moreover, the 
size of this problem will be aggravated to the extent that there is a lot of 
debt to begin with, and to the extent that the real rate of interest 
exceeds the real rate of growth of the economy - whatever its level. This 
complication's potential for creating longer-run difficulties should not be 
underestimated, much less ignored, as is the case when 'cyclical' adjust¬ 
ments are made to deficits to reveal the 'underlying' or FES position. In 
my judgment, the only practical way of dealing with this particular 
(dynamic) problem is to generate a sensible medium-term projection for the 
Ontario economy that incorporates tax and revenue functions and alterna¬ 
tive assumptions about real growth and interest rates. Then one can look 
at the longer-run evolution of the financial variables of interest (in partic¬ 
ular, deficits, debt/GNE, etc.) and see whether the trends do seem worri¬ 
some for those concerned about inflation and/or crowding-out. 

The inflation adjustment to the deficit 

Those concerned with monetization of deficits or crowding-out do recognize 
that the capacity of a deficit to increase total spending may be offset by 
increased private saving. This is where the inflation adjustment comes in. 
To the extent that inflation erodes the real worth of private assets (public 
liabilities), the private sector should save more (in principle out of 
inflated interest income) in order to offset this effect. 

In principle, this adjustment should also be made. Yet the practical 
problem is how to measure the increased saving caused by higher inflation. 
This is not an easy task, as those who have attempted it (in particular, 
Greg Jump) will attest. 1 Professor Auld's calculations use ex post 
inflation times the stock of existing debt, but this does presume that the 
demand for real debt is independent of the inflation rate and that all 
inflationary expectations are realized. These are very strong assumptions. 


109 








By way of a summary example, suppose that inflation had been unexpected 
and that interest rates increased only after the event. All existing 
bond-holders would be nursing memories of big losses. Is it obvious that 
they would now be inclined to buy still more government bonds because 
the real value of their government bond portfolios had fallen? While this 
must, of course, happen over the longer run, the transition path to that 
long run could be characterized by quite different savings behaviour. 

The capital stock adjustment to the deficit 


Those concerned about deficits leading to crowding-out of private invest¬ 
ment might well be tempted to adjust the deficit to the extent that it is 
financing public capital formation. This adjustment makes sense at least to 
the extent that private capital formation and public capital formation carry 
similar rates of return. 2 Again there are difficult measurement questions 
to answer (in particular, with respect to the treatment of human capital), 
but Professor Auld has mentioned most of them. 

Conclusion 


All of Professor Auld's deficit adjustments seem appropriate in principle if 
one is worried about the prospects of public sector deficits crowding out 
private spending. In practice, there are substantial difficulties involved 
in making such adjustments, and they should be treated with caution. 

I also wonder if concern about crowding-out provides the best 
starting place for the analysis of a provincial government deficit taken by 
itself. Does the Ontario government have significant capacity to raise 
levels of economic activity in the province (through an increased deficit), 
and hence levels of interest rates and the level of the exchange rate? The 
first link would seem somewhat problematical, given the degree of import 
leakage in an economy as open as that of Ontario's. As for raising 
interest rates generally, it seems more likely that Ontario government 
deficits would raise Ontario government bond rates, relative to others, 
than that there would be a general increase in the level of rates. While 
associated capital inflows (provincial bond issues abroad) might in fact 
raise the exchange rate and thus crowd out expenditures on domestically 
produced tradeable goods, I believe this would be a temporary rather than 
a permanent phenomenon. 3 The conclusion to be drawn from all of this is 


110 





that, adjusted or unadjusted, the size of the Ontario deficit should not be 
a source of great concern for those who are principally worried about 
crowding-out. Whether the same conclusion could be drawn were we to 
consider the deficit position of, not one, but all ten provinces taken 
together is, of course, a different issue. 

DEFICITS: CONCERNS ABOUT BOND RATINGS AND MANAGEABILITY 

Yet there are other reasons for being concerned about deficits, and it is 
with these reasons that an analysis of the Ontario budget position would, I 
think, more properly start. Deficits are important not because of the 
flows themselves, but because they contribute to a large stock of debt that 
may or may not be serviceable. From this perspective, adjustments to 
actual deficit figures are still warranted, but they may in some cases have 
to be made rather differently. What we are concerned with now are the 
costs of servicing the real net liabilities of the government sector, relative 
to the capacity of the government to tax. 

Inflation-adjustment 


From this perspective, the deficit should be adjusted for the effects of 
inflation (but in this case using ex post inflation figures); the real 
liabilities of the government have fallen to the extent the price level has 
risen. This phenomenon can, however, be controlled for in a simpler way 
by adding actual deficit numbers to the stock of existing debt and then 
looking at the ratio of debt to GPP (nominal) over time. Since both 
numerator and denominator will be growing with inflation, this ratio should 
be unaffected by inflation. This approach also focuses attention directly 
on the stock of debt relative to some proxy for servicing capacity, which 
by assumption is what we are really concerned about in this case. 

Capital expenditures 


Here too an adjustment to the deficit is warranted, the point being that 
such capital expenditures are presumed to have a rate of return over time 
that would warrant their being undertaken. That is, they are in some 
sense self-servicing and do not increase the government's net debt service 
burden. 


Ill 








But again, if we are primarily concerned with the stock of net liabili¬ 
ties (governments' net worth), it might be better to just measure them 
directly rather than to impute something about stocks by adjusting flows. 
However, to put the matter this way immediately raises the question of 
why all the other assets and liabilities of governments are not also 
included in this balance sheet exercise. A good example of the assets 
would be the future taxes to be levied on RRSPs, and a good example of 
the liabilities would be the various forms of unfunded debt noted by 
Professor Auld in the last part of his paper. 

A proxy for servicing capacity 


Finally, if the capacity to service debt is what is at issue, then we need 
to address the tax base question directly. It is not obvious how the 
deficit could be adjusted to take this into account. Professor Auld relies 
on such stock measures as debt per capita and debt/GPP to get some 
handle on this issue. While this does take us a long way, the problem is 

clearly more complex than this. The initial level of taxes relative to GPP 

must also be taken into account in identifying potential servicing problems. 
Tax rates that are relatively low (compared, say, to the tax rates of other 

provinces or states) can be raised, if need be, whereas increases in tax 

rates already at high levels threaten to be counterproductive if capital and 
labour respond by moving elsewhere. I look forward to the rest of today's 
proceedings in the hope of hearing other suggestions about judging the 
'creditworthiness' of Canadian provinces. 

SOME MISCELLANEOUS ISSUES 


Sign of fiscal multipliers 


There are two statements from Professor Auld's paper that attracted my 
attention: 

Attempts to maintain the $50 million surplus by higher taxes or lower 
spending in the face of a recession would only deepen the recession and 
possibly contribute to a higher deficit in the short run. 

For example, an actual deficit of $500 million would have been larger if a 
tax reduction had not been implemented in that same year. 


112 





Both of these statements imply that tax increases increase deficits, presum¬ 
ably because they result in lower incomes and, consequently, lower tax 
receipts. I would like to put on record my belief that while such feedback 
effects do reduce the size of fiscal multipliers, they do not reverse their 
sign. One can, of course, conceive of perverse signs arising if deficits 
cause financial markets to push real interest rates up sufficiently, but 
(whether plausible or implausible) that is quite a different thing. 

'Efficient' discretionary fiscal policy 


Professor Auld suggests that 'efficient' discretionary fiscal policy always 
leans against the wind of cyclical movements. I would add one qualific¬ 
ation to this, assuming one lives in a world in which automatic stabilizers 
are already designed to reduce the amplitude of cyclical movements. 
Logically, Professor Auld's suggestion seems in fact to imply that Ontario's 
automatic stabilizers have been inefficiently designed and should be made 
stronger. I wonder whether he would agree with this. 

Balanced budgets and capital transactions 


This is a small point but an important one. Professor Auld seems to imply 
that over the cycle the budget (perhaps after accounting for capital 
expenditures) should be balanced. However, in a growing economy there 
is a presumption that the demand for real financial assets, including the 
liabilities of governments, would grow at the same rate. A cyclically 
balanced budget would then, over time, result in a relative shortage of 
government securities in which to invest. 


NOTES 


1 The fact that personal savings behaviour has differed so widely in 
Canada and the U.S. since the middle 1970s, in the face of a broadly 
similar inflationary experience, also raises a warning. The implic¬ 
ation is that we should be hesitant in making inflation adjustments to 
savings rates on the basis of simple propositions about the absence of 
money illusion. 

2 This is fairly heroic in itself. Consider a range of public invest¬ 
ments from bridge maintenance to Mirabel and the associated rates of 
return on each. 

3 In my view, substitution between various assets should suffice over 
time to wash out the effects of increased supplies of any particular 
form. 


113 










Discussion 


QUESTION: I'd like to ask Professor Auld if he thinks there’s a case for 
legislating a requirement that some official agency of the provincial govern¬ 
ment be asked to carry out these adjustments in the best manner possible 
on a regular basis and publish them for public scrutiny. 

D.A.L. AULD: The provincial government made some of the adjustments 
up to the period 1977-8. We have had discussions with people in Treasury 
with regard to the preparation of this paper, and I think the concern is 
the one that Bill White mentioned. That is, how can we get a handle on 
what is the appropriate full employment, full capacity operation of the 
Ontario economy right now? I don't think there is very widespread agree¬ 
ment as to what benchmark one would use for that particular adjustment. 
On the capital side, I guess I welcome the thrust in last year's budget 
speech, which made a strong case for public fixed capital expenditures. 
But I think there are still a number of issues to be resolved there. For 
example, do you include net investment in human capital as part of public 
capital formation? How do you get a comparable rate of return on capital? 
These are difficult issues, but I still think it would be useful in terms of 
accountability if provincial governments and the federal government too, 
were to establish separate capital accounts. This is done in a number of 
other countries. The one I am most familiar with is Australia, where for 
years they have had a separate capital account and, as well, a relationship 
between the current budget and the capital budget, which gives a much 
better idea of the role of the public sector in the formation of net capital. 

I guess my answer would be yes, I think these adjustments should be 
made and these changes should be made. If there is difficulty in making 
them, then wave the yellow flag while you are doing it. But I do think if 
there were at least some measure, we would have a better accountability of 
where taxes are going and the role of deficit financing. 

QUESTION: I have a question for Professor Auld. It wasn't clear to me 
whether your amount for funded debt included the amount owed to the 
Canada Pension Plan. 

D.A.L. AULD: The direct funded debt, which is derived from the Ontario 


114 



government’s budget deficit, is funded in part by borrowing from the 
Canada Pension Plan and in part by borrowing from other government 
pension and superannuation funds. The arrangements for the borrowing, 
with respect to the term of the debt and the calculation of the interest 
rates, are set out in legislation. As you probably know if you have read 
the budget speeches for the last few years, the Ontario government has 
done no borrowing from the private capital markets for the purpose of 
financing its deficit. Of course, you can simply switch the thing around 
and say, 'if Ontario is borrowing from the Canada Pension Plan and 
superannuation funds, then those funds are not available to the private 
sector'. 


115 


Deficits and the economy to 1990: 
projections and alternatives 

D. P. Dungan* and T. A. Wilson** 


Much of today’s concern with the federal deficit is not so much concern 
with what the deficit is right now - large though it is - but a concern 
about what the present figures mean for the future. 

Will the current high deficit be even higher? Will it continue to 
balloon? Will it become runaway? Will it crowd out investment, stifle 
future recovery and growth, and possibly lead to a new burst of inflation? 
Or is it a transitory phenomenon that will disappear with recovery and 
subsequent economic growth? 

Of course, the answers to these questions have great importance for 
the conduct of current fiscal policy. Is there room to manoeuvre or not? 
If so, for how long, and for what types of manoeuvres? 

In an attempt to shed some light on these questions, we have devel¬ 
oped a trend projection of the Canadian economy from current conditions 
through 1990. 1 As part of the overall projection, we obtain projections for 
the federal and consolidated government sector deficits that are consistent 
with the overall outlook. 

It is important to note that this basic trend solution assumes no new 
policy initiatives. It does, however, make use of several key policy and 
external assumptions, which are described briefly below. In later sections 
of the paper, we consider some possibilities for new fiscal initiatives. 2 
These policy 'experiments' indicate not only the effects of these initiatives 
on economic activity, but also their effects on inflation and on the deficit; 


* Assistant Professor, Department of Economics, and Associate Director, 
Policy and Economics Analysis Program, University of Toronto. 

** Professor and Chairman, Department of Economics, University of Toronto, 
and Research Associate, Institute for Policy Analysis. 


116 


hence they help to indicate how much room for fiscal manoeuvre may now 
exist. 

Our method puts heavy weight upon econometrically estimated relation¬ 
ships within a consistent accounting framework. It is our belief that it is 
most important to guantify and use numbers to describe the relevant 
factors in a consistent fashion. 

THE BASE-CASE PROJECTION 

The base projection derives from two key elements. The first is the 
FOCUS model itself. 3 FOCUS is a medium-scale quarterly econometric 
model of the Canadian economy. It is by no means a simplistic Keynesian 
Model. In fact, it tends to be rather less Keynesian than most Canadian 
econometric models 4 , in that it has strong neoclassical properties in longer- 
term analysis. It represents an accumulation of human capital, in the 
sense that what is placed on the computer is a codification of our learning 
about the economy over the years. 

The second key element is the assumptions that we feed into the 
model for the current and future years of the projection. 

The first set of assumptions, critical for this kind of exercise, is 
about what is going to happen in the rest of the world. Here we have 
drawn heavily upon a long-term trend projection for the United States that 
was released by Data Resources Inc. in December 1982. 5 In this pro¬ 
jection, real growth in the U.S. for the 1983 to 1990 period averages about 
3.5 per cent, with price inflation over the same period at about 6 per 
cent. 

Given recent world oil price developments, these figures may repre¬ 
sent an unduly pessimistic outlook for inflation in the United States. In 
an alternative experiment, reported in Appendix B, we examine the sensi¬ 
tivity of the base projection to a lower inflation rate. 

With respect to world oil prices, we assume a 12 per cent decline this 
year, flat prices in nominal terms in 1984 and 1985, and a resumption of 
gradually rising relative oil prices over the remainder of the projection 
interval. 6 

As for domestic energy prices, we assume that there will be a federal- 
provincial compromise on the oil-price guestion, with domestic oil prices 
rising to 90 per cent of world prices from the current 75 per cent, phased 
in over a three-year period in 5 per cent steps. Under these assumptions 


117 


S’ 


the domestic oil price in Canada never has to decline. 7 

Of the oil mega-projects, only Hibernia is incorporated in the base 
case, since it is assumed to be viable even with the projected lower prices. 
We also include a number of smaller-scale oil projects coming on stream, as 
well as pilot projects and experimental projects, which help to keep up the 
level of energy sector investment. 

We assume that the Bank of Canada will conduct monetary policy to 
accommodate about 6 per cent inflation with about 4 per cent real growth. 
Note that this result does not imply a steady 10 per cent rise in the money 
supply every year. While the increased demand for money that will accom¬ 
pany disinflation must be accommodated by additional growth of the money 
supply over the next two years, we assume that continued innovations in 
the financial markets will allow the income velocity of Ml to increase there¬ 
after at about 2.5 per cent per year. 

As noted above, we assume no new fiscal initiatives. We have incor¬ 
porated the six-and-five restraint program, subsequent income tax index¬ 
ing, the UIC contribution rate increases that went into effect on 1 January 
of this year, and, subsequently, a return to the system of annual adjust¬ 
ments to the UI contribution rate. 

However, we have had to make some assumptions about future govern¬ 
ment expenditures. These assumptions are depicted in Figures 1 and 2. 
Figure 1 projects government expenditures on real goods and services as a 
percentage of real GNP. As is apparent, this ratio declined over time 
until 1981. During the recent severe recession, real GNP went down, so 
that the ratio took a sudden upturn in 1982. We assume that it will decline 
again as recovery takes place. Thus, over the projection interval, we 
assume a gentle downward trend for government expenditures in relation to 
aggregate output. The important thing to note here is that we are not 
assuming anything radically different from the recent past. For several 
years now, the share of real government spending in real GNP has been 
gradually declining. 

When we examine total government expenditures, the story is some¬ 
what different. These expenditures include all transfer payments and 
interest on the public debt as well as purchases of goods and services. 
In relation to nominal GNP, total expenditures experienced a more dramatic 
upsurge in 1982 than expenditures for goods and services alone: not only 
did GNP go down, but interest payments and some transfer payments 
(notably UIC payments) went up. 


118 


Figure 1 

Real government current and capital spending as percentages of gnp 

Per cent 



Figure 2 

Government total expenditure as a percentage of gnp 


Per cent 



The recent large increases in interest payments on the public debt 
are in part a reflection of the impact of inflation upon interest rates. As 
noted below, when deficits are adjusted to remove this inflation-induced 
accounting distortion, the picture changes dramatically. Given a reduction 
of interest rates and gradual recovery in the projection, total expenditures 
also begin to decline as a percentage of GNP. 

So much for key assumptions. Let us now briefly review the results. 


119 















Figure 3 

Real gnp growth rate 


Per cent 



Figure 3 plots real GNP growth over the projection interval. As is appar¬ 
ent, there is a recovery, albeit a very mild one, this year, with real 
growth averaging just 1.3 per cent. However, 1984 witnesses a more 

vigorous recovery, with real growth of 4.1 per cent. Over the remainder 
of the projection interval, real growth averages 3.9 per cent. These 

growth rates are somewhat above Canada's estimated potential growth rate, 
but it is important to recognize that, after a deep recession such as that 
of 1982, several years of above-potential growth are required to bring 
unemployment down to its equilibrium or natural level. 

The depth of the recent recession is also apparent in the projections 
regarding capital formation. Under the base case scenario, real investment 
growth looks quite vigorous for every year after 1983; nevertheless, we 
are not back to 1981 investment levels until 1987 or 1988. We have gone 
down so far that it takes a long time to come back, and indeed for the 

near term this recovery is led by consumption and housing rather than by 

business fixed investment. 

Figure 4 graphs the projections of the key policy objectives of infla¬ 
tion and unemployment. The spectacular rise in unemployment in 1982-3 
stands out in the figure. Unemployment is projected to peak this year at 
just above 13 per cent, and to decline only gradually as recovery takes 
hold. (The most current data indicate that this rate might be slightly too 
high.) At the projection horizon in 1990, the unemployment rate is 8.6 


120 





Figure 4 

Inflation and unemployment 


Per cent 



per cent, about 2-3 percentage points above its estimated equilibrium level 
for that year. The important story here is that there remains considerable 
excess supply in labour markets over the medium term. 

The inflation rate has been coming down, and we project it to decline 
to a rate somewhat below 6 per cent in 1984 when a slight uptick returns 
it to about the 6 per cent level. 8 As noted above, this inflation rate may 
perhaps be too pessimistic in the light of recent price developments. 

Figure 4 also depicts our long-term estimate of the 'full employment' 
or natural unemployment rate. As is apparent, this rate gradually dimin¬ 
ishes over the decade. The primary reason for this result is demographic: 
there will not be as many people in the labour force in the high unemploy¬ 
ment groups as there were before. Note that the unemployment rate in 
the base case scenario is above the estimated full employment unemployment 
rate at all times, indicating continued slack in the labour markets. This 
trend suggests that the inflation rate may come down somewhat more than 
we have projected - in the absence, of course, of any major external price 
shocks. 

What are the implications of this scenario for the deficit? As Figure 5 
shows, the projected federal deficit increases to about $28 billion on a 
National Income Accounts basis this year, and then gradually declines to 
about $18 billion at the projection horizon in 1990. However, the absolute 
deficit gives a somewhat misleading picture. A portion of this deficit 
simply represents the inflation component of interest payments on the 


121 









Figure 5 

Federal deficit (nia basis) 


$ Billions 



public debt. In addition, aggregate nominal GNP more than doubles over 
the projection interval. 

Figure 5 also presents the inflation-adjusted federal deficit, which 
changes the impression of the problem. Although the adjusted federal 
budget does not return to the surpluses of a few years ago, the deficit 
does approach zero by 1990 under sustained moderate growth. Since the 
unemployment rate remains above its natural level at the projection hori¬ 
zon, the inflation-adjusted full-employment budget would be in a small 
surplus position. 

There is another, more obvious adjustment to make: that is, to 
express both the unadjusted deficit and the adjusted deficit as a per¬ 
centage of GNP, as shown in Figure 6. Obviously, the adjusted deficit 
still approaches zero in 1990, but the important thing is that the unad¬ 
justed number as a percentage of GNP does come down guite a bit over the 
period. It is still high - it still is obviously a positive deficit - but it 
comes down considerably more as a percentage of GNP. 

Figure 7 presents the debt/GNP ratio, which is an approximate mea¬ 
sure of the real debt burden. 9 As is apparent, this number fell from 
about 20 per cent in 1970 to about 10 per cent or less in 1974-5. Since 
then it has been rising. In our projections, it continues to rise through 
the mid-eighties as we continue to have very large deficits. However, it 
does reach a peak near 50 per cent in 1988. 

There is no reason to believe that a debt/GNP ratio this high will 


122 







Figure 6 

Federal deficits as a percentage of gnp 


Per cent 



Figure 7 

Federal debt as a percentage of gnp (estimate) 


Per cent 



necessarily stifle our growth. We have had ratios higher than this before; 
indeed, they were higher for much of the entire postwar period - boom 
years for investment in Canada. The projected debt/GNP ratios are, 
therefore, in keeping with historical experience. 

Figure 8 compares the aggregate deficit with the federal deficit, 
revealing the net surplus of the combined non-federal sector. This sur- 


123 











plus drops near zero in the current recession but then increases again, 
partly on the strength of resource revenues. 

There are some indications that, beyond 1990, and perhaps even in 
the late 1980s, the Canada Pension Plan contribution rate should be in¬ 
creased to make sure that the plan retains a comfortable surplus. Since 
we did not carry the simulation beyond 1990, we did not incorporate such 
an increase in the CPP contribution rate. 

Figure 9 compares corporate and government financing, presenting a 
rough estimate of corporate finance requirements against the aggregate 
government deficit, both expressed as percentages of GNP. As is readily 
seen, the increase in this year's and next year's deficit is offset by a 
large decline in corporate financing requirements. Corporations have been 
burned badly by the recession and are carefully rebuilding balance sheets 
and restricting new investment commitments, at the same time that reces¬ 
sion-induced reductions in tax revenues and increases in transfers swell 
the deficit. By the mid-eighties, corporate financial requirements are 
projected to return to the levels that we saw before the extraordinary 
run-up of the late 1970s and 1980s. When this occurs, it is important that 
the federal deficit be coming down. Indeed, it is necessary for the smooth 
growth indicated in this projection that the two lines cross in about 1985, 
with the federal government's borrowing requirements declining (in relation 
to GNP) to make room for the corporations as they begin to invest more 
and hence need to borrow more. 

POLICY IMPLICATIONS OF THE BASE CASE 

The base-case projection represents a scenario in which continued down¬ 
ward demand pressure on the price level is generated by persistent excess 
supply. 

The existence of considerable excess supply in the near term, as 
evidenced by both unemployment and capacity utilization rates, indicates 
that there is room for near-term fiscal stimulus with little danger of signi¬ 
ficant short-term crowding-out through financial market or price level 
effects. 

This does not mean that the medium-term consequences of increased 
federal deficits are no cause for concern. Efficient capital markets should 
take into account the longer-term consequences of federal fiscal actions. 
It is therefore important to design policy initiatives that have minimal 


124 



Figure 8 

Aggregate and federal deficits as percentages of gnp 


Per cent 



Figure 9 

Government and corporate financing requirements as percentages of gnp 


Per cent 



‘Estimate: non-residential investment minus CCA’s and retained earnings 

effects on the structural deficit over the medium term. 

Fortunately, there exists a variety of policy initiatives that would 
provide near-term stimulus without significant effects on the medium-term 
structural deficit. Expenditure programs that are explicitly temporary and 
temporary tax cuts to stimulate aggregate demand are two obvious exam- 


125 









pies. The problem in both cases is to make the temporary design credible 
to the financial markets. 

There are also more permanent measures that would provide additional 
stimulus with little or no increase in the medium-term structural deficit. 
Measures to improve the automatic stabilizers - such as a reform of the 
Unemployment Insurance system to eliminate the perverse tax increases 
generated by the current statutory formula - represent one such set of 
policies. Another set is represented by tax measures designed to augment 
aggregate supply as well as aggregate demand over the longer run. 

In the alternative projections, we consider examples of each of these 
three types of policy initiatives. In two of the experiments, we examine 
variations based on assumptions regarding price behaviour, the exchange 
rate, or monetary policy. 

We first consider what might be called a traditional temporary demand 
shock via increased government expenditures - a 'standard' Keynesian 
policy. An attractive feature of this policy is that it can give fast pain 
relief; but not long-lasting relief - it is palliative only. 

The second type of policy we examine is something more permanent - a 
policy that is directed towards the supply side in such a way that, by 
increasing potential supply, it could reduce its own inflationary impact and 
also mute its future deficit impacts. 

The third type of policy we consider is one example of how an exis¬ 
ting automatic stabilizer could be improved - giving a boost now but swing¬ 
ing the other way when a period of recovery with higher employment is 
well established. 

POLICY EXPERIMENTS INVOLVING TEMPORARY EXPENDITURE INCREASES 

The first pair of experiments basically involves an expenditure increase of 
$2 billion for 1983 and for 1984, partly in current expenditures and partly 
in capital expenditures. This is an example of a type of short-term job- 
creation policy. 10 

In the first experiment, the expenditure increase is bond financed. 
That is, we assume that the Bank of Canada holds to the same monetary 
course it follows in the base case - it does not let the additional govern¬ 
ment expenditure in any way alter the monetary growth rate. With no 
monetary response, some increase in interest rates is likely, and possibly 
some crowding-out of investment will occur. 


126 


Figure 10 depicts the economic effects of this bond-financed expend¬ 
iture increase. As is apparent, under this policy GNP would be a half 
percent higher in 1983 than it would have been otherwise. In absolute 
terms, the policy yields approximately an additional $3 billion worth of 
aggregate output in 1984. The CPI would be a half per cent higher than 
it otherwise would have been in 1984. 

Because this policy is a temporary one, it raises real GNP initially; 
however, once the policy stimulus is removed, real GNP actually dips below 
its base case levels for a short period, then gradually returns toward the 
levels of the base case. On the other hand, there is a permanent increase 
in the price level but obviously not in the inflation rate . 

Figure 11 shows the impacts on the unemployment rate and rates of 
growth of output and prices. Note that the three variables are graphed 
as differences in percentage rates. For example, the figure indicates that 
in 1983 the real GNP growth rate (at 1.8 per cent) would be a half per¬ 
centage point higher than it would have been under the base case (1.3 per 
cent). 

The inflation rate would rise somewhat less than real growth in 1983, 
but it would continue to rise thereafter, adding almost an extra half per¬ 
centage point on the base-case inflation rate in 1984. The inflation rate 
would then also return to base-case levels. 

So the expenditure shock generates a temporary increase in the 
inflation rate as well as in GNP, and generates a temporary decline in the 
unemployment rate of somewhat less than half a percentage point. This 
assumes that a fair number of discouraged workers come back in as new 
jobs are created, but nevertheless it implies additional jobs on the order of 
60,000 to 70,000 over the next two years. So even though the results are 
not spectacular, they indicate that we could at least get some additional 
output and some additional jobs from this policy, without a great cost in 
terms of inflation. Essentially, the policy involves a backward shifting of 
growth from 1985-6 to 1983-4. Provided that the policy is indeed revers¬ 
ible, it should have the contra-cyclical pattern of effects shown in the 
figures, with minimal effects on financial market expectations. 

The impact of this policy on the deficit is not all that large. Figure 
12 presents the impact of the policy on the federal deficit and federal debt 
expressed as a percentage of GNP. As is apparent, the federal deficit as 
a percentage of GNP initially goes up by something less than a quarter of 
a percentage point, and then comes back to the base-case ratio. 


127 








Figure 10 

Bond-financed expenditure increase: 
change in gnp and cpi 


Per cent 



Figure 11 

Bond-financed expenditure increase: 

change in growth, inflation, and unemployment rates 

Percentage 

points 



The debt/GNP ratio goes down initially because nominal GNP goes up 
more than the debt does, so that the ratio actually comes down. In other 
words, the government has stimulated more nominal GNP than it has gen¬ 
erated debt - although some of the GNP increase is real and some of it is 
simply inflationary. 

In the second variation of the temporary expenditure-increase exper- 


128 















iment, we assume that the expenditure increase is accompanied by an 
accommodative monetary policy that adjusts the money supply so as to hold 
short-term interest rates at their base-case levels. Such a policy should 
be feasible, given the transitory nature of the government expenditure 
increases. 11 

The economic implications of this 'money-financed' expenditure in¬ 
crease are portrayed in Figures 13 and 14. As is apparent, the impacts 
on all three policy objectives are somewhat larger than they are in the 
bond-financed experiment. The peak impact on real output is now a full 1 
per cent in 1984, and the price level effects are above 1 per cent in 1985. 

There is a larger short-term reduction of almost one-half a percentage 

point in the unemployment rate. 

Once the stimulus is removed, both real growth and inflation diminish, 
and indeed drop below base-case rates over the medium term. At the 
projection horizon, the economy is moving back towards equilibrium, though 
at a slower pace than in the previous experiment. 

The fiscal implications of such a 'money-financed' expenditure increase 
are presented in Figure 15. As is clear, the immediate deficit effects are 
much weaker than they are in the bond-financed expenditure experiment, 
and the deficit/GNP ratio drops for a period during the mid-1980s as a 
result of the larger stimulus. It is notable that the debt to GNP ratio is 

below its base-case level in every year of the projection. This outcome 

reflects not only the larger stimulus to GNP provided by this alternative, 
but also the fact that more of the deficit is financed by money creation 
and hence does not enter into funded debt. 

POLICY EXPERIMENTS INVOLVING INDEXING OF CAPITAL COST 
ALLOWANCES 

We next consider a somewhat different policy, one that is intended to 
provide stimulus to supply as well as to demand. The particular policy we 
examine is full indexation of capital cost allowances under the corporate 
income tax. This approach effectively removes the negative impact of 
inflation on the real value of capital cost allowance deductions, thereby 
providing a powerful stimulus to increased capital formation. As Figure 16 
shows, the resulting extra boost to business fixed investment generates a 
fairly healthy rise in real GNP, but this rise takes longer to occur than it 
does in the two temporary expenditure-increase experiments. Indeed, the 
peak impact on real output is not obtained until 1985. 


129 











Figure 12 

Bond-financed expenditure increase: 
change in ratio of federal deficit and debt to gnp 

Percentage 

points 



Figure 13 

Money-financed expenditure increase: 
change in gnp and cpi 


Per cent 



130 












Figure 14 

Money-financed expenditure increase: 

change in growth, inflation, and unemployment rates 

Percentage 

points 



Figure 15 

Money-financed expenditure increase: 
change in ratio of federal deficit and debt to gnp 

Percentage 

points 



131 











Unfortunately, the demand stimulus resulting from this more per¬ 
manent shock also increases the rate of inflation over the whole projection 
interval, as is shown in Figure 17. However, this result is obtained using 
the model under a short-term 'markup' price setting mechanism that does 
not allow the productivity effects of the additional capital formation - i.e., 
the additional potential created - to feed through into additional output 
and generate downward price pressures. If instead we equip the policy 
with a more neoclassical price mechanism, whereby price is determined by 
demand and supply conditions, the inflationary impacts are not nearly as 
great (see Figures 19 and 20). As a matter of fact, the rate of inflation 
actually comes down for a while relative to base-case rates. We do 
observe an upsurge in inflation in the last year of the projection, but that 
turns out to be related to an exchange rate movement. 

A third variation on the indexing of capital cost allowances combines 
the neoclassical price mechanism with a monetary policy regime that 
neutralizes any exchange rate impacts. As is clear from Figure 22, the 
economic impact of the policy is now a gradual increase in real aggregate 
output, combined with a gradual decrease in the price level. The impacts 
on the respective rates of growth are very modest, as is shown in 
Figure 23. 

It is clear from these experiments that if the increase in potential 
stimulated by the policy change is allowed to feed through to productivity 
and unit costs, much of the inflationary impact is eliminated, while real 
output increases over the medium term. 

Another interesting aspect of these experiments is that the deficit 
actually declines in the initial years. It does so because the indexing of 
capital cost allowances promises firms future lower taxes for investing now . 
Their investment today boosts tax revenues today and cuts transfers in 
the near term (via increased investment activity), thereby reducing the 
federal deficit instead of increasing it. In other words, by promising to 
reduce taxes later, when they would otherwise have increased through 
inflation, but by generating the benefits of the real stimulus that results 
from this reduction in the present, this approach actually produces a small 
downward movement in the near-term federal deficit. This improvement in 
the deficit relative to the base case naturally generates a near-term reduc¬ 
tion in the federal debt/GNP ratio (see Figures 18, 21, and 24). 12 


132 






Figure 16 

Indexing of cca with markup price rule: 
change in gnp and cpi 

Per cent 



Figure 17 

Indexing of cca with markup price rule: 

change in growth, inflation, and unemployment rates 

Percentage 

points 



POLICY EXPERIMENT WITH IMPROVED UNEMPLOYMENT INSURANCE 
SYSTEM 


The final experiment involves a major change in the Unemployment Insur¬ 
ance system to restore it as an effective automatic stabilizer in the tax/ 
transfer system. Note that under the current UI system the UI contribu- 


133 
















Figure 18 

Indexing of cca with markup pricing rule: 
change in ratio of federal deficit and debt to gnp 

Percentage 

points 



Figure 19 

Indexing of cca with flexible price rule: 
change in gnp and cpi 

Per cent 



tion rate is adjusted every year to eliminate (or partially eliminate) the 
previous year's deficit or surplus. There may be some advantage to 
adjusting the UI system to pay its own way over the longer term, but the 
current system adjusts it far too rapidly, reducing its automatic stabilizing 
properties. 

In the present experiment, the rate of adjustment of the UI contri- 


134 








Figure 20 

Indexing of cca with flexible price rule: 

change in growth, inflation, and unemployment rates 

Percentage 

points 



Figure 21 

Indexing of cca with flexible price rule: 
change in ratio of federal deficit and debt to gnp 

Percentage 

points 



bution rate is limited to the level required to eliminate only 20 per cent of 
the previous year's surplus or deficit. As is clear in Figure 25, the net 
result of this change is a stimulus to GNP over the next few years while 
unemployment is high. However, the adjustment mechanism gradually 
balances the UI account as the economy gets closer and closer to full 
employment. 


135 














Figure 22 

CCA indexation under flexible price rule and ‘fixed’ exchange rate: 
change in gnp and cpi 


Percent 



Figure 23 

CCA indexing with flexible price rule and ‘fixed’ exchange rate: 
change in growth, inflation, and unemployment rates 

Percentage 

points 



This change has little impact on the price level. One reason why an 
inflationary impact does not occur is that over half of the total contribu¬ 
tion to the UI fund is paid directly by employers, and any increase in 
employer contributions represents an increase in unit labour costs that 
tends to be marked up in higher prices. If this year's scheduled increase 
in UI contributions were reduced, there would be less of an increase in 


136 




















unit labour costs and consequently a downward cost push effect on infla¬ 
tion that would offset the upward pull on prices generated by the asso¬ 
ciated increase in output and employment. As a result, the inflation rate 
would hardly be affected at all, as is seen in Figure 26. 

Figure 27 shows the impact of this change in policy regime on the 
deficit. For the next few years, obviously, the deficit increases, but as 
time passes the system adjusts automatically to bring the deficit down 
towards the base-case level. In fact, if we carried the projection to 1995, 
and got closer to full employment, the deficit would most likely be reduced 
below the base-case level. 

OVERVIEW 

Taken together, the six policy experiments indicate that there is room for 
manoeuvre by the fiscal authorities, given that the measures adopted do 
not increase the medium-term structural deficit. This room for manoeuvre 
exists because of the enormous excess supply created by the past two 
recessions. In the base-case simulation, this excess supply diminishes 
gradually over the decade, indicating that there would be little risk of 
re-accelerating inflation under these conditions. 

Of course, should recovery proceed at a more vigorous pace, this 
margin of safety from rekindled inflation would disappear faster. 
However, in most of the above policy experiments, the price inflation 
effects moderate or reverse in the second half of the decade. The results 
therefore indicate that a carefully designed fiscal stimulus could help to 
speed recovery from the recent deep recession, without necessarily 
building in inflationary pressures in the later years of the decade. 




137 







Figure 24 

CCA indexing with flexible price rule and ‘fixed’ exchange rate: 
change in ratio of federal deficit and debt to gnp 

Percentage 

points 



Figure 25 

Modified ui system: 
change in gnp and cpi 


Per cent 



138 










Figure 26 

Modified ui system: 

change in growth, inflation, and unemployment rates 

Percentage 

points 



Figure 27 

Modified ui system: 

change in ratio of federal deficit and debt to gnp 

Percentage 

points 







139 











APPENDIX A 

BACKGROUND NOTES AND TABLES TO PROJECTIONS AND 
ALTERNATIVES 


TABLE A1 

Base case projection: summary table 


History 


Projection 


Average 



1981 

1982 

1983 

1984 

1985 

1986-90 

Growth rates (% change) 

Real GNP 

3.1 

-4.8 

1.3 

4.1 

4.6 

3.7 

Consumption 

Non-residential 

1.9 

-2.5 

1.1 

3.6 

4.5 

3.9 

investment 

6.4 

-13.0 

-17.6 

11.3 

9.9 

8.1 

Residential investment 

5.6 

-23.8 

14.1 

16.0 

12.3 

4.1 

Government 

.7 

2.2 

1.7 

1.6 

1.5 

1.5 

Exports 

1.6 

-1.5 

1.8 

3.7 

5.4 

6.2 

Imports 

2.6 

-10.4 

3.5 

6.4 

7.8 

6.6 

Employment 

2.6 

-3.3 

-1.1 

1.7 

3.1 

2.5 

GNP deflator 

10.1 

10.7 

6.9 

5.5 

6.0 

6.0 

CPI 

12.5 

10.6 

7.3 

5.4 

6.4 

6.2 

Labour productivity 

.4 

-1.7 

1.7 

2.3 

1.5 

1.2 

Money supply (Ml) 
Corporate profits 

4.3 

2.3 

10.3 

9.1 

7.5 

7.5 

(pre-tax) 

Levels (percentages) 

-10.4 

-32.1 

22.6 

50.2 

24.8 

15.4 

1990 

level 

Unemployment rate 

7.6 

11.1 

13.2 

13.1 

11.8 

8.6 

90-day paper rate 

18.3 

13.9 

9.9 

9.6 

10.0 

9.6 

Industrial bond rate 

16.3 

15.7 

12.3 

10.7 

10.3 

10.9 

Exchange rate 

(US$/C$) 

$ billions 

.834 

.811 

.815 

.82 

.826 

.84 

Current account of 

balance of payments 

-5.3 

2.7 

6.7 

5.1 

2.4 

-2.2 

Aggregate government 

deficit 

4.0 

17.5 

27.6 

22.4 

18.1 

12.5 

Federal deficit 

8.0 

21.1 

28.3 

25.3 

23.9 

18.9 


140 








TABLE A1 (continued) 



History 

1981 1982 

1983 

Projection 

1984 

1985 

1990 

Federal deficit 
adjusted for 
inflation 

-4.2 

8.5 

17.9 

16.3 

11.5 

1.2 

Per cent 







Federal deficit 

as % of GNP 

2.4 

6.1 

7.5 

6.1 

5.2 

2.5 

Inflation-adjusted 

deficit as % of GNP 

-1.3 

2.4 

4.7 

3.9 

2.5 

.2 

Aggregate deficit as 

% of GNP 

1.2 

5.0 

7.3 

5.4 

3.9 

1.7 

Federal debt as 

% of GNP 

29.4 

34.0 

40.1 

44.5 

46.8 

47.0 


1981 

1982 

1983 

1984 

1985 

Average 

1986-90 

Foreign (US) assumptions: 
(per cent) 

U.S. real growth 

1.9 

-1.7 

2.2 

4.4 

4.2 

3.2 

U.S. inflation 
(GNP deflator) 

9.4 

5.9 

5.2 

5.7 

6.2 

6.3 

U.S. industrial 
bond rate 

15 

13.9 

10.6 

10.4 

10.6 

10.3 

Adjusted import 
price of crude oil 
(US$ at Montreal) 

35.0 

34.2 

30.0 

30.0 

31.5 

1990 

47.1 

NOTES ON EXPERIMENTS 

Experiment 1: temporary 

fiscal 

expansion 






The figures in Table A2 are actual changes made in the model. They 
represent a current-dollar expenditure of $1.5 billion over the last three 
quarters of 1983 and the four quarters of 1984. The current-dollar capital 
expenditures are $250 million in 1983 and 1985, and $500 million in 1984. 
The employment figures were calculated from the above expenditures, 


141 




















TABLE A2 

Changes in exogenous variables to represent temporary fiscal expansion 


Federal current 
expenditure 
(millions of 1971$, 
annual rates) 

Federal capital 
expenditure 
(millions of 1971$, 
annual rates) 


Federal 

employment 

(’000) 

1983:2-4 607 

• 


111. 


61 

1984:1-4 429 

• 


159. 


43 

1985:1-4 0.0 




76. 


10 

TABLE A3 

Experiment 1: Temporary Expenditure - 

'Bond-Financed ’ 



Change in: 

1983 

1984 

1985 

1986 

1987 

1990 

Real GNP (millions 
of 1971$) 

756 

858 

86 

-376 

-464 

57 

% of base 

.58 

.64 

.06 

-.26 

-.30 

.03 

CPI (% of base) 

.24 

.60 

.57 

.40 

.38 

.46 

Unemployment rate 
(per cent) 

-. 26 

-.28 

- .04 

.14 

.20 

.03 

Aggregate government 
balance ($ million) 

-731 

-742 

-547 

-1096 

-987 

-455 

TABLE A4 

Experiment 1-A: Temporary Expenditure 

- 'Money-Financed’ 



Change in: 

1983 

1984 

1985 

1986 

1987 

1990 

Real GNP (millions 
of 1971$) 

862 

1341 

942 

460 

-49 

-653 

% of base 

.67 

.99 

.67 

.31 

-.03 

-.39 

CPI (% of base) 

.34 

.96 

1.29 

1.27 

1.19 

.42 

Unemployment rate 
(per cent) 

-.27 

-.37 

-.23 

-.05 

.12 

.30 

Aggregate government 
balance ($ million) 

-525 

177 

1161 

478 

-204 

-1177 


assuming that approximately half the economy-wide average real wage 
would be paid. 


142 











Experiment 2: capital cost indexing 


At present, capital cost allowances (CCAs) for tax purposes are not adjus¬ 
ted for the effects of inflation on the replacement costs of capital goods. 
In principle, capital cost allowances could be adjusted to reflect inflation, 
thereby giving an on-going tax benefit to corporations. Experiments 2 
and 2-A implement a scheme of full indexation of CCAs, in which terms 
reflecting the indexation enter into both the 'user cost' variables - and so 
into investment - and corporate tax collections. (For further details, see 
the FOCUS model documentation.) The indexing feature is assumed to 
apply only to capital goods purchased after the scheme is put into effect; 
in this way it can generate new investment at little cost in near-term tax 
revenue. 

Experiment 3: restoration of the unemployment insurance 'automatic 
stabilizer' 


At present, the rate of contribution to the Unemployment Insurance system 
is adjusted at the beginning of each year in light of the fund's perfor¬ 
mance in the previous year and expectations for its performance in the 
year to come. Thus, the rate was slightly reduced in 1982 after the 
'surplus' year of 1981, while it was increased by about 40 per cent in the 
wake of the fund's 1982 deficit. Increasing UI premiums on this basis is, 
of course, equivalent to increasing taxes during a recession. On the 
other hand, the current legislation requires the contribution rate to be cut 
as the unemployment rate falls, an arrangement that could fuel a boom. 

In the FOCUS model, the mechanism for adjusting the contribution 
rate is only approximate. It is as follows: 

New rate = Old rate * a * Payment s . R p rev j ous year. 

Receipts 

In practice, a value for a of 1.0 seems to best approximate the past 
adjustments of this rate. Under this formula, a year of zero net deficit in 
the fund will leave the rate unchanged. An excess of payments over 
receipts will, of course, raise the rate. 

For the experiment making the UI system more of an automatic stabil¬ 
izer again, the value of a was changed to .2 (from 1.0), starting with the 


143 






TABLE A5 

Experiment 2: CCA indexing ('mark-up' price) 


Change in: 

1983 

1984 

1985 

1986 

1987 

1990 

Real GNP (millions 
of 1971$) 

393 

1321 

1637 

1434 

1088 

836 

% of base 

(-30) 

(.98) 

(1.2) 

(.97) 

(-71) 

(.49) 

CPI (% of base) 

.05 

.45 

1.1 

1.3 

1.6 

2.9 

Unemployment rate 
(per cent) 

-. 06 

-.31 

-.52 

-.55 

-.46 

-.41 

Aggregate government 
balance ($ million) 

360 

1417 

1653 

408 

-534 

-2324 

TABLE A6 

Experiment 2-A: CCA 

indexing 

(market- 

clearing 

price) 



Change in: 

1983 

1984 

1985 

1986 

1987 

1990 

Real GNP (millions 
of 1971$) 

368 

1193 

1698 

1790 

1437 

1959 

% of base 

(.28) 

(.88) 

(1.2) 

(1.2) 

(.94) 

(1.2) 

CPI (% of base) 

.11 

.71 

1.2 

.94 

.34 

1.4 

Unemployment rate 
(per cent) 

-.07 

-.33 

-.54 

-.51 

-.29 

-.33 

Aggregate government 
balance ($ million) 

424 

1645 

1571 

-306 

-2371 

923 


1983 adjustment. As a result, the rate rises less in 1983 and 1984, but 
falls less in subsequent years when unemployment improves, thus 'damping' 
the downturn-upturn fluctuation of the 1980s. 


144 








TABLE A7 


Experiment 2-B: CCA Indexing 
held at base by monetary poli 

(market-clearing 
cy) 

price with 

exchange 

rate 

Change in: 

1983 

1984 

1985 

1986 

1987 

1990 

Real GNP (millions 
of 1971$) 

212 

481 

454 

687 

881 

899 

% of base 

(.16) 

(.36) 

(.32) 

(.47) 

(.58) 

(.54) 

CPI (% of base) 

.05 

0.0 

-.26 

-.25 

-.30 

-.39 

Unemployment rate 
(per cent) 

-.04 

-.11 

-.07 

-.11 

-.18 

-.19 

Aggregate government 
balance ($ million) 

230 

21 

-898 

-460 

-761 

-2976 

TABLE A8 

Experiment 3: 'Damped 

1 UI contribution 

adj ustment 



Change in: 

1983 

1984 

1985 

1986 

1987 

1990 

Real GNP (millions 
of 1971$) 

345 

1254 

1621 

1080 

190 

-562 

% of base 

(.27) 

(.93) 

(1.1) 

(.73) 

(.12) 

(-.33) 

CPI (% of base) 

.06 

.12 

.19 

.28 

.46 

1.1 

Unemployment rate 
(per cent) 

-.13 

-.44 

-. 56 

-.37 

-.04 

.27 

Aggregate government 
balance ($ million) 

-1068 

-2583 

-1980 

-1486 

-893 

39 


APPENDIX B 

LOWER-INFLATION ALTERNATIVE 

At present, projections of inflation for the 1980s vary widely. The pro¬ 
jection presented above, in which both U.S. and Canadian inflation 
average about 6 per cent over 1984-90, may be unduly pessimistic in light 
of current information and a continuing large output gap. The 6 per cent 
rate could still apply, given at least one major supply shock during the 
decade (of course, this would lead to a much less regular pattern of 
inflation than in the projection) and continued 'catch-up' wage and price 


145 












TABLE B1 

Impact of reduced inflation 


Change in: 

1984 


1985 

1986 

1987 

1988 

1989 

1990 

Inflation rate 
(percent) (CPI) 

-.25 


-.5 

-.75 

-1.0 

-1.25 

-1.5 

-1.75 

Federal deficit 
($ billions) 

-.040 


■.109 

-.354 

-.868 

-1.462 

-2.129 

-2.930 

Aggregate deficit 
($ billions) 

-.068 


-.112 

-.372 

-1.091 

-1.955 

-2.923 

-4.055 

Federal deficit 
adjusted for 
inflation 
($ billions) 

.343 


.783 

1.168 

1.415 

1.705 

2.027 

2.342 

Federal deficit 
as percentage 
of GNP 

.10 


.31 

. 61 

.95 

1.31 

1.68 

2.08 

pressure left over 

from 

the 

'great 

recession' and 'six 

and five 

.' But 

inflation could also 

be lower, 

with perhaps 

the most 

likely 

scenario 

being a 

gradual reduction 

below 

our 

base-case 6 

per cent rate 

as the 

decade 

progresses. This 

appendix 

briefly 

examines the 

implications of 

a lower 


inflation rate for the deficit. 

The alternative projection differs from the base case in that the 
inflation rate, both in the U.S. and Canada, is reduced by 1/4 of a per¬ 
centage point each year starting in 1984; thus the 1990 inflation rate is 
about 4 per cent. Interest rates are reduced correspondingly, so as to 
keep real rates at base-case levels, and the same real changes in world oil 
prices are maintained; but the real components of GNP are fixed at base- 
case levels - i.e., they in no way respond to the reduced inflation. By 
holding real expenditure components at base levels, we are more easily 
able to isolate the impact on the deficit of lower inflation alone. Of 
course, reduced inflationary pressures could open more room in which to 
manoeuvre between the twin goals of stimulating the economy and keeping 
inflation low. More robust growth would itself have implications for the 
deficit. 

As can be seen from Table Bl, the results of reduced inflation are 
mixed. In purely nominal terms, the sizes of both the federal and aggre¬ 
gate deficits are reduced by lower inflation. By 1990, this improvement is 


146 





by about $3 billion for the federal government and about $4 billion in 
aggregate. However, the inflation-adjusted federal deficit increases by a 
little over $2 billion by 1990 under a lower inflation scenario. (This 
result, combined with the figures in Table B1 yields a total inflation-adjus¬ 
ted federal deficit of about $3.5 billion in 1990.) As a percentage of GNP, 
the projected federal deficit also rises under reduced inflation; nominal 
GNP declines slightly more than the deficit. The extent of the deficit 
increase is about 2 percentage points; nevertheless, as in the base case, 
the ratio of debt to GNP still peaks in 1988. 

This alternative indicates that while the tax and expenditure system 
is by no means fully inflation-proof, assuming a different and lower infla¬ 
tion rate does not alter the basic lessons of the projection described in the 
text. It is instructive how the unadjusted and inflation-adjusted deficits 
move in opposite directions in this alternative, underscoring the importance 
of taking the inflation adjustment into account. 

NOTES 

1 This projection was developed using the FOCUS model maintained by the 
Institute for Policy Analysis, University of Toronto. 

2 For details of the base-case simulation and the policy alternatives, 
see Appendix A. 

3 FOCUS is described in detail in Institute for Policy Analysis (1982) 
Focus-User's Manual . 

4 This is apparent from a review of the longer-term comparative simula¬ 
tion results from several Canadian models published by the Bank of 
Canada and the Department of Finance (1982). 

5 The DRI simulation, TRENDLONG 1282, is described in detail in Data 

Resources Inc. (1982) U.S. Long-Term Review . 

6 The recent OPEC price adjustment represents a 15 per cent decline in 
world oil prices this year. 

7 This refers to the wellhead price of 'old' conventional oil. The 

blended price paid by buyers will of course be somewhat higher. 

8 Clearly, the Canadian inflation rate can deviate from the U.S. infla¬ 
tion rate in the shorter run. However, we assume that the Canadian 
and U.S. inflation paths will roughly track together over the longer 
run, as they have in the past. 

9 The term used for federal debt in FOCUS is only an approximation, 

designed to be easily constructed from other model variables. In 
essence, it is simply the accumulation of past federal deficits (sur¬ 
pluses reduce debt) net of any monetization of the debt. That is, 

bond holdings of the Bank of Canada are not included. 


147 





10 For further detail see Appendix A. Increases in government expendi¬ 
tures to mitigate the recession have recently been analyzed by DRI and 
Informetrica. 

11 For a more permanent policy change, the central bank may not be able 
to control short-term interest rates without serious problems of 
instability. 

12 The debt/GNP patterns generated by the three experiments differ sig¬ 
nificantly over the medium term. This circumstance mainly reflects 
the differing impact of the stimulus on the aggregate price level 
under the different price and exchange rate regimes. 


Leo de Bever* 

Given the limited time, it does not seem fruitful to comment on details of 
the simulations performed by Dungan and Wilson. I want to focus instead 
on the key issues. 

Is the base case outlook plausible? 

Do we like what it tells us about the future? 

Are we focusing on the right problem (i.e., the deficit)? 

Is the deficit as harmless as it is made out to be in the adjustments 
carried out by the authors? 

We at Chase Econometrics are in rough agreement with the long-term 
scenario under discussion, although we are somewhat more pessimistic 
about growth prospects (about 3 per cent as opposed to 3.7 per cent). 
We are forced to the conclusion that 6 per cent inflation is about the best 
one can reasonably expect over the next decade, on average. We arrive 
at equally pessimistic labour productivity estimates, but see the federal 
deficit ending up somewhat lower at $8-10 billion by 1990, probably 
because we built in a fiscal response to continuing deficits on the tax 
side. 

The main thing that disturbs me in all of this is that for the period 
1970-90, there are very few years in which the federal government is in 
surplus, even after adjusting for inflation. Without wishing to comment on 
the appropriateness of the specific adjustments made, I find the implicit 
debt position very disturbing. The authors mention that we saw a debt/ 

* Director, Chase Econometrics Canada. 


148 


GNP ratio of 50 per cent in the 1950s, but if I remember correctly, this 
was because of a war. To many of us 1982 may have been a war, but the 
situations are not really comparable. 

I basically agree that a $30 billion deficit in 1983 is unavoidable, and 
a quick reduction difficult to achieve, and probably undesirable. I am not 
particularly against short-term relief for the unemployed, although I would 
like to see some strings attached. There is again talk of stimulating 
construction, one of the sectors characterized by high wage levels and 
high unemployment rates. In sectors where markets are obviously not in 
balance, we should insist on some self-help in the form of reduced wage 
levels. 

The results of the alternative policy scenarios clearly indicate some¬ 
thing that disturbs me about a lot of proposals for stimulus: they buy 
temporary, short-term gains in output, at the cost of a long-term increase 
in the price level. The only exception to this is the proposal for indexa¬ 
tion of capital cost allowances. 

Our own work has indicated that fiscal stimulus has to operate on the 
productivity nexus to have lasting positive effects on prices, deficits, and 
debt levels. We are not convinced that massive government expenditure is 
required. The main task at hand may be to provide a stable environment 
in which human and fixed capital formation will be undertaken. If we are 
to get ready for the competitive environment of the eighties, as Mr. 
Lalonde urged us to do, both of these forms of investment are essential. 
Therefore, we deplore recent proposals to further cut education expendi¬ 
tures. On the fixed capital formation side, fiscal policy seems to have 
made the worst of a bad situation over the last few years: to the adver¬ 
sity caused by monetary upheavals, we have added confusion (and expec¬ 
tations of future confusion) in taxation policy. This is in part responsible 
for a replacement of inflationary expectations with expectations of low 
growth, which may be just as hard to eradicate. Stability in fiscal policy 
would go a long way towards correcting this problem, and avoiding a 
future of low growth, high unemployment, and high deficits. In closing, I 
was struck by a comment the authors made about Canadian performance 
being tied to U.S. performance. Without ignoring obvious linkages, it 
strikes me that there is no reason why we could not set our aims somewhat 
higher, as Lester Thurow recently suggested. The U.S. projections 
underlying the Dungan-Wilson study do not portray a future that is parti¬ 
cularly worth emulating, and I suggest that there is no reason to be 
satisfied when we do no worse than the U.S. 


149 


What can macroeconomic theory 
tell us about the way deficits should 
be measured? 

Michael Parkin* 


INTRODUCTION 

Deficits in the federal government accounts of $30 billion - approximately 
10 per cent of GDP - have generated two diametrically opposed reactions. 
Both are reactions of alarm, but the perceived sources of that alarm stand 
in direct conflict with each other. 

Perhaps the most common reaction, certainly the most common one 
among non-economists, is that of seeing the deficit as being too big, of 
seeing it the source of high real interest rates and (until recently) stub¬ 
born (possibly to be subsequently renewed) inflation. High interest rates, 
in turn, are held responsible for low levels of expenditure on capital 
equipment and household durable goods, and low levels of expenditure are 
identified as the source of high unemployment. Thus, according to this 
view - a view rarely found in academic writings but common in less formal 
or journalistic discussion - the deficit is too big, and as a consequence 
interest rates, inflation, and unemployment are all too high. 

The second reaction, which comes mainly from economists, reaches 
exactly the opposite conclusion. Using traditional concepts of opportunity 
cost and the newer (though by no means new) ideas of post-Keynesian 
macroeconomic theory, economists suggest that the conventionally calculated 
deficit requires adjustment for (at least) three clearly identifiable factors 
that are mistreated in the financial accounting procedures employed by 
national income and public sector accountants. One of these factors would 
make the deficit, as correctly measured, larger than the accounting 
measure makes it, but the other two factors would make the deficit 
smaller. 

* Professor, Department of Economics, University of Western Ontario. 


150 


It is the government sector accounts treatment of pension operations 
that makes the conventionally measured deficit smaller than the correctly 
calculated one. This result arises from the fact that the financial accounts 
treat the pension funds on a cash-receipts rather than an accruals basis. 
In the early stages of a pension program, such as the Canada Pension 
Plan, cash receipts from the large number of working contributors sub¬ 
stantially exceed the cash payments to those already retired and eligible 
for a pension. However, the government incurs a liability to existing 
contributors, just as it incurs a liability when it sells a bond. At present 
there is a surplus in the cash flow of the Canada Pension Plan in excess 
of $1 billion annually. In the absence of changes in either benefit rates 
or contributions (or subsidies from general taxation), this surplus will 
become a substantial (and increasing) deficit as the plan reaches maturity 
and the stock of eligible pension recipients increases. Thus the cash flow 
approach understates the true deficit, compared with what would be re¬ 
vealed by an accruals approach. Economists have suggested that the best 
way of handling government pension plans is to remove them completely 
from the government sector accounts, making it possible to measure the 
deficit net of any surplus contributed by the pension plan. 

The factors that economists believe misleadingly swell the deficit are 
inflation and unemployment. In part, the high market rates of interest 
that the government is currently paying on its debt merely reflect compen¬ 
sation to bond holders for the falling value of money. This 'inflation 
component' of interest payments on the government's debt should be eli¬ 
minated from calculations of the deficit. Furthermore, the real value of 
the government's monetary - non-interest bearing liability - is falling at 
the rate of inflation, and this factor too should be allowed for. In terms 
of opportunity cost (as opposed to financial) accounting, there can be no 
question that these adjustments to the financial accounts are required in 
order to establish the current real resource transfers that are taking place 
between the government and private sectors. 

The rationale for adjusting the deficit to take unemployment into 
account derives not from opportunity cost considerations but from 
Keynesian macroeconomic theory. Central to Keynesian economics is the 
proposition that high unemployment and low real income result from insuf¬ 
ficient aggregate demand. If aggregate demand were at an adequate level, 
there would be 'full employment.' High unemployment and low income swell 
the government deficit by inducing extra unemployment compensation pay- 


151 





ments and lowering tax receipts. These induced components of the deficit 
are transitory and will disappear when the economy returns to its full- 
employment state. They should not, therefore, be counted as part of the 
'core' (or 'chronic') deficit. 

Allowing for all these factors results in the transformation of what at 
first appears to be an alarming deficit into an equally alarming surplus. 
According to the calculations of Bossons and Dungan (1983, 1), 'if the 
economy were at full employment and the government surplus or deficit 
were correctly measured, the current taxation and expenditure programs of 
all governments combined would yield a surplus of approximately $6 
billion.' This 'discovery' of a surplus has led these and perhaps a majority 
of economists to conclude that 'as a result, the current fiscal position of 
Canadian governments is depressing rather than stimulating the Canadian 
economy.' (Bossons and Dungan 1983, 1.) That is, far from the big 
deficit being the source of high interest rates, inflation, and unemploy¬ 
ment, it is the big adjusted surplus that is responsible for these problems. 

Which view is correct? What do economic theory in general and 
macroeconomic theory in particular tell us about the way in which deficits 
should be measured? These are the questions to be addressed here. The 
first section following this introduction reviews the alternative ways in 
which the conventional accounting deficit may be adjusted to arrive at a 
variety of alternatively defined deficits (one of which will be the 'full- 
employment' real deficit/surplus) conventionally calculated by economists. 
I shall then pose the question: Which of these alternatives (if any) is the 
right one? The answer to this question will be of the 'it-depends-on' 
variety. In this case, it depends on which of various questions concern¬ 
ing macroeconomic performance we wish to address, and also on which 
model of economic behaviour best describes the world in which we live. 

The subsequent four sections of the paper identify and address four 
questions of substance concerning macroeconomic performance. These 
questions concern the influence of the deficit on aggregate demand, the 
real rate of interest, the levels of output and unemployment, and inflation. 
In addressing these questions, I shall use a theoretical framework that is 
sufficiently general to permit an examination of the consequences for the 
way in which deficits should be measured of alternative views about macro- 
economic behaviour. The final section of the paper provides a brief 
summary of the main conclusions. 

The main conclusion reached is worth anticipating here. It is that 


152 


macroeconomic theory is much less clear-cut in its implications for how the 
deficit should be measured than most economists seem ready to believe. 
However, to the extent that macroeconomic theory does offer guidance, it 
suggests that we face a potentially serious problem arising from an evolv¬ 
ing deficit that has been too big, is too big, and promises to remain too 
big. I offer no calculations in this paper, so I am not able to say how 
large a problem we face. However, the rosy picture painted by the real 
full-employment deficit calculations is clearly misleading and probably 
seriously so. 

I also offer no policy prescription. (This may be inferred as imply¬ 
ing a preference for a neutral policy stance under current circumstances.) 
My only prescription is for more research. We need to know more not only 
about macroeconomic behaviour, but also about the political processes 
whereby fiscal policy and the resulting deficit are generated. I suspect 
that until we have a better knowledge of these matters we shall not make 
any serious progress in improving the performance of fiscal policy. We 
shall be able to inject additional (unwanted) noise, but we shall not have 
the capacity to change the underlying structural process. 

FINANCIAL AND ECONOMIC MEASURES OF THE DEFICIT 

There are two readily available bases for measuring government economic 
activity. They are the National Accounts basis and the Public Accounts 
basis. The National Accounts basis is more comprehensive than the Public 
Accounts basis. The two bases differ substantially in their measure of 
total revenues and total expenditures; the difference between their assess¬ 
ments of the deficit is much smaller. In what follows I shall (implicitly) 
regard the National Accounts measures of government activity as the 
appropriate ones. 

A starting point for discussing the deficit is the government's 
receipts and payments in current dollars. Purchases of goods and ser¬ 
vices, transfer payments, the payment of interest on outstanding bonds, 
and the redemption of bonds all have to be financed by tax revenues, the 
sale of new bonds, and the creation of new money balances. This neces¬ 
sary equality between outlays and receipts may be expressed as follows: 

G t - V t + (iSp B t - = °< 


153 








where 


G t is nominal government spending on goods and services, 

V t is nominal tax receipts net of transfer payments, 

R t is the nominal rate of interest, 

B t ^ is the value of bonds redeemed at t, 

is the value of bonds to be redeemed at t+1, 
is the nominal money supply outstanding at the end of the 
previous period, and 

M t is the nominal money supply outstanding at the end of the 
current period. 

The first two terms in eguation (1) are self-explanatory. The other 
terms reguire some elaboration. To keep things simple, I am supposing 
that the government issues only one-year bonds. 1 Because of this 
assumption, the amount paid out in the current period to redeem old bonds 
is the entire outstanding stock, B^_^. The analytically convenient fiction 
used is that in the previous year the government sold commitments to pay 
out B t _- L one year hence. In the current year the government has sold a 
commitment to pay out B^ one year hence. The price for which the gov¬ 
ernment has sold such a promise is l/Cl+R^) - the price of a one-period 
bond in the current period. Thus, R^B^/O+Rj.) represents the implicit 
interest payments on the current year's bonds. 2 The final two terms in 
the above equation capture the contribution to the government's budget of 
newly created money. This is to be thought of as the liability of only the 
central bank (Bank of Canada) and not of the chartered banks as well. 
In other words, equation (1) consolidates the Bank of Canada and the 
government. Government bond transactions with the Bank of Canada are 
not included in B. 

Care is required to interpret equation (1) correctly for an open 
economy. In an open economy, the money supply has to be redefined as 
the equivalent of the stock of domestic securities held by the central bank 
(domestic credit) and not as the total high-powered money stock. Under a 
cleanly floating exchange rate regime where the stock of foreign exchange 
reserves were either zero or constant, the change in domestic credit 
(domestic credit expansion or DCE) would, of course, be equal to the 
change in the nominal stock of high-power money (IV^-M^). 3 

Although equation (1) describes government receipts and outlays on a 
cash basis, a suitable redefinition of the equation's terms would allow us to 
describe government income and expenditure on an accruals basis instead. 


154 


An accruals-base definition of income and expenditure differs from a cash- 
receipts definition in exactly the same way that an accountant's profit-and- 
loss account differs from his cash account. It reflects changes in 
liabilities even if those liabilities are not represented by an explicit 
contractual transaction. As was noted in the introduction, one aspect of 

government activity - its involvement in pension plans - makes it important 

to recognize the distinction between cash receipts and accruals for the 
purpose of calculating the government's budget position. When the budget 
is calculated on a cash-accounts basis, the government's pension activities 
appear in the term V in eguation (1). V will include all of the payments 
made to existing pension recipients minus the amounts received from 
existing contributors. When the government's budget is calculated on the 
more appropriate accruals basis, these contributions are subtracted from V 
and added to the current-period receipts from the sale of bonds. 
Although the government does not actually sell an explicit bond to a 
member of its pension plan, it may nevertheless be thought of as selling 

an implicit bond. It is not enough to simply remove the net cash flow 

arising from pension plan operations from V. A further reduction in V 
(with a conseguential increase in the value of implicit bonds outstanding) 
has to be made to allow for any true (accruals base) deficit in the pension 
program. With this discussion in mind we may now reinterpret eguation 
(1) as an accruals eguation. 

Converting from a cash-flow to an accruals basis does not affect 
purchases of goods and services or the creation of money (the first and 
last two items in the equation). It does, however, involve juggling items 
between taxes net of transfers and bonds outstanding. A correct calcu¬ 
lation, one that allowed for the actual features of existing pension plan 
arrangements, would lower V and increase both the level of and growth 
rate in the stock of bonds outstanding to include the implicit bonds that, 
like actual bonds, represent claims on the government's future resources. 
Respecting this accruals-base reinterpretation of equation (1), care has to 
be taken to note that neither V nor the conventionally reported bond- 
market data represent the accruals measures which are the appropriate 
ones. 

The next adjustment to be made to the government accounts involves 
deflating to allow for the effects of inflation. Instead of being considered 
in current dollars, the government's account may be considered in constant 
dollars by dividing each item in the above equation by the general price 


155 





level (the GDP deflator). Denoting the real values of variables (nominal 
values divided by the GDP deflator) by the corresponding lower case letter 
and defining 1 + r t = (l+R t )/(l+7i t ), where n is the rale of inflation, gives 4 


9t' v t + I+r^ b t ¥t-l) - - n t m t = °' 


( 2 ) 


The first two terms in (2) are real government expenditure on goods and 
services and real tax receipts (net of transfers), respectively. The third 
term is best interpreted in two parts: r^ is the real interest rate, and 
b^/O+rp is the current real value of the stock of outstanding bonds. The 
next two composite terms represent the growth in the stock of real bonds 
(b t -b t _^) and of real money balances (m^-m^). The final term repre¬ 
sents the inflation tax - the real revenue accruing to the government as a 
result of the decline in the outstanding real stock of its monetary 
liabilities. 

Various alternative definitions of the government's deficit may now be 
obtained by, in effect, simply rearranging equation (2). The first and 
narrowest definition of the deficit is 


d it H s t‘ v f 


(3a) 


This is the real deficit exclusive of debt interest. It is the magnitude 
focused on in several papers, notably those by McCallum (1982) and by 
Sargent and Wallace (1981a). It is equivalent, of course, to 


d it = (b t - b t-i + m t • m t-i ) + n t m t 


1+m 


(3b) 


What equation (3b) says is that the deficit (excluding debt interest) has to 
be financed by the growth in the stock of real bonds and real money 
balances outstanding (the first term in the equation), and by the inflation 
tax (the second term in the equation). The final term in equation (3b), is 
simply the interest payments on the existing stock of debt that have to be 
subtracted from the growth of liabilities and inflation tax. 

A second fairly natural definition of the deficit is one that includes 
real interest on the real debt outstanding. That is, 


156 







(4a) 


d 2t E 9 t - v t + b f 

This, of course, is equivalent to 

d 2t = (b t - b t-l + m t ' “t-P + W < 4b > 

What this says is that the deficit inclusive of real interest payments has to 
be covered by the inflation tax and the growth in the real value of the 
government's bond and money liabilities. 

A third definition recognizes that the government's real monetary 
liabilities are falling as a result of inflation and subtracts that fall in the 
real value of money from the above deficit to give 



v t + 



w 


This, of course, is equivalent to 


(5a) 


d 3t = b t ' Vi + m t ' m t-r (5b) 

Thus, this definition of the deficit amounts to simply the change in the 
real value of the government liabilities between t-1 and t. 

A further definition is suggested by the famous Ricardian equivalence 
proposition. If government bonds are perceived to generate a future tax 
liability with the same present value as the market value of the bonds, 
then the above discussion concerning the distinction between a cash- 
receipts and accruals definition of the government's account is seriously 
incomplete. While a bond sale does represent a cash receipt, it does not, 
under the conditions of Ricardian equivalence, generate a revenue on an 
accrual basis. A definition of the deficit that takes this distinction into 
account is 


r t 

d 4t E 9 t - v t + b t' Vt' b t + b t-i 
or, equivalently, 

d 4t = m t - m t-r 


(6a) 


(6b) 


157 





What this says is that if bonds and future tax liabilities are equivalent to 
each other, then bonds are not part of private sector wealth and any 
change in the stock of bonds outstanding represents a change in neither 
private sector wealth nor government liability. It is only the change in 
real money balances (equation (6b)) that represents a change in private 
sector wealth. 

Each of the above alternative definitions of the deficit takes opportun¬ 
ity cost notions into account. These definitions do not, however, pay any 
attention to the macroeconomic arguments concerning the need for cyclical 
adjustment of the deficit. A cyclically adjusted deficit ('full-employment' 
deficit) could be defined for each of the four alternative deficit definitions 
presented above. It is simplest, however, to examine the full-employment 
version of the deficit in connection with the d^ definition. 

The basic idea behind the full-employment adjustment is that govern¬ 
ment expenditures and tax receipts (g^-v or / equivalently, d^) will 
reflect both the programs and tax regulations in place as well as the 
current state of the economy. We could represent this idea as 



(7) 


Here d t denotes the level of the deficit that is independent of the current 
state of the economy; y represents the state of the economy at the 
current moment in time; reflects the effects of the existing tax and 

spending regulations on the way in which the current state of the economy 
influences the deficit. In principle, a and y are vectors of high dimen¬ 
sionality. It is convenient, however, to regard y as simply real GDP and 
a as the effects of real GDP on the deficit. Using this simplification, we 
could define the full-employment deficit as the deficit that occurs when 


income is at its full-employment level. Calling y^ full-employment real 
income at time t, the full-employment deficit then becomes 



( 8 ) 


What this equation tells us is that the deficit is equal to its full-employ¬ 
ment counterpart plus an adjustment for the deviation of income from full 
employment. Each of the other definitions of the deficit have a full- 
employment equivalent that involves precisely the same adjustment - the 


subtraction of a. (y -y ) from the definition given above. 


t w t x t 


158 


Thus, we now have eight alternative definitions of the deficit (actual 
and cyclically adjusted) for four definitions of varying breadth. Which, if 
any, of these deficit measures are interesting from a macroeconomic per¬ 
spective? The answer to this question clearly depends on the underlying 
question of substance regarding macroeconomic performance that one seeks 
to address. Some possible questions of substance are: 

1. Is the deficit stimulating or retarding aggregate demand? 

2. Is the deficit raising real interest rates? 

3. Is the deficit raising or lowering output and employment (relative to 
some 'full-employment' levels)? 

4. Is the deficit inflationary? 

Note that questions (1) and (3) are not the same: aggregate demand 
may or may not be equivalent to actual output. 

None of these questions can be answered solely by reference to the 
government accounts (somehow arranged). All need to be addressed 
through analysis of the behaviour of both the private sector and the 
government. In the pursuit of such analyses, some definitions of the 
deficit will turn out to be useful; some will not. To see this, however, it 
is necessary to address the specific questions posed above. 

THE DEFICIT AND AGGREGATE DEMAND 

Is fiscal policy, as reflected in the deficit, stimulating or retarding aggre¬ 
gate demand? This is a question of interest to Keynesians and non- 
Keynesians alike. The Keynesian interest stems from the direct link 
between aggregate demand and actual output and employment (as well as 
unemployment) in the simplest Keynesian frameworks. The non-Keynesian 
is interested in the question because of the potential effects of aggregate 
demand on both real income and the price level (or, more generally, the 
inflation rate). 

Most economists would agree that any analysis of the effects of the 
deficit on aggregate demand must start from a theory of aggregate de¬ 
mand. A widely accepted general proposition is that aggregate demand 
increases with the quantity of real money balances, the stock of real 
government bonds outstanding, the expected rate of inflation, and the 
level of government purchases of goods and services. Note that the level 


159 



of government expenditure, as well as the deficit and the way in which it 
is financed, is crucial to the level of aggregate demand. The composition 
of public expenditure is also of vital importance. The implications for 
aggregate demand of expenditure on profitable (revenue creating) activities 
are clearly different from those of pure waste. For present purposes, 
however, I shall treat both the level and the composition of public expend¬ 
iture as given. I shall also ignore considerations such as time lags and 
foreign influences on demand. 

For a given expected rate of inflation and a given level of government 
expenditures on goods and services, aggregate demand will change with 
changes in real money balances and the real stock of government bonds 
outstanding. There is no disagreement concerning the role of real 
balances. However, Ricardian economists would hypothesize that the 
effects of government bonds on aggregate demand are unimportant because 
of the offsetting change in future tax liabilities, and most economists would 
agree that bonds have less effect on aggregate demand than do real money 
balances. 

If the level of aggregate demand depends upon the levels of real 
money and real bond holdings, then which of the alternative deficit defin¬ 
itions are needed in order to understand the effects of the deficit on 
aggregate demand? The answer is immediate and direct. It is a combin¬ 
ation of definitions d^ and d^, given above. If the economy really is 
Ricardian, then d^ is the appropriate definition. If bonds and money have 
equally important effects on aggregate demand (an extreme position to 
which hardly anyone would subscribe), then definition d^ is appropriate. 
In general, a weighted average of d^ and d^ is required; the weighting 
will reflect the degree to which bonds constitute net worth. 

It is worth emphasizing that it is the deficit as measured by a com¬ 
bination of d^ and d^ on an accruals rather than a cash-receipts basis that 
is appropriate. Changes in individuals' perceived wealth through their 
participation in public sector pension schemes are just as important as any 
other perceived wealth changes. Of course, if the world is Ricardian, 
then this consideration loses its force. 

We have seen, then, that definitions of the deficit d^ and d^ are the 
appropriate ones for telling us about the effects of the deficit on changes 
in aggregate demand. However, changes in real money balances and real 
bonds outstanding are induced in part by changes in the current level of 
real income. This suggests the need for a measure of deficit that is inde- 


160 




pendent of the actual level of income. The full-employment version of 
deficits d^ and d^ are, therefore, suggested as the appropriate ones. 
Which particular level of employment is selected as the full-employment 
level is a matter of secondary importance. All that is needed is a constant 
level of real income, so that the effects of changes in income on induced 
changes in the growth rates of real money and real bonds outstanding are 
removed.. 

It is important to note that the level of the deficit yielded by this 
calculation tells little in and of itself. It is the changes in the deficit 
yielded by this calculation that provide information about the direction and 
magnitude of the effects of fiscal policy on the growth rate of aggregate 
demand. 

It is noteworthy that the measure of the deficit required to indicate 
the effects of the deficit on aggregate demand, the full-employment version 
of either d^ or d^, is very close to the adjusted deficit calculated by 
Bossons and Dungan (1983). There are, however, two features of their 
work that make their calculations and inferences slightly different from 
those reached above. First, their calculations of the deficit ignore the 
government's pension activities - which means, of course, that they fail to 
convert the pension program from a cash-receipts base to an accruals 
base. The second difference has to do with levels versus changes in 
levels. It was remarked above that it is changes in the full-employment 
deficit (d^, d^ definition) that provide information about changes in the 
growth rate of aggregate demand induced by active fiscal policy. Bossons 
and Dungan seek to draw inferences about aggregate demand from the 
level of the full-employment surplus. Of course, to say that a change in 
the deficit (appropriately defined) induces a change in the growth rate of 
aggregate demand is equivalent to saying that the level of the deficit 
induces a change in the level of aggregate demand. There is, however, a 
crucial difference between these two propositions, at least as empirically 
implemented by Bossons and Dungan. In order to make statements about 
the effects of a change in the deficit on the change in the growth rate of 
aggregate demand, one does not need to be sure that one is using the 
correct level of unemployment in calculating the full-employment deficit. 
In order to make statements about changes in the level of aggregate de¬ 
mand on the basis of statements about the level of the deficit, one does 
need to know that the full-employment level used in the calculated adjust¬ 
ments is the correct one. 


161 


















Those who advocate fiscal stimulation in current conditions assume 
that any rise in aggregate demand would have a sufficiently large output 
effect and a sufficiently small price effect for such stimulation to be worth¬ 
while. The possible effects of the increased deficit (or decreased surplus) 
on the price level and on aggregate supply are either ignored or down¬ 
played. I shall return to this issue later in the paper. 

THE DEFICIT AND REAL INTEREST RATES 

There are certain situations (some actual and some hypothetical) in which 
the real rate of interest is a parameter and not subject to any influence 
from either the government's fiscal policy or any other domestic source. 
Such a situation occurs, for example, when a small, open economy faces 
effectively perfectly elastic supplies of capital from the world capital 
market. Canada, in current conditions, probably fairly well approximates 
such an economy. Another situation of this type is one in which bonds 
are not treated by their holders as part of net wealth - when they are 
perfect substitutes for future tax liabilities. A third situation is where 
there is a constant rate of time preference that bolts down the equilibrium 
real rate of interest (in stationary equilibrium) and forces adjustments of 
the capital stock to ensure that the marginal product of capital equals that 
parametrically given rate of time preference. In all these cases, it is 
evident that, regardless of how it is defined and measured, the deficit has 
no effect on the real rate of interest. 

In other situations, when the rate of interest is an endogenous vari¬ 
able responding to current domestic economic conditions, it is of some 
importance to enquire how it is affected by the deficit (on some appro¬ 
priate definition). Although this question may be of limited practical 
importance for Canada, it is clearly of importance for the world economy as 
a whole and for large economies such as the United States. 

A moment's reflection will show that the question is almost the same 
as the question of the deficit's influence on aggregate demand, for the 
level of aggregate demand and the level of the real rate of interest are 
determined simultaneously by goods market and money market equilibrium. 
Except in the special cases noted above, and ruling out perverse cases, 
the standard reduced form predictions arising from a variety of alternative 
structures would have the real rate of interest fall as the quantity of real 
money balances increased and rise as the stock of real bonds increased. 


162 


If we suppose that we are dealing with a situation in which the expected 
(and actual) rate of inflation is constant, so that real balances are also 
constant, the only remaining influence (stemming from fiscal policy) on the 
real rate of interest is the size of the real stock of government bonds 
outstanding. If that stock is rising (in per capita terms), then, except in 
the special cases that we are not considering here, the real rate of 
interest will also rise. Such a rise in the real rate of interest will induce 
intertemporal substitution in consumption of both goods and leisure. It 
will also induce a smaller equilibrium capital stock and, temporarily, a 
slower rate of accumulation. 

Which is the appropriate measure of the deficit for judging the influ¬ 
ence of fiscal policy on the real interest rate? The answer appears to be 
definition d^. Given a constant inflation rate and real balances constant in 
equilibrium, that measure of the deficit would be a direct measure of the 
increase in the stock of real government bonds outstanding. If we are 
interested in removing any effects stemming from changes in real bonds 
outstanding induced by the state of the economy, it will be the full- 
employment version of d^ that provides an indication of the qualitative 
effects of the deficit on the real rate of interest. 

The importance of these considerations for Canada taken alone is 
probably negligible. (See Ronald Wirick's paper later in this volume.) 

THE DEFICIT AND THE LEVEL OF REAL ECONOMIC ACTIVITY 

The level of real economic activity may be viewed as being determined by 
the interaction of aggregate supply and aggregate demand. Even the most 
eclectic introductory textbooks in economics now adopt this approach. 
(See, for example, Lipsey, et al. (1982).) In this view, if aggregate 
demand and aggregate supply both rise, but aggregate demand rises more, 
then income will rise and the price level will rise. If both rise, but 
supply rises by more than demand, then income will fall in the face of 
rising prices - stagflation. If both aggregate demand and aggregate 
supply fall, but demand falls by less than supply, then income will rise 
but the price level will fall. Finally, if both fall, but if aggregate demand 
falls by more than aggregate supply, then both income and the price level 
will fall. Therefore, to determine the effects of the deficit on the level of 
real economic activity, it is necessary to examine its effects on both aggre¬ 
gate supply and aggregate demand. I have already discussed its effects 


163 


on aggregate demand. I shall now focus on aggregate supply. 

The most general approach to an analysis of aggregate supply is to 
suppose that the amount that will be supplied at any given point in time 
depends on the price level prevailing at that point relative to the expec¬ 
tations of the price level formed at various dates in the past. In a special 
case - the so-called New Classical version of the theory of aggregate 
supply - it is only the most recently formed expectation of the current 
price level that matters. 5 In the alternative so-called New Keynesian or 
long-term contract approach, what matters are the expectations of the 
current price level that were formed on the dates when all the labour 
market (and other input market) contracts still in force at the present time 
were entered into. 6 To the extent that current changes in aggregate 
demand were anticipated on those past dates, such changes will have no 
effects upon real economic activity and will have all their effects on the 
price level. However, to the extent that such changes in aggregate de¬ 
mand were not anticipated, they will have real effects. Thus, the crucial 
thing for output determination is not the level of aggregate demand, but 
rather its level relative to the anticipated level (where the anticipations in 
general have been formed at various dates in the past). It immediately 
follows that it is not the level of the deficit that is important for the 
determination of the level of real economic activity, but rather the pre¬ 
dictability of the deficit. If the deficit is currently at a level that was 
anticipated in the past, then its effects will come through aggregate de¬ 
mand to influence the price level but not the level of output and employ¬ 
ment. Therefore, regardless of how one measures the deficit, it is not 
possible, by simply examining the deficit and its effects upon the level of 
aggregate demand, to reach any conclusions concerning the effects of the 
deficit on the levels of employment and unemployment. 

These considerations have important implications for the policy advice 
that is often given concerning the deficit. Those economists who regard 
the current state of the deficit as being contractionary and who therefore 
advocate fiscal stimulus need to ask how it will be possible to generate 
fiscal stimulus that will not have been reasonably well anticipated by the 
time it actually occurs. In view of the lags between decision, legislation, 
and implementation associated with fiscal policy, it is hard to believe that 
much of what happens on the fiscal policy front can be regarded as unan¬ 
ticipated. 

The implication of the above remarks is that no particular measure of 


164 



the deficit has any major role to play in indicating what determines the 
level of economic activity. It is only the current level of the deficit 
relative to its previously expected (cyclically adjusted) level that is of 
concern. 

There is, of course, an alternative view, which holds that the price 
level (and inflation rate) are largely (though not entirely) independent of 
aggregate demand. If this is the case, variations in aggregate demand will 
indeed be associated with variations in actual output and employment. 
Econometric models can be built that both fit the facts and incorporate this 
view of the world. It may, therefore, justifiably be asked how I can 
spend so much time on the present topic and yet say nothing about this 
alternative (Keynesian) approach. This is not a suitable place to engage 
in a lengthy appraisal of the Keynesian tradition. Nevertheless, there do 
seem to be cogent reasons for doubting the ability of the alternative 
approach to offer more than a description - rather than an understanding 
and an explanation - of the facts. 7 

THE DEFICIT AND INFLATION 

'Is the deficit inflationary?' is an often-asked and clearly important ques¬ 
tion. Equally clearly, it is a question that cannot be answered with 
reference to either the d^ or d^ definitions of the deficit that I have 
identified as influencing aggregate demand and real interest rates, since 
these measures in no way depend on the rate of inflation. They are 

entirely real definitions. They could occur at any rate of inflation. 

What meaning can we attach to the concept of an inflationary deficit? 
Do any of the other definitions posed above help? In particular, what 

about d^ and d£? Each of these definitions contains the inflation rate and 
each says that the deficit increases with the rate of inflation (with real 
money balances being a factor of proportionality linking the two), other 
things being equal. However, each definition also indicates that there are 
sufficient degrees of freedom for the perpetual issue of bonds on an 

ever-increasing scale to sustain any inflation rate regardless of the value 
of the deficit (whether measured on the basis of the d^ or the 

definition). Thus, neither of these two definitions of the deficit provides 
a direct measure of its inflationary orientation. 

In reaching this conclusion, it was necessary to say that any inflation 
performance could be achieved provided the future path of bonds was 


165 






appropriate. This suggests that in trying to attach meaning to the con¬ 
cept of an inflationary deficit we need to pay attention not to some exist¬ 
ing state of affairs at one point in time, but to a process over time. 
Thus/ while there is no meaning to the notion of a state of inflationary 
deficit, there may be some meaning to the notion of an inflationary deficit 
process . This is the way in which several scholars with widely different 
viewpoints have sought to analyze the bond-financing of ongoing deficits. 8 

In an analysis of ongoing-inflation, no explicit attention need be paid 
to fluctuations in income and employment around their average levels. 
Consequently, it is a full-employment variant of the deficit definitions 
introduced above that needs to be considered. Some care is required, 
however, in the definition of full employment - a matter I shall return to 
later in this section. For now, let us assume that the full-employment 
adjustment is correct. The starting point for the analysis is a full- 
employment statement of the government's real accounts, i.e., 

* r t 

d it + rr?/ b t - (b t ' b t-i + m t - m t-i> ' n t m t = °- (9) 

* 

The variable d^ is the full-employment level of government expenditures 
on goods and services less taxes (net of transfers), expressed in real 
terms. This variable is treated as exogenous in the analyses of Sargent 
and Wallace (1981a) and McCallum (1982). 9 

* 

The government is presumed to choose a path for d it that is indepen¬ 
dent of the real rate of interest, the levels of real bonds and money 

* 

balances outstanding, and the inflation rate. (For McCallum, d^ is deter¬ 
ministic, whereas for Sargent and Wallace it is stochastic.) Note that the 

* 

assumption that d^ is an exogenous process is restrictive. It ignores any 
feedback effects from interest rates or inflation to the deficit. Neverthe¬ 
less, it is a useful starting point. It shows what would happen if govern¬ 
ments were totally unresponsive to any havoc they might create. It also 
perhaps shows tendencies that would emerge if governments were slow to 
respond to a deteriorating situation. For simplicity, the real rate of 
interest is treated as independent of the government's budget. This is 
probably a reasonable assumption for a small, open economy operating in 
the neighbourhood of solvency, but probably not otherwise. 10 

The rate of inflation is related to the level of real balances by way of 
a demand for money function and money market equilibrium. Thus there is 


166 





a given (inverse) relationship between real balances and the rate of infla¬ 
tion, the latter being determined by the rate of growth of the nominal 
money supply. If the government's budget equation (equation (9) above) 
is satisfied, if the real full-employment deficit is exogenous, if the real 
rate of interest is given, and if the inflation rate is linked to the level of 
real balances by a demand for money function, then a rule for the growth 
rate of the money supply implies a path for the stock of real bonds. The 
question that can now be posed is this: Can a rule for the money supply 

be specified that delivers a desired rate of inflation regardless of the 
* 

behaviour of d^? Or rather, is it the case that the entire sequence of 

deficits {d^} may be such as to prevent a desired inflation path under any 

feasible rule for the growth rate of the money supply? Notice that the 

* 

deficit concept relevant to this question is dj t viewed as a sequence or 

process rather than as a single-point-in-time event. 

Scarth (1980) and Sargent and Wallace (1981a) show that a Friedman 

(1959) k-percent rule for the money supply growth rate cannot work 

unless the deficit path is a deterministic one and of the 'right' size. They 

further show, however, that a Friedman (1948) 'monetize-the-deficit' rule 

can work, but that the deficit in that case determines the growth rate of 

the money supply and inflation. In that event, then, the rate of inflation 

* 

depends directly upon the deficit sequence d^. Sargent and Wallace go on 
to show - and this is the 'unpleasant' aspect of their results - that if, in 
the face of a given exogenous deficit, an attempt is made to control infla¬ 
tion by slowing down the growth rate of the money supply, the effect will 
be a rise in the rate of inflation. 

It is easy to understand why these results arise. The attempt to 
pursue a fixed money supply growth rule with a random deficit requires 
that the path of bonds be random. This requirement, however, induces 
additional noise in the requirement to service the bond stock. Interest 
has to be paid on a random stock of bonds, which is added to the noise 
generated by the deficit itself. The result is a stochastic process that has 
no equilibrium. An attempt to slow down inflation by lowering the growth 
rate of the money supply simply involves issuing more bonds at the 
present time. If more bonds are issued, then more interest must be paid 
on them, and this implies that at some later stage more money must be 
printed to retire those bonds and to pay the interest that has been 
incurred by them. 

Thus, tight money now implies more inflationary monetary growth 


167 





later. This in turn induces a higher inflation rate now by a very simple 
and natural mechanism. Expectations of more rapid money growth later 
imply expectations of more inflation later. This leads to an expectation of 
a drop in the demand for money (to avoid the inflationary losses), which 
in turn has the same effects on inflation as a rise in the growth rate of 
the supply of money. Avoiding the higher opportunity costs of holding 
money, when the money supply growth rate has finally increased, involves 
lowering the demand for money before the money supply growth rate 
increases. This in turn increases the inflation rate before the inflationary 
rise in the growth rate of the money supply occurs. This process (con¬ 
ceptual process) has to unwind all the way back to the present, for only 
by reducing holdings of money balances now can the consequences of 
future inflation be avoided. Thus, lowering money balance holdings now 
has the same effect now on the inflation rate as a rise in the growth rate 
of the money supply now would have. 

McCallum (1982) conducts an analysis similar to that of Scarth (1980) 
and Sargent and Wallace (1981) but uses a deterministic exogenous deficit 
path. He shows that in this case (implicitly in agreement with earlier 
findings) a Friedman k-percent rule can be followed provided that the 
deficit is not 'too large.' What 'too large' means in this context is that the 
growth rate of bonds implied by the deficit and the k-percent rule for the 
money supply must not exceed the sum of the real rate of growth of the 
economy and the rate of time preference. Clearly, the stock of bonds 
outstanding can grow at the growth rate of the economy without any 
consequence. They can, however, grow faster than the economy provided 
there is a finite present value to the entire future path of bonds out¬ 
standing. 

None of the foregoing analyses make predictions about the world in 
which we live. Rather they generate predictions that are the logical 
consequences of following particular rules for deficit financing proposed by 
prominent economists and of following those rules in the face of an exo¬ 
genous (unyielding) deficit process. It may be of greater interest than 
analyzing the consequences of rules advocated by economists to analyze the 
consequences of rules that would in fact be followed by rent-maximizing 
politicians and central bankers. 

Barro (1983), Barro and Gordon (1982), and Blinder (1982) have 
developed suggestive theoretical models in this direction, but no definitive 
results are available. Clearly, there is need for both work on the positive 


168 


theory of government and central bank behaviour and for statistical inves- 

* 

tigations of the behaviour of d^. The latter’s exogeneity, or its links 
with other economic variables, and the relationship between the money 
creation process and the deficit thus defined are matters requiring urgent 
attention. 

The preceding discussion, paid no attention to fluctuations in the 
deficit induced by fluctuations in the level of economic activity. Clearly 
such fluctuations do occur, so the definition of the deficit employed has to 
be one that is controlled for such fluctuations. Does the 'full-employ¬ 
ment' - perhaps more fittingly called the 'high employment' - adjustment 
already discussed do the required trick? There are good reasons for 
believing that it does not. In order to analyze the influence of the deficit 
on inflation, we clearly need to work with a definition of the deficit that is 
normalized for the average sustainable level of real economic activity. Is 
the average sustainable level the same as the full-employment (or high- 
employment) level? The answer is almost assuredly that it is not. There 
are two reasons for this answer, both of them probably important to the 
current Canadian situation. 

First, there is an important difference between the natural or full- 

employment level of employment and the average level of employment that is 

compatible with a fixed average rate of inflation. This difference arises 

from the (universally accepted) fact of a non-linear relationship between 

inflation and unemployment. 11 The point is most easily made with the help 

of a diagram. Figure 1 shows a conventional non-linear Phillips curve 

(labelled P). The inflation rate 71* is the desired or target inflation rate, 

and u* is the natural rate of unemployment. The vertical line above u* 

can be thought of as the long-run Phillips curve. If it were possible to 

run the economy bang on the natural rate of unemployment at all times, 

then the target inflation rate n* could be delivered. If, alternatively, 

* 

there are fluctuations in activity about u*, then u cannot be the average 
rate of unemployment unless inflation is continuously accelerating. To see 
this, consider a special case. 

Imagine that the economy bounces between a high unemployment rate 
(Uj_j) and a low unemployment rate (u^) and spends half the time at each. 
In that case, the average unemployment rate will indeed be u*. What will 
the average inflation rate be? Clearly, it will be the average of (asso¬ 
ciated with u u ) and (associated with u T ). That average inflation rate 
H. hi -L 

is found where a cord joining the two relevant points on the non-linear 


169 




Figure 1 

The Phillips curve and the sustainable average unemployment rate 


Inflation 



Phillips curve intersects the long-run or vertical Phillips curve above u*. 
Thus the average inflation rate is n^. Evidently, is greater than n*. 
If the economy operated at the average unemployment rate of u*, clearly 
the inflation rate, 71*, could not be the expected inflation rate, so the 
short-run Phillips curve labelled P could not be the relevant short-run 
Phillips curve. In fact, no equilibrium steady-state inflation rate exists in 
this situation. However, there obviously does exist an average rate of 
unemployment greater than u* but less than u^ that would deliver n* as 
the average rate of inflation. That is, if the range of unemployment 
between u^ and u^ were held constant but the absolute levels of unem¬ 
ployment increased, it would be possible to find a combination of u^, and 
Upj, averaging between u^ and u* that delivered 71* as the average infla¬ 
tion rate. The more the Phillips curve is curved and the greater is the 
range of variability in unemployment, the higher is the average sustainable 
unemployment rate relative to the natural unemployment rate. In adjusting 
the measured budget deficit for the effects of unemployment, it is the 
average sustainable unemployment rate and not the natural rate of unem- 


170 










ployment that should be used. 

The second set of factors that need to be taken into account in 
calculating the full-employment deficit consists of the structural and other 
non-cyclical factors that affect the natural rate of unemployment and that 
may operate for relatively long periods to keep the natural rate of un¬ 
employment at historically high levels. For the past several years, the 
natural rate of unemployment has been unusually high and it will probably 
remain high for several years to come. The natural rate of unemployment 
is of course nothing other than the rate of unemployment that reflects 
turnover in the labour market when all expectations concerning nominal 
magnitudes are fulfilled. As such, the natural rate of unemployment 
depends crucially upon the amount of resource reallocation (and therefore 
labour force reallocation) taking place in the economy. In recent years, 
the amount of reallocation in most industrial economies and certainly in the 
Canadian economy has been unusually large. Unusually big changes in 
relative prices associated with (but not exclusive to) the energy sector 
and with unusually large movements in real rates of interest have drawn 
resources away from capital- and durables-oriented industries and 
energy-intensive technologies towards other sectors of the economy. In 
the face of such large-scale reallocation, it is inevitable that the natural 
rate of unemployment should register a rise. Measures of the natural rate 
of unemployment for the United States based on an explicit measure of the 
amount of sectoral reallocation taking place suggest that much of the 
fluctuation in unemployment in that economy can be accounted for by these 
forces and that, in recent years, the amount of reallocation has been 
unusually large (see Lillien 1982). 

Taken together, the two considerations described above indicate that 
a great deal of caution must be exercised in adjusting the budget deficit 
for the level of employment. 12 It is clearly inappropriate to calculate 
unemployment-adjusted deficits for unemployment rates in the neighbour¬ 
hood of 5 or 6 per cent. Bossons and Dungan (1983), in their calcula¬ 
tions, accept that judgment. They use unemployment rates between 7 and 
1\ per cent (with 8.2 per cent for 1982). Perhaps such values seem high 
for the no-recession unemployment rate. However, when one takes account 
of the factors described above, it seems likely that even these numbers 
are too low. Unemployment rates closer to 10 per cent, at least for the 
foreseeable future, appear to be more appropriate given the current possi¬ 
bilities concerning the sustainable constant-inflation average unemployment 


171 




rate. 


I have suggested that the entire sequence of deficits as measured by 
the real deficit exclusive of debt interest (d^), adjusted appropriately for 
the average sustainable unemployment rate level (but with a smaller adjust¬ 
ment than that performed by Bossons and Dungan), is the appropriate 
concept of the deficit for judging the inflationary stance of fiscal policy. 
If this is correct, then we still have some potential for concern about the 
deficit. There appears to be a political process in place that is generating 
ever-increasing expenditures and ever-growing deficits - even on an 
employment-adjusted basis. Consequently, there is a substantial inflation 
potential built into the deficit process. 13 

It is also worth emphasizing that viewing the deficit as the outcome of 
an ongoing political process makes the cash-flow basis and not the accruals 
basis definition of the deficit the relevant one. The current value of the 
accruals-basis deficit may contain information about the future cash-flow 
deficit, but it is the sequence of cash-flows that has to be financed. 

Although the deficit looks serious when it is viewed as an exogenous 
process, a very different conclusion emerges if the deficit is seen as 
responsive to monetary policy constraints. Specifically, as was shown by 
Sargent and Wallace (1981), if monetary policy can be placed on a 
k-percent growth rule and if the deficit itself can become an endogenous 
variable responsive to the constraints imposed on it by the growth rate of 
the money supply, then all will be well. Thus the real issue is not the 
size of the deficit but whether it is exogenous or endogenous. If the 
deficit is exogenous, with money supply growth reacting to it, we are, on 
current trends, in for substantially more inflation in the future. If the 
money supply growth process is or can be made exogenous and fiscal 
policy endogenous, then neither the size of the deficit at any given point 
in time nor, indeed, its historical path is relevant to the determination of 
the future course of inflation. 

Whether the deficit is exogenous or is responsive to money growth, 
real interest rates, inflation, and a variety of other possible influences is, 
of course, an empirical question. Unfortunately, it is a question to which 
we do not know the answer. Obtaining that answer and, in particular, 
establishing whether or not the existing tax and spending programs and 
their built-in dynamic paths are consistent with returning us, eventually, 
to an acceptable level of steady-state inflation is an urgent and difficult 
research objective. 


172 


SUMMARY AND CONCLUSIONS 


I have developed various alternative definitions of the deficit and analyzed 
their role in answering four questions of concern for macroeconomic policy. 
We have discovered that the full-employment deficit, defined to include the 
real interest payment on the government debt but excluding the inflation 
tax, is the appropriate concept of the deficit for studying the effects of 
fiscal policy on aggregate demand and (with qualifications) on the real rate 
of interest at a given price level. I have also suggested that aggregate 
demand as such has no simple effect on the level of real economic activity. 
Rather it is the level of aggregate demand relative to the level anticipated 
when existing wages and prices were determined that affects the level of 
actual output and employment. The concept of the deficit that helps us 
understand the determinants of aggregate demand is, therefore, of limited 
value in helping us understand what determines the actual level of econo¬ 
mic activity. Furthermore, that measure of the deficit tells us nothing 
about the inflationary stance of fiscal policy, since it is a measure that is 
independent of inflation. 

For understanding the effects of the deficit on inflation, the relevant 
deficit concept is one that focuses on the deficit process rather than the 
deficit state . The central issue is not the size of the deficit but whether 
it is exogenous or endogenous. If the deficit, measured to exclude fluctu¬ 
ations of the level of economic activity from its average sustainable level 
(and not from its full-employment level), is indeed exogenous, then it will 
dominate inflation and the relevant deficit measure is the difference 
between real government expenditures on goods and services and transfers 
net of taxes (appropriately adjusted for average sustainable activity 
levels). If, on the other hand, monetary growth is exogenous and the 
deficit ultimately responds to that exogenous monetary policy, then no 
matter what the deficit is at any given moment it is monetary policy that 
dominates the inflation path. We know too little about these processes. 
Clearly we need much more research, both theoretical and empirical, on 
government policymaking and its interactions with private economic be¬ 
haviour. 

NOTES 

1 A similarly simple formulation could be achieved by assuming that all 


173 






the bonds are perpetuities. In fact, the present average term of 
maturity for Canadian government bonds is about six years. Nothing of 
substance turns on this simplification. 

2 A more direct way of writing (1) is 


G t' 


h + \-i 


1+R. 


B t + 


M 


t-1 


- = 0 


( 1 ' ) 


The equation in the text is (1') with B added and subtracted. This 
has the advantage (useful later) of directly displaying the (implicit) 
interest on the government debt. In (1'), there is no distinction 
between interest payments and bond redemptions. The amount (B ) is 
paid out on last period's bonds, and the amount (B^) that will be paid 
out next period has generated a revenue this period of (B^/Q+R^)). 

3 I am grateful to Brian Scarfe for pointing out my omission of open- 
economy considerations and to William Scarth for his kindness in 
providing me with an efficient and thorough guided tour through the 
literature on this topic. For a much fuller discussion of this and 
related issues, the reader is referred to Scarth (1975). Scarth's 
open-economy equation (7) in that paper (p. 11) is equivalent to my 
equation (1) above. Scarth's bonds are perpetuities, however, and he 
models tax collections (my V) endogenously. 


4 The steps from equation (1) to equation (2) are as follows: First, 
define the GDP deflator at t as P . Then divide equation (1) by P^ to 
give the government's budget in constant dollars as 


V R t B t A 

p t (i+R t ) r t l P t 


B t-i M t 

_L_L) _ (_P 

p J 

t t 


M 


t-1. 


= 0 


(a) 


Next, multiply and divide the last and third last terms by P^ use 
the fact that P -/P = 1/(1 + 7T ) and let lower case letters denote 
the real value of the relevant variable to give 


R. 


e - v + - 

g t t 1+R. 


b t-l Vl 

b t - (b t - a^o > - (m t - (TWJ ) - o 


(b) 


Next, use the definition (1+R ) = (1+r )(l+n ) and multiply (b) by 
(1+71^) to give 

(l+n t )(g t -v t ) + b t - ((l+7t t )b t - b t . 1 ) - ((l+7i t )m t -m t -l) = 0. (c) 

Noting that R -71 -r t =0 and treating t as a short interval so that 
7t^(g t “ v t ) vanishes gives 


174 










0 . 


(d) 


o -V + - 

t 1+r. 


b t - (V b t-i> 


(m t -m t _i) - 7t t m t = 


which is equation (2) in the text. 

5 See especially Lucas (1973) and Barro (1976). Although neither of 
these papers incorporates fiscal policy and therefore deficits, both 
could fairly readily be modified to do so. 

6 See especially Fischer (1977) and Taylor (1979). For a general dis¬ 
cussion of this and the new classical model, see Parkin (1982). 

7 For a fuller development of this theme see especially Parkin (1982). 

8 See, for example, Blinder and Solow (1973), Scarth (1976), and Christ 
(1979) for Keynesian analyses in which either the price level or the 
expected rate of inflation is given; Tobin and Buiter (1976) for both 
Keynesian and perfect foresight, full-employment analyses; and Scarth 
(1980), Sargent and Wallace (1981), and McCallum (1982) for perfect 
foresight or rational expectations analyses. 

9 In the Keynesian models referred to in the previous footnote, devi¬ 
ations of output from full employment are a crucial feature of the 
analysis and tax collections are endogenous. Indeed, if an equili¬ 
brium exists in these models, it is achieved by an equilibriating 
change in income. For present purposes, it is appropriate to focus on 
the full-employment studies already referred to. 

10 See Wirick (1983). 

11 This analysis was first suggested by and draws on an important but 
unfortunately unpublished paper by Gray and Lipsey (1974). 

12 An interesting additional cyclical adjustment may be necessary due to 
cyclical movements in the ratio of money to bonds. If that ratio 
rises in a recession, then the normal, cyclically corrected bond 
interest payments will exceed the actual payments. 

13 What I am saying here in general and abstract terms David Slater put 
in more detailed and graphic terms in his remarks from the floor at 
the conference. He noted (as I interpreted him) the pressures for 
increased government spending resulting from the effects of an aging 
population on pension revenues and outlays and on health programs, the 
pressures for ever better education programs, and demands by civil 
servants for easier budgets and yet better pay and conditions for 
themselves. 


REFERENCES 


Barro, R.J. (1976) 'Rational expectations and the role of monetary policy.' 

Journal of Monetary Economics 2, 1-32 
- (1983) 'Inflationary finance under discretion and rules.' Canadian 
Journal of Economics 16, 1-16 


175 







Barro, R.J. and D.B. Gordon (forthcoming) 'A positive theory of monetary 
policy in a natural-rate model.' Journal of Political Economy 
Blinder, A.S. (1982) 'On the monetization of deficits.' National Bureau 
of Economic Research Working Paper No. 1052 
Blinder, A.S. and R.M. Solow (1973) 'Does fiscal policy matter?' Journal 
of Public Economics 2, 319-37 

Bossons, J. and D.P. Dungan (1983) 'The government deficit: too high 
or too low?' Canadian Tax Journal 31, 9 
Christ, C.F. (1979) 'On fiscal and monetary policies and the government 
budget restraint.' American Economic Review 69, 526-38 
Fischer, S. (1977) 'Long-term contracts, rational expectations and the 
optima] money supply rule.' Journal of Political Economy 85, 191-206 
Friedman, M. (1948) 'A monetary and fiscal framework for economic stabi¬ 
lity.' American Economic Review 38, 245-64 

- (1959) A Program for Monetary Stability (New York: Fordham University 

Press) 

Gray, M.R. and R.G. Lipsey (1974) 'Is the natural rate of unemployment 
compatible with a steady rate of inflation?' Queen's University, Discus¬ 
sion Paper No. 147 

Lilien, D.M. (1982) 'Sectoral shifts and cyclical unemployment.' Journal 
of Political Economy 90, 777-93 

Lipsey, R.G., D.D. Purvis, G. Sparks and P.O. Steiner (1982) Econo ¬ 
mics , 4th ed. (New York: Harper and Row) 

Lucas, R.E., Jr. (1973) 'Some international evidence on output-inflation 
tradeoffs.' The American Economic Review 63, 326-34 
McCallum, B.T. (1982) 'Are bond-financed deficits inflationary? A Ricard¬ 
ian analysis.' National Bureau of Economic Research Working Paper 
905 

Parkin, Michael (1982) Modern Macroeconomics (Scarborough, Ont.: Pren¬ 
tice-Hall of Canada) 

- (1983) 'The inflation debate: an attempt to clear the air.' Forthcoming 

in Douglas D. Purvis, ed.. Proceedings of a John Deutsch Round 
Table . 

Sargent, T.J. and N. Wallace (1981a) 'Some unpleasant monetarist arith¬ 
metic.' Federal Reserve Bank of Minneapolis Quarterly Review 5, 1-17 

- (1981) 'The fight against inflation: how much can the fed do on its 

own?' Federal Reserve Bank of Minneapolis, mimeograph 


176 



















Scarth, W.M. (1976) 'A note on the "crowding out" effects of bond-finan¬ 
ced increases in government spending.' Journal of Public Economics 
- (1980) 'Rational expectations and the instability of bond-financing.' 
Economic Letters 6, 321-7 

Taylor, J.B. (1979) 'Staggered wage setting in a macro model.' American 
Economic Review, Papers and Proceedings 69, 108-13 
Tobin, J. and W. Buiter (1976) 'Long run effects of fiscal and monetary 
policy on aggregate demand.' in Jerome L. Stein, ed., Monetarism 
(New York: North-Holland) 

Wirick, R.G. (1983) 'Does the deficit crowd out private investment?' In the 
present volume 


Brian L. Scarfe* 

First, may I state that it is both a pleasure and a privilege to be invited 
to discuss Michael Parkin's fascinating and thought-provoking paper. 
Parkin demonstrates how a number of different measures of the real gov¬ 
ernment sector deficit are related and shows us how these individual 
measures need to be used in order to address four important macroecon¬ 
omic issues, namely the effect of the deficit on aggregate demand, real 
interest rates, output and employment, and the inflation rate. 

Parkin agrees with other economists that the cyclically adjusted deficit 
is a more appropriate concept to use in addressing these issues than the 
actual deficit, but he cautions us that there are serious dangers in over¬ 
correcting, and therefore understating, the size of the cyclically adjusted 
deficit. These dangers result from two circumstances: first, that over 
the business cycle the average level of unemployment that is consistent 
with a fixed or non-accelerating inflation rate exceeds the so-called natural 
rate of unemployment by an amount that depends not only upon the curva¬ 
ture of the underlying short-run inflation-unemployment trade-off function 
but also upon the amplitude of cyclical fluctuations; second, that the 
natural rate of unemployment may itself be underestimated in times that 
necessitate significant structural changes and resource reallocation in the 
economy. 

Although Parkin deals explicitly with a closed-economy framework, it 


Professor, Department of Economics, University of Alberta. 


177 








should be pointed out that in an open-economy case the important rela¬ 
tionship between the real government-sector deficit and the growth of the 
stocks of money and bonds that underlies the whole of his analysis would 
have to be altered to allow for the fact that surpluses in the overall bal¬ 
ance of international payments must generate growth in the stock of high- 
powered money or (to the degree in which the surplus is temporarily 
sterilized) in the stock of government bonds held by the private sector. 
Similarly, a deficit on the balance of payments must reduce the stock of 
high-powered money or (again to the degree in which the deficit is temp¬ 
orarily sterilized) the stock of government bonds held by the private 
sector. It is not clear, however, that allowing for the impact of these 
foreign sector transactions on an open economy would substantially alter 
Parkin's conclusions. Certainly, if these conclusions are correct, they 
would at least continue to apply to the analysis of the macroeconomic 
impacts of the U.S. federal deficit. 

Putting these preliminaries to one side, Parkin argues quite correctly 
that the most appropriate measure of the deficit for answering the question 
'What is the effect of the deficit on aggregate demand at a constant infla¬ 
tion rate?' is one that is equal to the change in the real value of govern¬ 
ment sector liabilities that would accrue at the cyclically adjusted level of 
output and employment. Notice that this measure of the deficit includes 
real interest payments on the government debt but excludes the inflation 
tax on money-holding (or the fall in the real value of money). To the 
extent that tax-discounting of the future real interest burden of bond- 
financed debt occurs, this measure is larger than the deficit-generated 
growth in real private-sector wealth. 

The impact of the real government sector deficit on aggregate demand 
and the real interest rate will depend upon the bond versus money compos¬ 
ition chosen when the deficit is financed. The higher the proportion of 
the real government sector deficit financed by bond-creation, the larger 
will be the upward pressure on the real rate of interest and the smaller 
will be the overall positive impact on aggregate demand at a constant 
inflation rate. The larger the proportion of the increase in the real value 
of government sector liabilities that is financed by new money-creation, the 
smaller will be the upward pressure on the real rate of interest and the 
larger will be the overall positive impact on aggregate demand. On these 
points, I believe, there would be little disagreement among economists. 

I turn now to the areas in which major disagreements may occur, 


178 


namely with respect to the impact of the real government sector deficit on 
real output and employment on the one hand and on the rate of inflation 
on the other. Parkin seems to argue that in this context the most useful 
measure of the deficit would be simply the total financial requirements of 
the government expressed in real terms. If one accepts the twin mone¬ 
tarist assumptions that the equilibrium inflation rate is ultimately deter¬ 
mined by the growth rate of the nominal stock of money, and that only 
unanticipated policy prescriptions acting on the economy can influence the 
volume of output and employment, then the cyclically adjusted real govern¬ 
ment sector deficit only affects the rate of inflation to the degree that it is 
expected, on a cumulative basis, to be monetized, and it can only affect 
the overall level of output and employment to the degree that private 
agents are unable to anticipate it. 

To put the matter differently, except insofar as government budget 
deficits and surpluses lead directly to changes in the nominal stock of 
money, their only long-run impact is to alter the equilibrium real rate of 
interest. In so doing, government deficits and surpluses have important 
and often crucial longer-run effects on the structure of the economy. But 
as far as changes in the overall level of employment are concerned, over 
the longer term, fiscal policy is rendered ineffective by the complete 
'crowding-out' of its impact. 

Parkin then goes on to argue that the real issue does not concern the 
size of the cyclically adjusted deficit, but rather its degree of exogeneity. 
To the extent that an exogenously determined structural deficit exists, 
adherence to a continuously restrictive monetary policy may lead to an 
unstable growth path for the stock of government bonds outstanding, and 
perhaps also for the real interest rate. As more bonds are issued to 
finance the exogenously determined structural deficit, more interest must 
be paid in each period of time, and this may well eventually require that 
the money supply be permitted to grow in order to retire the larger stock 
of government bonds and pay the accumulated interest charges on them. 
This process will inevitably lead to higher rates of inflation in the future 
and, to the extent that it is currently anticipated by market agents, to 
higher inflationary expectations in the present. On the other hand, if the 
cyclically adjusted deficit is endogenously determined, continuous adher¬ 
ence to a restrictive monetary policy will eventually bring it into line. 
John Grant comes to a similar conclusion in his paper to this conference: 
'As far as capital markets are concerned, the most important treatment for 


179 


deficits is likely to be steadfast monetary restraint: a slow-acting cure, 
but a certain one.' 

Thus, if Parkin's analysis is correct, the question of how the deficit 
is to be financed is crucially important to its potential impact on the rate 
of inflation. The mix of monetary and fiscal policies is therefore of vital 
importance, since it is possible for a combination of sustained large-scale 
fiscal deficits and tight monetary policy to lead to adverse effects on the 
asset structure of the economy and possible associated instabilities. In¬ 
deed, a twist towards a smaller cyclically adjusted deficit and perhaps a 
somewhat more elastic monetary policy might well have been advisable some 
three years ago (and was in fact advocated by myself, among others), 
though it is not totally clear that this would be the appropriate policy 
combination to suggest today. 

Equally important, however, is a point that Parkin ignores. It is 
crucial not only how the deficit is financed; it is also crucial how the 
associated funds are spent. Real output and employment are surely depen¬ 
dent not only upon unanticipated policy manipulations; they are also depen¬ 
dent upon the growth and structure of the real capital stock. Surely 
deficits used to finance public sector capital formation and increase the 
productive potential of the economy are much less of a cause for concern 
than deficits used to finance public consumption. Thus, when Parkin 
concludes from his analysis that 'we face a potentially serious problem 
arising from an evolving deficit that has been too big, is too big, and 
promises to remain too big,' in my view he is actually also saying that the 
government sector itself is not only too big but also guilty of dissipating 
our wealth potential through an excess of bureaucracy and regulation - or, 
more generally, through government consumption as opposed to public 
sector investment in productive capital assets, including human capital 
assets. In this, I am inclined to agree. 

Finally, although the current public sector deficit is overstated, 
given the absence of appropriate inflation-accounting for the interest paid 
on the government debt, and is not out of line when considered in relation 
to the depth of the current economic recession (although it is still per¬ 
ceived to be out of line, and perceptions are important to the behaviour of 
capital markets), I am inclined to advise that we nut increase it further in 
the current round of federal and provincial budget preparations. A no¬ 
change fiscal scenario would be my recommendation and I believe I share 
this view with Parkin and with the OECD, which has indicated that Canada 


180 




is in no position to actively follow policies that are more stimulative than 
current policies. The overall federal deficit needs to be reined in sharply 
over the next two or three years as economic recovery proceeds. 


John Bossons* 

Michael Parkin’s paper provides a useful clarification of the various pur¬ 
poses for which the government deficit may be measured and, as well, of 
how it should be measured in each case. I want to emphasize the import¬ 
ance of recognizing the different purposes for which deficit measures may 
be used, as well as to discuss some of the issues which Parkin raises in 
his comments on my recent paper with Peter Dungan on this subject. 1 

Our paper was motivated by a belief that much of the concern over 
the size of the deficit expressed by non-economists reflects a widespread 
fear that the long-run rate of growth of the government debt has become 
insupportable. If we are correct in assuming this fear to be a major 
element in the current reactions of financial market participants, then it is 
of signal importance to obtain quantitative measures of the deficit that 
show what is happening to the long-run growth rate in the total real net 
liabilities of the government sector (or, alternatively, in the long-run 
growth rate of the ratio of government debt to GNP, including consolidated 
liabilities of the Bank of Canada as part of the government debt). To do 
so, it is necessary to estimate the empirical magnitude of the real full- 
employment surplus or deficit, as specified by the cyclically adjusted 
version of the third measure (d| t ) defined in Parkin's paper. 

Parkin and I are in total agreement that this is the correct measure 
to use in evaluating whether the long-run fiscal position of the government 
has changed. 2 Depending on the question posed, either the level or rate 
of change of d^ may be of interest. The rate of change of d| t provides 
information on whether the government's longer-run fiscal stance is becom¬ 
ing looser or tighter. The level of d^ is of interest in evaluating whether 
the existing 'social contract' represented by the current tax, transfer, and 
expenditure policies of the government sector is sustainable in the long 
run. Such an evaluation does not in itself presume that future govern- 

* Professor of Economics, Institute for Policy Analysis, University of 
Toronto. 


181 


ments (or the citizens that elect them) will necessarily want to continue 
these policies in an unchanged form, but is simply concerned with the 
feasibility of such continuance. From this viewpoint, a level of d| t (ex¬ 
pressed as a fraction of current real GNP) that exceeds some critical value 
may be viewed as eventually unsustainable. 3 This critical value is what 
would imply an unchanged average ratio of total government debt to GNP 
in the long run; it is equal to the product of some given target debt/GNP 
ratio and the likely average long-run growth rate of real GNP due to 
productivity increases and population growth. 

Note that I have here defined the purpose of d| t as one of evaluating 
the feasibility of a given set of tax, transfer, and expenditure policies. 
This underlying issue is one of political economy, not macroeconomics. 
Nevertheless, it has macroeconomic implications to the extent that a set of 
government policies that is perceived as unsustainable causes investors to 
project an increased likelihood of future increases in taxes on income from 
capital. Other decisions may also be affected by fears of other tax in¬ 
creases (or of cutbacks in transfers or expenditures) that may occur in 
response to the political pressures induced by an increasing debt burden. 

Given the potential macroeconomic (and political) importance of public 
perceptions of the long-term sustainability of government revenue and 
expenditure policies, it is clearly useful to provide accurate measures of 
the long-run fiscal balance of the government sector that are corrected for 
the short-term effects of macroeconomic fluctuations. From a social accoun¬ 
ting viewpoint, the corrected measure (d| t ) is a more useful indicator than 


its unadjusted counterpart (d^). This is as much true during a recovery 
as it is during a recession. During a recession, the correction prevents 
the recession-induced increase in the unadjusted deficit from being per¬ 
ceived as a change in the long-run fiscal policy of the government sector. 
During a recovery, focusing attention on d| t helps ensure that a reduction 
in the unadjusted deficit that is smaller than the recovery-induced deficit 
reduction is correctly perceived as a worsening of the government's long- 
run fiscal position. 4 

Turning to the actual measurement of the real structural deficit 
(d| t ). Parkin and I differ strongly on our interpretations of what is 
implied by 'full employment.' Dungan and I used 1979 factor utilization 
rates as a standard for this purpose, in part because most of the supply- 
side shocks that occurred during the 1970s were already reflected in these 
utilization rates. The sources of the 1981-2 recession were in monetary 


182 


policy, not supply shocks. Moreover, the 7.5 per cent unemployment rate 
experienced in 1979 was significantly above the high end of estimates of 
the long-run natural equilibrium unemployment rate for that year. 5 To 
suggest, as does Parkin, that the natural unemployment rate is now in the 

neighbourhood of 10 per cent is casual empiricism taken to the extreme. 

This comment should not be taken as a denial of the potential signifi¬ 
cance of structural shifts in employment due to changes in relative prices. 
Particularly when investment is generally depressed, as it is at present, 
such structural shifts in demand may be an important factor contributing 
to the persistence of unemployment. But one must be careful in interpre¬ 
ting reduced-form estimates adduced as evidence for the empirical signifi¬ 
cance of such effects. 

Let me now turn from social accounting to deal with some of the 

issues raised by Parkin in describing the macroeconomic effects of the 

deficit. I will first deal with the empirical question of how fiscal policy 
changed during the onset of the 1982 depression and then turn to some 
theoretical issues. 

An activist counter-cyclical fiscal policy would be one in which the 
real structural deficit (d| t ) was increased during an economic depression. 
In this, contrary to the impression created by Parkin's remarks at the end 
of the third section of his paper, Parkin and I are in agreement. What 

Dungan and I concluded in our paper (pp. 2-3) was that 'the tax in¬ 

creases and reductions in government expenditures implemented in the past 
two years (1981-2) have effectively eliminated the "chronic" federal deficit 
of the late 1970s.' In 1979, our measure of d| t for the federal government 
amounted to approximately 1.5 per cent of GNP. By 1982, government 

policy changes had reduced this structural federal deficit to approximately 
zero. From a stabilization viewpoint, this change in fiscal policy was 

perverse. 6 

Whether the decrease in aggregate demand caused by the move to a 
more contractionary fiscal policy can be said to have resulted in a decrease 
in aggregate output is questioned by Parkin in the fifth section of his 
paper. This is not the place to deal with this issue at length. However, 
it should be noted that Parkin assumes that an aggregate supply function 
dependent on the ratio of actual to expected prices implies (1) that only 
unanticipated shifts in aggregate demand matter and (2) that only unanti¬ 
cipated deficits can affect economic activity. Moreover, Parkin argues that 
deficits must generally be regarded as anticipated. While the first of 


183 



Parkin's assumptions is generally accepted as a characterization of the 
determination of aggregate prices and output in a single-commodity macro- 
economic model for teaching purposes, it requires very strong restrictions 
on the process by which expectations are formed to derive the 'implication' 
that deficits have no effect on aggregate output. As with other 'rational 
expectations' models, Parkin's presumed policy ineffectiveness breaks down 
quickly once the real-world complications of multiple commodities, informa¬ 
tion costs, and heterogeneous decision-maker expectations are introduced. 

Parkin's analysis of the potential inflationary effect of deficits correct¬ 
ly emphasizes the necessity of modelling the linkage between structural 
deficits and longer-run rates of monetary growth. The structural deficit 
as defined above (d| t ) is a real variable that in itself has no monetary 
significance; a long-run inflationary impact can only be derived if a given 
sustained value of the ratio of d| t to real GNP can be shown to be suffi¬ 
ciently high to imply unstable monetary growth. To put this in perspec¬ 
tive, it is useful here to note than even in the United States, which is 
currently experiencing an unusually high structural deficit, the current 
ratio of the real structural deficit to GNP is generally estimated to be on 
the order of 2 per cent. It would take over 30 years of continued deficits 
on this scale to bring U.S. debt/GNP ratios back up to where they were 
at the end of the Second World War. At that point, the growth rate of 
real government liabilities would approximately equal the growth rate of the 
U.S. economy, thus satisfying the McCallum condition for potential price 
stability. 

It is in fact a mistake to assume that the structural deficit is exo¬ 
genously determined. Some interesting evidence on the longer-run behav¬ 
iour of governments in response to deficits is presented in Section 2 of my 
paper with Dungan. The hypothesis advanced there (based in part on the 
Canadian response to the federal deficits created during the 1970s by the 
indexation of personal income tax rates in 1973) is that 'chronic' deficits 
are gradually eliminated, and that the 'steady state' to which Canadian 
governments tend to converge is a zero structural deficit. 7 Equally inter¬ 
esting, we show that the size of the deficit has had an important influence 
on the rate of growth of the government sector (as measured by the ratio 
of total government expenditures to GNP). 8 

I do not believe that it is a correct characterization of the political 
process to say, as does Parkin, that it appears to generate ever-increasing 
expenditures and deficits. While this view is a popular one, it is not 


184 


supported by empirical evidence. A better characterization is that the 
political cost of expanding the relative size of the government sector is an 
increasing function of the size of the deficit and a decreasing function of 
the growth rate of revenues produced by the currently defined tax system. 
Moreover, deficits are in themselves politically costly, and these costs 
create pressures to reduce expenditure growth (and/or increase taxes) in 
order to lower the deficit. Viewed in this context, the U.S. structural 
deficit generated by the Reagan administration may well be politically the 
most effective way of lowering the growth of the U.S. government sector. 
To define the central issue in modelling the inflationary impact of the 
deficit as one of the exogeneity of the deficit is to ignore the relationship 
between the size of the deficit and other important features of fiscal policy. 

NOTES 


1 See J. Bossons and P. Dungan (1983) 'The government deficit: too high 
or too low?' Canadian Tax Journal 31, 1-29. 

2 It should be emphasized that for this purpose d* t must be defined on 
an accrual basis and so should include the change in the unfunded 
pension liabilities of the government sector. On this Parkin and I 
also agree. For a contrary view, see the comment by Asimakopulos and 
Ascah in the May-June 1983 issue of the Canadian Tax Journal . 

3 The word 'eventually' should be underlined. In the short run, temp¬ 
orary growth in the ratio of total government liabilities to GNP is 
clearly feasible without affecting the sustainability of any given 
'social contract' in the long run, even if such temporary growth is 
not subsequently reversed. 

4 The estimates presented in the Bossons-Dungan paper indicate that, to 
keep the federal government's long-run real fiscal position unchanged, 
the real federal deficit would have to fall by $14 billion if the 
economy recovered to a level of factor utilization associated with a 
7.5 per cent unemployment rate. In addition, the federal deficit as 
conventionally measured should fall by a further $6 billion if the 
inflation rate falls below 3 per cent (see Bossons and Dungan 1983, 
Table 11). Anything less than a $20 billion reduction in the federal 
deficit as the economy recovers to more normal levels of unemployment 
and inflation (here defined as 7.5 and 3.0 per cent, respectively) 
should thus be taken as evidence of a worsening in the long-run fiscal 
position of the federal government. 

5 See the discussion of this issue in Bossons and Dungan (1983) 19, fn. 
19. The estimates reported there suggest that the long-term natural 
rate for 1983 is likely below 6 per cent. While the estimates pre¬ 
sented in Table 4 of the paper by Lilien cited by Parkin show a sub¬ 
stantial increase in Lilien's estimated natural rate from the mid- 
1960s to the late 1970s, this increase is limited to 2.5 percentage 


185 







points. Moreover, Lilien's results are based on a reduced-form model 
that does not incorporate changes in the supply of labour due to the 
changing age/sex/race composition of the U.S. labour force or due to 
work-incentive effects of changes in transfer payment programs. Some 
portion of the secular changes in unemployment rates that Lilien 
attributes to structural shifts in the demand for labour is in fact 
attributable to structural shifts in labour supply of the kind reflec¬ 
ted in the Dungan-Wilson estimates reported in Bossons and Dungan, 
ibid. Indeed, it is noteworthy that Dungan and Wilson’s estimates of 
the increase in the natural rate over the 1965-79 period are almost 
identical to the estimates obtained by Lilien for the U.S. but 
ascribed by Lilien to different causal factors. 

6 A somewhat different picture is presented by recently published 
estimates prepared by federal government officials (see Department of 
Finance (1983) The federal deficit in perspective , Chart 7). These 
estimates indicate that fiscal policy was contractionary to the degree 
indicated by the Bossons-Dungan paper during the 1979-81 period, in 
that the real (inflation-adjusted) federal deficit moved from approx¬ 
imately 1.7 per cent of GNP to zero over this period. However, the 
federal government estimates show a move to greater stimulus during 
1982. This difference partly reflects a different treatment of 
changes in petroleum taxes, which are treated on more of a cash basis 
in the federal estimates. 

7 This implies that, on average, the ratio of total real government debt 
to real GNP declines with time under peacetime circumstances. This 
has been empirically the case in both Canada and the United States 
during the post-Second-World-War period. 

8 In general, the government sector has grown fastest when the govern¬ 
ment has had a ’chronic' structural surplus. Eliminating the motiva¬ 
tional effect of the chronic surplus caused by inflation-induced 
increases in effective tax rates was one of the principal goals under¬ 
lying the introduction of indexation of personal income tax rates in 
1973. The shift to a chronic deficit caused by the resultant fall in 
the growth of government revenues coupled with built-in momentum in 
expenditure programs was an important element of the process which led 
to a halt in the growth rate of the ratio of federal expenditures to 
potential output (see Bossons and Dungan 1983, Table 2, 9). 


MICHAEL PARKIN 

There is a considerable measure of agreement between John Bossons and 
myself. Nevertheless, several issues divide us. 

The first broad area of disagreement concerns the calculation of ’full 
employment’ for the purpose of assessing 'full-employment deficit.' There 
appear to be three specific points of contention. 

First, John Bossons says that 'the sources of the 1981-2 recession 
were in monetary policy, not supply shocks.' This leads him to the con- 


186 



elusion that the natural rate of unemployment did not increase materially in 
the recent recession and that the present rate is the same as the rate that 
would have been assessed in the late 1970s. I am less confident than 
Bossons about the sources of the 1981-2 recession. I certainly do not 
disagree that monetary policy played a critical role. However, I also am of 
the view that unusually large fluctuations in international relative prices 
and unusually high real rates of interest generated unusually large 
amounts of resource reallocation during that recession. Consequently, I 
would conjecture, for reasons identical to those set out by David Lilien 
(1982) (and cited in my paper), that the natural rate of unemployment has 
risen in recent years. By how much I do not know. 

Second, Bossons argues that I suggest, in contradiction to what I 
have just said, that the natural unemployment rate is now in the neigh¬ 
bourhood of 10 per cent and that to make such a remark is 'casual empiri¬ 
cism taken to the extreme.' What I in fact said is that unemployment rates 
closer to 10 per cent, at least for the foreseeable future, appear to be 
more appropriate given the current possibilities concerning the sustainable 
constant-inflation average unemployment rate.' I take this to indicate that 
I am unsure of what the natural rate of unemployment is and believe that 
the sustainable average rate (a rate higher than the natural rate by virtue 
of the analysis presented in my paper) may be approaching 10 per cent. I 
certainly do not object to Bossons' attaching the label 'casual empiricism' to 
this number. I do object to his calling it my estimate of the natural rate. 
It is a conjecture about a number bigger than the natural rate towards 
which the sustainable average rate may be tending. 

Third, we apparently disagree about the appropriate way in which to 
calculate the average sustainable unemployment rate. Bossons' comments 
do not deal with this issue. I assert that the Bossons-Dungan calculations 
are incorrect. They are based on the presumption that the economy is 
deterministic and not stochastic. Such an assumption is clearly untenable. 
How important this error is in quantitative terms I do not know. The 
question clearly requires further investigation before assertions as confi¬ 
dent as Bossons' can be regarded as having sufficient credibility to pro¬ 
vide a basis for the design and conduct of policy. 

A second broad matter on which we apparently disagree concerns the 
use of macroeconomic theory for the purpose of reaching policy judgements 
and conclusions. In my own work I use quite explicitly specified, well 
understood macroeconomic models. I generate predictions from those 


187 








models and make policy statements in the light of those predictions. 
Bossons asserts that 'as with other "rational expectations" models. Parkin's 
presumed policy ineffectiveness breaks down quickly once the real-world 
complications of multiple commodities, information costs, and heterogeneous 
decision-maker expectations are introduced.' It is difficult to know where 
to begin commenting on this amazingly obfuscating statement. I can only 
conclude that Bossons is attempting to muddy the waters by pretending 
that because the real world is complicated and not exactly described by my 
model that in some way it is characterized and described by his. The 
currently available generation of rational expectations macroeconomic models 
may well be inadequate. But their inadequacy pales into insignificance 
beside the inadequacy of the Keynesian models on which Bossons bases his 
policy conclusions. Where in the IS-LM expectations augemented Phillips 
curve models do we see 'multiple commodities, information costs, and 
heterogeneous decision-maker expectations?' I leave the question in 
rhetorical form. 

Finally, Bossons and I disagree on how the currently evolving deficit 
should be read. Bossons sees this deficit as a transitory deviation from a 
long-run sustainable position that is compatible with macroeconomic stabi¬ 
lity. I am less sure. I certainly do not want to assert that Bossons' view 
is demonstrably wrong. I do worry, however, when conclusions such as 
his are reached in the absence of an appropriate, careful, statistical 
investigation of the stochastic processes governing the evolution of the 
deficit, revenues, real output, monetary growth, and, most important of 
all, the state of war and peace. Historically, this last factor more than 
any other has governed the evolution of deficits. To compare current 
debt-income ratios with those that prevailed at the end of the greatest war 
that mankind has ever seen hardly seems appropriate. 

In summary, I believe that Bossons is too confident in his assertions 
about the natural rate of unemployment; that he has made analytical errors 
in translating those beliefs into calculations of the average sustainable 
high-employment deficit; that he is too cavalier in his dismissal of the best 
macro-economic models that the profession has so far managed to develop; 
and that his conclusions concerning the seriousness of the current deficit 
situation are insufficiently grounded in careful statistical analysis. My 
'bottom line' is not that Bossons is wrong. It is that we do not know 
enough to state conclusions with as much confidence as Bossons is 
apparently ready to state them. 


188 



What does the public think about 
deficits? 

What does Bay Street think about 
deficits? 

Ian McKinnon* 


The subject of this paper makes it substantially different from the other 
papers given at this conference. Faced with the sophisticated econometric 
and theoretical work being presented here, one is tempted to be apologetic 
about discussing public opinion data and to denigrate it as representing 
what Professor Modigliani might refer to as 'naive' views. Particularly 
when we are examining government actions, however, few things are more 
important than public opinion. It is perhaps a crude analogy, but none¬ 
theless an accurate one, to say that, for governments, public opinion is 
the bottom line. 

It seems to me, as a public opinion analyst, that many attempts to 
explain government actions underestimate the importance of people's per¬ 
ceptions and, particularly, their preferences as voters. Much economic 
analysis posits rational models of what should happen as opposed to what 
might happen, given the vagaries of human behaviour. On the other 
hand, the traditional opinion approach to the studying of consumer behav¬ 
iour - market research - is premised on the notion that behaviour is 
motivated by product consideration alone - as if a downturn in consumer 
activity were somehow a function of growing dissatisfaction with product 
choice in the marketplace. Clearly, the events of the past two years have 
brought both of these traditional forms of analysis into question. 

Of course, I have a vested interest in believing that a fuller under¬ 
standing of how the individual behaves as a citizen - as opposed to how he 
or she behaves only as a consumer - can assist the more traditional forms 
of economic analysis. And it can assist the economist in understanding not 
only how the public behaves but how governments respond to that public 
behaviour. This understanding must begin with a knowledge of the cur¬ 
rent mood of the public. 

* Vice-President, Research Department, Decima Research. 


189 



The Canadian public has been traumatized by the course of the last 
two years. Canadians' traditional beliefs have been contradicted by their 
current experiences and perceptions - yet they tenaciously continue to 
cling to those traditional beliefs. They have seen our problems worsen 
and become increasingly more complex, yet they continue to view them as 
aberrations - problems that we should not have, that we don't deserve, 
and that are eminently solvable. Canadians are absolutely tenacious in 
their conviction that Canada has the resources - physical, human, and 
financial - to solve our problems. This conviction is the key to both the 
public's perception of today's economic problems and its expectations with 
respect to government. 

CANADIANS AND ECONOMIC ISSUES 

Let us begin by looking at what Canadians consider to be the most impor¬ 
tant problem facing the country. As we can see from Figure 1, concern 
about economic issues has climbed from the 50 per cent level three years 
ago to the 70 per cent level. The economy now dominates the political 
agenda. 

Figure 2 shows that this increase in concern over economic issues has 
gone hand-in-hand with a precipitous decline in the public's assessment of 
the economy. Three years ago, 20 per cent of the adult Canadian popula¬ 
tion described the state of the economy as 'poor'; in June of 1982, 65 per 
cent used this description of the economy. Although matters have im¬ 
proved considerably since then, only a small minority have positive views 
of the economy. 

While an exceptionally weak economy is unquestionably a liability for 
any government, the amount of pressure that it puts directly on the 
government depends upon the extent to which people hold the government 
responsible for the economic situation. This, in turn, is closely related to 
beliefs about the intrinsic solvability of economic problems - I have already 
said that Canada's problems are viewed as aberrations and solvable, but 

for government policymakers the vital fact is that three-quarters of the 

population believe that they can be solved in whole or in part, by the 

federal government (see Figure 3). 

The result of this firm conviction that government can at least ame¬ 
liorate Canada's problems, and of the growing concern with economic 
issues, has been strong pressure on government and persistent dissatis- 


190 



Figure 1 

Canada’s most important problems 


% 



’80 ’80 ’80 '80 '81 '81 '81 '81 '82 '82 '82 ’82 '83 '83 '83 '83 


Figure 2 

Perceptions regarding Canada’s economy 



Generally speaking, how 
would you describe 
Canada’s economy today? 
Would you say it is 
excellent, good, only fair 
or poor at this time? 


Current 


Data 



% 



35 


Poor 

49 


Only fair 

0 


No opinion 

1b 


Good 

1 

■ • • • 

■ • • ■ 

Excellent 


191 





























































































































































































































































Figure 3 

Perceptions regarding the federal government’s ability to solve 
Canada’s most important problems 


% 

100 


How much do you think 
the government of Canada 
can do to solve this prob- 
lem? Do you think they 
can solve it totally, solve it 
partially, not do very much 
to solve it, or not do any- 
thing at all to solve it? 

Current 

Data 

III 

m 

m 

W'* 


m 

• ) / 

7 7y7\ 

OYY 


yM 

Ay/A 

W/A 



y/ M 

2 

m 







90- 






m 
























1 1 



i 















QH- 



; • 


J ‘- r 


m l 

I 















OU 

1 

1 " 

-ip-., 


1C 


i: jh? 

m 







7 








70- 


J-r 



P 

— 

| i 



kki 


ll 



1 









ikk 

Xk\ v\ 


J } . 

vvvspN 

P 

P 

P 

lkk 

h 

H 


kkk 



$ 

1# 

1 








60- 

> 

lii 

Htl 

Hk 

Tkk 

P 

P 

Hi 

Ikk 

Hk 

ikk 

■ -i. 

\\Nl 

1 


il 



P 










111 

111 

Gh 

Hk 

P 

P 

kK\Ki 

p 

' > > 

\V\N 

§ 



ii 

ii 

l 

P 

1 








Cf) . 

p 

\ \ 

SJSKl 

\X\X\ x 

P 

P 

P 

Ikk 






kk 

C\ ks 


P 

k! 

|| 







O vJ 

il 

SSlkl 

\V\X\' 

\ A : 

P 

ll 

Gy, 

nxNKl 

;! | > 

OsK\ 

k 



\\ \N 

11 


P 









40- 

P 

Hk 

Hk 

Ikk 

P 

P 

s\KxkN 

AKknN 

\V\nK> 

V\/VNV s 

\VSN 

VfS 


PfPV' 


11 

ll 


P 

1 










OsK\K\ 

> •; 

|1 

P 


Hk 

Ikk 

H 

H 


kkk 



§ 


k 








% 

8 

14 

2 

65 

11 

m 

Nothing at all 

Not very much 

No opinion 

Partially 

Totally 

00 - 

p 


> > 

P 

P 

HJvtX 

Hk 

Ikk 

PS 


P 




P 

1 









HI 

Hk 

Hk 

Hk 

Hk 

1H 

1h 

P 

\\Knn> 

SahjS 

\ V 

nfviv 

■ "J ■ 

§ 

il 

il 



li 









in 

i a a a a a 
i a a a a a 
• a a a a ■ 
•aaaai 

on - 

ip 


\ 1 

WVsSN 

P 

v vCv 

\V\VV 

Hk 

P 

vyv 



kkk 



P 


II 







10- 

ii 

■ •HIM 

P 

\M\xV 

w 

|1 

■■•■«■■ 
■ •Miia 

p 

■ ■ • ■ ■ •(■ 

■••■■as 

p 

■••■«■■ 
• a a a a a cs 

•Hk 

■ ■ • ■ 1 a a 

h 

OvJVS 

\V\N 

■ l«M 

k 

• • 

ll 

11 

\\\1 

Ik 

1 

a a 
a a 

P 

life 

1 

II 

ii 








lllltll 


■ iMMa 

Vi!!** 

> ■ a a a a ■ 

■■»■■■■ 
■ iMMa 

■ a • ■ ■ ■ a 
• a • ■ ■ ■ 4 

i a 4 a * a a> 


i a a a a a a 
i a 4 a a a a 
i a 4 a a a 

fTa a a a a 
■ a a • a a a 

i a ■ 

• a a ■ 

a a a a a • 
a a a a 4 a ■ 
a a 4 a*a• 

> a 

9 a 

i a i a a 
i a i a ■ 







0 

■■■••os 

■■■■■■■ 


::::::: 

■■»■■■• 
■ ■«■«■ » 

■ItlMl 

■■•■■•a 

• • • a a ■ • 
iiiiaia 

■■•■■as 

rrtmr 

i a 1 a a a a 

iiiirn 

'••■Mi 

ttliin 

iitaM* 
> a 4 a a a a 

■ •Mi 
l a a 

■ ■ 
a a 

aatiiai 

aaaajad 

9 a 

i a i a a 








Mar. Jun. Sep. Dec. Mar. Jun. Sep. Dec. Mar. Jun. Sep. Dec. Mar. Jun. Sep. Dec. 

'80 '80 '80 '80 '81 '81 '81 '81 '82 '82 ’82 '82 '83 '83 '83 J 83 


faction as adequate solutions have failed to emerge. As Figure 4 shows, 
the percentage of Canadians who were satisfied with the performance of 
the federal government dropped from 45 per cent in 1980 to 20 per cent by 
June 1982. A government faced with such a decline in public confidence 
had little choice but to act, or to be seen to act, to alleviate matters. 

It is clear that the electorate now ascribes responsibility for macro- 
economic performance to the government and the politicians. As a con¬ 
sequence, political survival may well depend upon being seen to act in the 
face of economic distress. Given these pressures, a status quo policy is a 
tremendous risk for any politician. Politicians are going to act; they will 
respond to popular wishes. The question remains: What does the public 
want? 

To answer this question, particularly in the context of this confer¬ 
ence's concern with the deficit, let us look in more detail at the current 
economic concerns of Canadians. As we saw earlier, almost three-quarters 
of Canadians view some economic concern as the most important problem 
facing Canada. Breaking the economic concerns into specific issues yields 
the results in Figure 5. Unemployment, mentioned by almost half the 


192 





































































































Figure 4 

General assessment of the federal government’s performance 


% 



Generally speaking, how 
satisfied are you with the 
performance of the federal 
government? Would you 
say you are very satisfied, 
somewhat satisfied, some¬ 
what dissatisfied or very 
dissatisfied? 


Current 

Data 

Very dissatisfied 

Somewhat 
dissatisfied 

No opinion 
Somewhat 
satisfied 

Very satisfied 



respondents as Canada's most important problem, now supersedes all other 
issues, and has replaced inflation as the primary concern. Governments 
will have to be seen to act on this concern. 

I would now, a la Sherlock Holmes and the dog that didn't bark, 
draw attention to something that is missing from this list of problems: the 
issue of the deficit or the national debt. While this issue has occasionally 
been mentioned by respondents, only in March 1981 did more than 1 per 
cent of the Canadian electorate polled in Decima's quarterly surveys 
believe that it constituted the most important problem facing Canada. On 
that occasion, 1.6 per cent of our respondents mentioned the national debt 
or the deficit. 

Let us summarize what we have found thus far: 


Canadians are convinced that their problems are solvable and, indeed, 
that the problems are aberrations. 

Their primary current concerns are economic. They are particularly 
concerned about unemployment, but not about the deficit. 

Canadians believe that the government can solve these problems, if 
not totally, then at least partially. 


193 




























































































Figure 5 

Canada’s most important economic problems 


% 

50 


40 


30 


20 


10 


0 





































■ * 

.« 

• • 









































•* 























































































/ 










































































-J. 

y 












































/ 


\ 












• 






























/ 




\ 










c 













































• 























































































•• 









































/ 


k 










































/ 





y 






































4 

f 








> 

. 


































-Jl 

✓ 

/ 


- 

—- 

i 





J 



















✓ 

* 

* 





•« 

o 

• 

V 





N 

• 

t 

f 

/ 

■' 





k 

N 



✓ 

r 



S 









... 






















11 e 

c. 


• 





y 

y 





N 

\ 





S 


- 





















/ 

r 

















k 

N 


























/ 

f 



















\ 

■> 














r 










* 


































What is the most important 
problem facing Canada to¬ 
day — in other words, the 
one that concerns you per¬ 
sonally the most? 


Current 

Data 

Inflation (10%) 


••• Unemployment (49%) 
- Interest Rates 


— Economy — 
General 


( 2 %) 

(9%) 


Mar. Jun. Sep. Dec. Mar. Jun. Sep Dec. Mar. Jun. Sep. Dec. Mar. Jun. Sep. Dec. 

'80 '80 '80 '80 '81 '81 '81 '81 '82 '82 ’82 '82 '83 '83 '83 '83 


Together, these beliefs put tremendous pressure on governments to 
be seen to act to solve economic problems. We can rest assured that 
governments will respond to those pressures. 


THE PUBLIC AND THE DEFICIT 

Canadians expect action from their governments, that we know. Now let 
us examine the actions that they believe would do the most good. 

In September 1982, we at Decima asked: 

As you know, every government is faced with different priorities on which 
it could be concentrating. Some of these priorities could include creating 
jobs through government spending, stimulating industry through tax cuts, 
lowering interest rates, reducing government spending by limiting public 
service wage increases or encouraging more foreign investment. Of these 
five things ... which do you think should be the first priority of govern¬ 
ment in Canada? And which do you think should be the second priority of 
government? 

Figure 6 shows the results of this poll. The biggest surprise in these 


194 

























































































results is the higher priority given to lowering interest rates than to 
'creating jobs through government spending' (this after unemployment had 
significantly surpassed interest rates and inflation as concerns). We 
should also note the distinct secondary interest in 'reducing government 
spending by limiting public service wage increases'. 

To probe this approval of the idea of retrenchment further, we speci¬ 
fically asked about the 'six-and-five' program: 

As you may also know, as one way of dealing with inflation, the June budget 
introduced a programme which would limit the size of salary increases for 
federal public servants to six per cent this year and five per cent next 
year. Generally speaking, would you say you approve or disapprove of the 
federal government's programme limiting wage increases for public servants 
to six and five per cent? 

While the nearly 3:1 approval shown in Figure 7 is not surprising, it is 
interesting to bear in mind the relatively low priority accorded such initia¬ 
tives in the previous question. The public seems to see such reductions 
in government spending as helpful and, perhaps, as a necessary first 
step, but it does not see them as immediate or vital prerequisites for 
economic recovery. 

Next we looked specifically at the public's preferences with respect to 
the deficit itself and asked the following question: 

There has been a lot of discussion recently about government deficits, the 
fact that governments spend more money than they collect in taxes. Some 
people say that governments must reduce their deficits in order to get the 
economy growing again. Other people say, that in tough economic times, 
governments should stimulate the economy to encourage growth, even if it 
means having large deficits. In your opinion, do you think it would help 
the economy a great deal, help it somewhat, not make any difference, hurt 
it somewhat or hurt the economy a great deal if governments reduce their 
deficits? 

As Figure 8 indicates, there is clearly a preference for fiscal restraint. 
Moreover, there is relatively little dispute about the mechanism that should 
be employed for reducing deficits. We asked: 

Governments are considering two different ways of reducing their defi¬ 
cits ... by increasing taxes or by reducing the services which they would 
provide. Which way do you think governments should reduce their defi¬ 
cits ... increase taxes or reduce services? 


The response to this question, shown in Figure 9, certainly does not 
depict a Canadian public given to Keynesian visions of governments spur- 


195 





Figure 6 

Perceptions of government policy priorities 



First and 
Second 
Priority 
(Combined 
Mentions) 


Lowering 
Interest Rates 

Creating 
Jobs 
Stimulating 
Industry 
Limit Public 

Service Wage Increases 
Encourage Foreign 
Investment 


As you know, every government is faced 
with different priorities on which it 
could be concentrating. Some of these 
priorities could include creating jobs 
through government spending, stimulat¬ 
ing industry through tax cuts, lowering 
interest rates, reducing government 
spending by limiting public service wage 
increases or encouraging more foreign 
investment. Of these five things... 

Which do you think should be the 
first priority of governments in 
Canada? 


And which do you think should be 
the second priority of government? 


Figure 7 

Levels of approval for the federal government public service wage 
restraint programme 


Approve 



72% 



No Opinion 




3% 


As you may also know, as 
one way of dealing with 
inflation, the June budget 
introduced a programme 
which would limit the size 
of salary increases for 
federal public servants to 
six percent this year and 
five percent next year. 
Generally speaking, would 
you say you approve or 
disapprove of the federal 
government’s programme 
limiting wage increases 
for public servants to six 
and five percent? 


ring aggregate demand through deficit spending. Instead, it strongly 
indicates a 'public household' view of what governments' actions should be. 
As individuals should try to economize and save in hard times, so too 
should governments. 

Does the Canadian public really want a much reduced level of gov¬ 
ernment services? To look at this, Decima offered a number of trade-offs 
in the following fashion: 


196 






























Figure 8 

Perceptions of the effects of reducing government deficits on 
the economy 


Help a 
Great 
Deal 


21 % 


Help 

Somewhat 



Not Make 
any 
Difference 



13% 



Hurt 

Somewhat 



11% 


Hurt a 
Great 
Deal 



9% 


There has been a lot of discussion 
recently about government deficits, 
the fact that governments spend 
more money than they collect in 
taxes. Some people say that 
governments must reduce their 
deficits in order to get the economy 
growing again. Other people say, 
that in tough economic times, 
governments should stimulate 
the economy to encourage growth, 
even if it means having large deficits. 
In your opinion, do you think it 
would help the economy a great 
deal, help it somewhat, not make 
any difference, hurt it somewhat or 
hurt the economy a great deal if 
governments reduced their 
deficits? 


Figure 9 

Preferred method of reducing government deficits 



Governments are considering 
two different ways of reducing 
their deficits... by increasing 
taxes or by reducing the 
services which they provide. 
Which way do you think govern¬ 
ments should reduce their 
deficits... increase taxes or 
reduce services? 


Now, I'm going to read you a list of some of the ways in which governments 
could cut their spending in order to reduce their deficits and I’d like you 
to tell me whether you would approve or disapprove of each one, if making 
that spending cut would significantly help to reduce government deficits. 
How about ... would you approve or disapprove of government doing that in 
order to reduce deficits? 


197 

















As Figure 10 shows, by far the most popular candidate for a cut in spend¬ 
ing is the public service. A full 83 per cent of Canadians would approve 
of governments reducing the number of people who work for them as a 
means of reducing their deficits. Although far from unanimous, Canadians 
are more inclined to approve (55 per cent) than disapprove (42 per cent) 
of reducing postal service to three days a week. Eliminating all financial 
support for the CBC would be the third most popular way of reducing the 
deficit, with 50 per cent of Canadians approving of such a move and 42 
per cent disapproving. The public is also more likely to approve (50 per 
cent) than disapprove (44 per cent) of a drastic reduction in the number 
of people eligible for unemployment insurance benefits as a way of reduc¬ 
ing deficits. 

Outside of the four areas just mentioned, Canadians are more likely to 
disapprove than approve of the spending cuts suggested. Least likely to 
meet with disapproval would be reducing passenger rail and ferry services 
(55 per cent), and eliminating support for research and development 
conducted by private industry (56 per cent). Interestingly, cutting 
student aid and subsidies meets with a greater degree of disapproval (75 
per cent) than cutting spending on job creation programs (65 per cent). 
The two possible methods of reducing government deficits that meet with 
the least amount of approval are the cancelling of all family allowance 
payments (20 per cent) and reducing access to health care (12 per cent). 

The public clearly does not want to see a dismantling of the basic 
social programs that have been put in place over the past thirty years. 
In other words, government is stuck between a public opinion rock and a 
fiscal hard place. There is simultaneously a preference for decreasing the 
deficit and a continued demand for the maintenance of many of the most 
expensive educational and social welfare programs. 

BAY STREET AND MAIN STREET 

To complement the opinion data that we had gathered from the Canadian 
public, we also interviewed twenty-five senior members of the financial 
community. These interviews were conducted with members of banks, 
trust companies, and brokerage firms, with bond specialists, and with 
representatives of groups that borrow in the bond market or supervise 
large investment funds. Although the intention was to avoid professional 
economists, a few crept into the sample and provided a very interesting 
contrast to many of the others with whom we spoke. 


198 


Figure 10 

Levels of approval for spending cuts in specific areas in order to 
reduce government deficits 

I’d like you to tell me whether you 
would approve or disapprove of 
each one, if making that spending 
cut would significantly help to 
reduce government deficits. 

How about... 


reducing the number of people 
working for the government? 


reducing postal service to three 
days a week? 




eliminating all financial 
support fortheCBC? 



drastically reducing the number of 
people who would be eligible 
for unemployment insurance 
benefits? 



reducing passenger rail 
and ferry services? 



cutting out support for research 
and development conducted 
by private industry? 



cutting spending drastically on 
job creation programmes? 




cutting student aid and subsidies so 
that the average post-secondary 
tuition fee would double? 


cancelling all family allowance 
payments? 



reducing access to health care? 



Approve % Disapprove 


199 




















Our first step was to repeat many of the questions we had already 
submitted to the public and to ask two things about them: 

How did the financial community respondent answer the questions? 

How did the respondent think the Canadian public had answered the 

same questions? 

The first two columns of Table 1 compare the responses of the public 
and the financial communities. It is interesting to note that the financial 
community is more favourably disposed towards stimulating industry and 
reducing government spending, while the public's preference is for lowered 
interest rates and job creation programs. The picture becomes more 
interesting, though, when we consider the third column in Table 1, which 
shows the financial community respondents' estimates of the public answer. 
What is particularly noteworthy is that 74 per cent of the financial com¬ 
munity respondents believed that job creation programs were the public's 
first priority; in fact, a plurality of Canadians wanted a lowering of inter¬ 
est rates to be the first priority. 

The tendency apparent in Table 1 - to underestimate the fiscal fru¬ 
gality of the average Canadian - is quite apparent in the responses to the 
question regarding the effects on the economy of reducing the deficit. 
The irony here lies in the financial community's believing that Canadians 
are not all that conservative fiscally. In Table 2, we see a public more 
conservative than the financial community. 

The same phenomenon was present when the public was asked to 
choose between a service and cutting that service, if by doing so it 'would 
significantly help to reduce government deficits.' In almost every case, the 
public was more willing to sacrifice than the financial community believed it 
was. With respect to one alternative - reducing postal service to three 
days a week - the public at large was significantly more approving than 
was the financial community. 

However, the financial community was much more conservative than 
the general population with respect to the three major social programs that 
enjoy the greatest popular support - health care, family allowance, and 
post-secondary education. Regarding the first two, the financial com¬ 
munity respondents recognized that their opinions were different from the 
public's; however, they did not recognize the strength of public resistance 
to the idea of cutting student aid (see appendix). 


200 




TABLE 1 

Top two priorities of government 


Financial 

community 

1 

Canadian 

public 

% 

Assessment 
of public 
% 

Creating jobs 

18 

46 

96 

Stimulating industry 

70 

35 

13 

Lowering interest rates 

35 

59 

74 

Reducing government 
spending 

57 

34 

18 

Encouraging more 
foreign investment 

22 

23 

0 


TABLE 2 

Economic effects of reducing 

the deficit 




Financial 

community 

% 

Canadian 

public 

% 

Assessment 
of public 
% 

Help a great deal 

22 

21 

22 

Help somewhat 

35 

46 

22 

Mot make any difference 

9 

13 

9 

Hurt somewhat 

26 

11 

22 

Hurt a great deal 

4 

9 

17 


In summarizing these assessments by Bay Street of its own as well as 
Main Street's opinions, a number of points emerge: 

The Canadian public displays a significant fiscal conservatism with 
respect to the deficit. 

The public is willing to accept cuts in many spending programs; how¬ 
ever, this willingness does not extend to some of the major 'safety 
net' social programs. 


201 












In general, the financial community underestimates the fiscal con¬ 
servatism of the public. 

ISSUES AND ANSWERS ON BAY STREET 

One of the first things that becomes apparent after interviews with a 
group of people in a single industry or organization is whether, on a 
particular issue, there is a consensus, a division into distinct opposing 
groups, or a range of opinion that reflects diverse but not systematically 
divided views. When we asked the financial community the most basic 
question in our interview - 'Given the current economic situation, do you 
believe that the current deficit of the federal government should be 
increased or decreased?' - it rapidly became apparent that there was no 
consensus. Moreover, there was not even a division of opinion into dis¬ 
crete camps; instead, opinions were arrayed across a broad spectrum. 

As we have shown, there were significant differences in opinion 
within the financial community regarding the effect of a decreased deficit. 
That range of opinion extended to recommendations as to the optimal size 
of the deficit. The highest figure given was a recommendation that the 
deficit be increased by between $6 and $7 billion to a level of approxi¬ 
mately $35 billion. At the other extreme was a recommendation that the 
deficit be reduced to $10 billion in the coming fiscal year. Overall, main¬ 
taining the deficit at its current level was the most popular option; it was 
favoured by approximately one-half of the respondents. Approximately 
one-third of the respondents wanted it reduced, while one-sixth wanted in 
increased. 

In discussing this question, it became clear that one specific issue - 
the distinction between a 'cyclical' and a 'structural' deficit - has become a 
source of particular interest and concern in the financial community. The 
topic was mentioned spontaneously by about half of the respondents. Some 
of them even made a specific reference to a structural deficit as a full 
employment deficit. 

Without exception, the respondents who raised this issue believed that 
the federal government had, over the past ten to fifteen years, created a 
very sizable structural deficit. The estimates of the size of that struc¬ 
tural deficit ranged from $5 billion to 'nearly $20 billion.' Again without 
exception, the elimination of this structural deficit over the mid-term - 
usually three to five years - was accorded a very high priority for gov- 


202 


ernment. The existence of this structural deficit was often - cited as a 
reason for the federal government's not being able to take more vigorously 
counter-cyclical actions right now. 

Implicit in this description and critique of the government's actions is 
a distinctly Keynesian view of the world. Indeed, the vast majority of 
respondents in the business community were of the view that it is a gov¬ 
ernment duty - more particularly, the federal government's duty - to 
stimulate the economy during recessions by running deficits. However, 
what was particularly interesting were the exceptions to this generaliza¬ 
tion. While a broadly Keynesian view of fiscal policy has become - if you 
will - the accepted view in the financial community, the opposition to this 
view in the financial community comes largely from the economists in the 
community. 

In some question areas, there was no such diversity of opinion. 
Virtually without exception, respondents wished to see the scope of gov¬ 
ernment involvement in the economy reduced - or, at least, kept from 
growing. While government involvement was not seen as an immediately 
pressing problem, reducing that involvement was generally viewed as a 
major long-run priority. This view arose from a profound conviction that 
governments are not as efficient as the market in allocating resources and 
that their presence distorts market signals. One example, DREE grants, 
came up in several of the interviews as the epitome of government's swim¬ 
ming against the tide of market signals and decisions. Finally, in this 
same context, a link was sometimes made between the inefficiency of gov¬ 
ernments and their persisting structural deficits. 

The issue of crowding-out also produced a consensus among a large 
majority of the respondents. The following response was typical: 'It's not 
very close, but only because the economy is so weak and the demand is, 
relatively, so low.' Those who did not see crowding-out occurring may be 
divided into optimists and pessimists according to their estimation of the 
likelihood of its occurring if a recovery begins. The optimists believe that 
crowding-out may be avoided if the recovery is gradual, while many add 
that government must also rapidly reduce its demands on financial markets 
as recovery proceeds. Many respondents doubt that governments will or, 
indeed, are capable of, reducing their demands quickly enough. The 
pessimists, perhaps taking the large government presence as a given, 
assume that crowding-out will occur. A typical response was: 'Any type 
of economic upturn and the corporate sector will want to fund out the debt 


203 



that they have in the short end and fund it out to term ... then you're 
going to see it.' 

Whether or not crowding-out would occur with recovery, the question 
remains of just how large a deficit the Canadian bond market could handle 
without considerable dislocation or a very sharp increase in interest rates. 
The answers we obtained seemed to depend, in large measure, on the 
respondent's proximity to the bond market. The closer his involvement 
with the bond market, the more apocalyptic his response tended to be. 
The extreme view, held by a number of respondents, was that the bond 
market could not sustain current levels of government borrowing and any 
private sector demand. If the government's fiscal needs do not contract 
very quickly as recovery begins, then either the recovery will promptly 
die as a result of higher interest rates or the government may even run 
the risk of touching off hyper-inflation. 

As with so many of the topics covered, the crowding-out question 
elicited a very interesting minority opinion. According to some respon¬ 
dents, crowding-out was occurring already, in the sense that the high real 
rates of return in the bond markets were keeping corporate borrowers out, 
and that these rates were maintained by the federal government's massive 
presence in those markets. 

Further to the matter of the government's financing of its debt, 
respondents were asked whether the government should monetize any of its 
debt and whether this was taking place currently. The first question 
brought near unanimity: monetizing was seen as exceptionally harmful to 
the economy - far worse, for example, than crowding-out. As to whether 
it was occurring currently, there was an even division of opinion. 

Response to the question of monetizing deficits frequently led to 
discussions of trust in government. Although these discussions tended to 
be very wide-ranging, the following views could be considered dominant: 

Trust between government and business has declined sharply over the 
past twenty-five years; this has affected their capacity to undertake 
cooperative activities. 

The markets largely ignore the federal government's announcements of 
its intentions - unless the news is adverse for the markets. In 
essence, the markets believe bad news, but have an 'I'll wait until I 
see the figures' attitude towards good news. 

A quotation will encapsulate another aspect of the relationship: 


204 


'There's a tremendous waste of time and effort taken up by second- 
guessing governments.' 

Finally, the restoration of confidence in government is seen to be a 
very lengthy process that will only result from an ongoing display of 
'fiscal integrity.' 

Respondents saw the Bank of Canada's role with respect to confidence 
in government in quite ambivalent terms. Many respondents saw the bank 
very nearly in the role of accomplice in monetizing the debt, others saw it 
as a hope for independent control of the money supply. 

Given these sometimes conflicting views of the government's role and 
performance, what advice did members of the financial community have for 
the budget drafters? As we have seen, the responses as to the appro¬ 
priate level for the deficit ranged very widely, with the modal response 
being in the mid to high end of the $20-$30 billion range. A frequent 
theme was that higher deficits would have a perverse effect: they would 
frighten business, which would reduce its spending by an even greater 
amount; thus the result would be a contraction. 

There was also a tremendous range of suggestions for the specific 
forms that stimulation might take. Some respondents wanted to spur a 
consumer-led recovery by actions, such as sales tax reduction, that would 
affect consumer demand. Some wanted a range of industrial incentives, 
while others argued that such measures would be futile, given current 
capacity utilization. Almost all of the respondents who recommended a 
particular stimulative policy added the caveat that the program be limited 
in its duration. Sunset provisions were seen as vital to the creation of a 
more healthy economy. 

A very substantial minority of respondents argued that there was 
virtually nothing that the government could do in the short run. These 
respondents felt that the government could not - and probably should 
not - reduce its deficit in the short run. On the other hand, the debt 
markets were so strained that further expansion of the debt would have 
perverse consequences. This 'immobilism,' brought on by past fiscal 
irresponsibility, could only be cured through a longer-term commitment to 
more responsible fiscal policies. In essence, the financial community did 
not see any rapid solutions that government could bring to bear on 
Canada's economic problems. 


205 


CONCLUSION 


Where does all this put us in terms of the public's and the financial com¬ 
munity's expectations with respect to government? 

The public believes that the government can do a great deal to ame¬ 
liorate economic conditions, and that it can do it quite quickly; the 
financial community disagrees. 

The public, with a vision of the 'public household,' wants govern¬ 
ments to retrench in hard times. A plurality of the financial com¬ 
munity respondents believe that the government has a duty to con¬ 
duct a policy of counter-cyclical stimulation. 

Both groups attribute significant responsibility to past government 
actions, many of which are seen to be wasteful. 

Finally, there is very strong public resistance to major changes in 
some of the most basic - and most expensive - of the social policies 
set in place over the past three decades. 

This is hardly a picture to please governments: they are blamed for 
problems and charged with solving them, and yet the preferences for 
short-term policies and advice on them vary tremendously. On one thing 
both groups agree: the government is not an efficient instrument for 
ensuring economic growth. 

Governments may be in virtually a no-win situation. Further stimula¬ 
tion may be both dysfunctional and unpopular. Retrenchment, while 
popular with the public, may worsen the situation and be seen to threaten 
programs that are almost inviolable. 

It is hardly an easy time for either the economy or the government. 
The government, in seeking solutions, can only try to minimize the damage 
that it will suffer. Nonetheless, governments must act and, it is clear, 
that the public and the financial community support and expect initiatives 
that will lead in the longer term to leaner governments and lower deficits. 
Only the future will tell us whether governments can meet these demands. 

APPENDIX 

The Ontario Economic Council commissioned Decima Research Limited to 
sample opinions on government deficits among representatives of the finan- 


206 


cial community. This survey was conducted during the first two months of 
1983. The same questions were posed to financial community respondents 
as had been asked by Decima in a national public opinion survey (n=l,500) 
conducted in September 1982. The financial community respondents were 
also asked to estimate the public's response to each question. The precise 
wording of each question and a statistical breakdown of the responses are 
presented below. 


207 


4 

00 

G 

CO 

G 

00 

4 

*H 

•H 

-G 

4 

4 

44 

4 

4 


3 

O 

4-1 

o 

44 

O 

1—-t 

4 

4 

•H 

4 

E 

CO 

O 

O 

44 

E 

CO 

G 



4 

00 



G 

. 

X 

'H 

00 

G 

oo 

G 

44 

G 

'H 


4 

44 

-G 

G 

G 

00 

o 

4 

G 

4 

44 

o 

G 

G 

4 

4 

4 

-G 


4 

44 

4 

G 


O 

O 

>3 


4 

4 



44 

CO 

4 

CO 

4 

i-Q 

3 

CO 


GD 

G 

GD 

G 

4 

r-H 

■ H 

4 

3 


4 

o 

00 

G 

4 

c 

*H 


• 4 


44 

44 

4 

■rH 

G 

oo 


r—3 

G 


2 

u 

• H 
2 
3 

G 

o 

co 

0) 

•H 


3 

S 

•r-H 

2 

CO 


CO 


•H 
4 
O 

■H 2 
4 C 

2 i 


> 
- 4 
00 4 
d oo 

•H 

GD O 
G -H 
4 H 

Oh 

G 


00 

G 


2 

G 

4 

4 

4) 

<2 


a 4J 

^ *H 
OJ E 
> -H 
O rH 
00 

0-1 >, 
•H Ji O 
GD 00 
3 

2 o 

H—> O 
•H -G 

4-> 


00 

G 


2 
C 

4 

GD CO Oh 
4 O CO 
4 O 
G •!-) 

G 
4 


CO 


00 

G 

1 rH 

2 

G 

4 

4 

u 


4-1 

G 
4 
E 
C 

*"* CD 

> 3 

0 r—4 

oc o 

>n -H 
4 

4 _ 

> 2 
4 'g 

o 

4 


c_> 

1—I 

2 

PQ 

03 

Oh 


cj 


E-h 

!zi 

§ . ■ 

00 )-3 

CO tin CO 

w o oo 
co Co 

CO 

< 


03 >4 

< H 

i—i i—i 

CD 2 


t-4 o 
04 CO 


co 

00 

G 

•r4 


> -G 

O 

OO 

4 

00 > 

G 

•H o-i 
4 

3 <U 
GD C0 

4 4 

3-4 -C 
4-J 


3 

o 

G 

2 

3 

O 

0^ 


Oh 

O 


2 

G 

4 

E 


co 4 

< Oh 


CO 
4 
S-H CO 
4 4 
4-3 > 
G C 

•H *H 


<j- r—- cm om 
<N t—i CO '— 1 


00 <N 


CN CO O' LO 
CN 2 CN t—h 


lO CO 


4 03 03 O 

r~- 


o o 


CN 

CN 


-O m o 
2 


o o 


03 4 CM nT 
CM (N hJ 


O O 


03 co co 

--G r—I r —1 


CN 

CN 


CD 

2 

i—i 

Q 

2 


CD 


CO 

2 2 




CO 

2 

CD 





2 

2 

2 



2 

2 

44 





2 

2 

3—1 



H 

2 

a 





E-h 

2 

2 


>4 

< 


2 




t>4 

<C 


2 


2 

2 

E-4 

O 




2 

2 

E-h 

O 


H 


2 

2 




H 


2 

2 


CO 

H 

w 





2 

E-4 

2 



2 

CO 

2 

2 


/-3 


2 

2 

2 

2 


Q 

W 

z 

2 


2 


Q 

a 

z 

2 


Z 

2 

2 

O 


O 


2 

2 

2 

O 

CO 

1-4 

W 

2 

2 


> 

2 

i—i 

2 

a 

2 

CQ 


E-h 

> 


E-h 


CQ 


E*h 



O 

cd 

2 

O 

2 



O 

2 

2 

o 

2 

>“5 

2 

1-4 

CD 

2 

w 

2 

2 

2 

44 

2 

2 


1—4 



44 


O 


i—i 



4H 

o 

E-h 

CD 

CD 

2 

H 

i—i 

2 

E4 

2 

2 

2 

2 

< 

2 

2 

<C 

C/3 

2 

2 

< 

2 

2 

<£ 

1—4 

2 

(—i 

r-H 

2 

W 

44 

i—i 

2 

44 

i—i 

2 

H 


2 

cj 

2 

£> 

2 

E-h 

2 

2 

2 

5 

< 

G" 

CxJ 

2 

O 


O 

< 

2 

a 

2 

o 

W 

1-4 


Q 

2 

»—i 


2 

i—i 


Q 

2 

2 

H 

O 

fjj 

2 


O 

2 

E-h 

o 

a 

2 

CJ 

CO 

2 

2 

2 


2 

2 

2 

2 

2 

2 


H 

2 


hJ 

O 

> 


e-t 

CO 


2 

O 


CO 

O 


CO 


4 


4 

• H 


2 

2 


2 


O* 


4 

03 

4 

2 

TD 

2 

2 

03 

C- 


a 

GD 2 

4 

03 

t2 G 

2 

CJ 

G 4 



O <= 

2 

d 

2 a 

rH 

*H 

CO 4 

G 


4 

O 

2 

2 > 

2 

G 

G O 

CO 

4 

•H 00 


g 

2 

2 

c 

2 2 

G 

4 

O 

•H 

4 

G 

2 

> 

O >> 

2 

O 

>> 2 


oo 

•H 

G 


O 4 

O 

2 

GD O 

>> 

O 

•H 



2 4 

o 

>. 

4 2 

2 

2 

•H 


■r4 

2 GD 

2 

4 

3 C 

U 

O 

O 

•H 

•r—i 

GD 4 

2 

4 

G 4 


2 

<C co 

< 


P3 

rH 


t-H 


G 

•r-t 

4 

G 

•rH 

> 

o 

3-4 

Oh 


o 

Oo 

44 

4 

•H 

3-4 

G 

> 

G 

3-4 

O 

4-1 

to 

4 

G 

o 

E 

co 
4 
GO 
• rH 
> 
o 

3-4 

Oh 


G 

4 

E 

G 

4 

4 

> 

O 

oo 


G 

4 

4 

XJ 

4 

4-1 

4 

-G 


4 

E 


44 

G 

4 

CO 

4 

4 

Oh 

4 

GO 


44 

< 


C_> 

1—4 

l-G 

CO 

2 

04 


co 

E-H 



4 

ts 



E 

a 


CD 

£ 

s 


44 

G 

CO 


2 

4 

CO 

pH 

PD 

OO 

w 

o 

2 

o 

CO 


2 

4 

CO 



2 

< 



2 




C 




4 




g 



>h 

c 


< 

E4 

4 


HH 

2 

4 


CJ 

2: 

> 



2 

O 



p- 

oo 


5z: 




HH 

o 

rH 


Ph 

CD 


o o o 

O- i—i CN 


-d -G 


CO 


4- 00 -G 
r-H r— 


CM 


Do you think the amount of money which INCREASED 

the federal government gives to the DECREASED 

provincial governments to pay for post- NO OPINION (VOL) 

secondary education should be increased 
or decreased? 














cj 

2 

i-4 

PQ 

2 

O-i 


2 

z; 

a 


CJ 

2 


1—1 

on 


21 

on 

Pm 

P 3 

a 

o 

2 

CO 


Cu 

CO 

< 




2 >h 
< 2 

i—i i—i 

CJ 2 


2 O 
0-1 a 


T—1 Q> O 
r- CN 4- 


00 O'! <1" o> 
r-~ 


CM LD O 
CM SO 


cn 


o cn n- 
SO 1— 1 cni 


\0 co <f 
cn m 2 


-S’ 04 -S 
<r cn 


cn 





,—^ 




/~N 




2 




2 




O 




O 




> 




> 



2 

s_✓ 



2 




O 




O 




> 

a 



> 

a 



“w* 





2 




<0 




< 



2 

CO 



2 

CO 

Q 

Q 

O 


Q 

Q 

O 


a 

W 

2 

a 

W 

W 

2 

fxJ 

CO 

CO 

2 

E 

CO 

CO 

2 

2 

< 

< 

2 

£-i 

< 

< 

:—i 

2 

a 

a 

a 


a 

a 

a 


oo 

2 

o 

>4 

2 

a 

o 


a 

a 


< 

CJ 

CJ 


< 

2 

a 

o 

2 

2 

a 

o 

2 

t—i 

a 

2 

co 

2 

Q 

2 

CO 

3 

2 







2 

3 




2 



2 

'H 




3 


3 


O 




2 


2 

40 

3 

co 


2 

r -4 



CJ 

*r 4 

3 


3 

3 


2 

•H 

> 

3 


•H 

3 


r—i 

2 

o 

•H 



1 

2 

3 


2 

> 


5 

O 

O 

O 


a 

2 

O" 


co 


2 

>. 


3 

2 

in 


E 

co 

3 

3 

CO 

3 

3 

o 

2 


3 

2 


CO 

3 

2 

O 

co 

O 

2 

2 

3 

O 


2 

2 

E 


2 

3 

E 

co 


3 


O 

i2 

2 


3 

3 

3 

+4 

2 

3 

U 

2 

> 

43 

E 

O 


3 

3 

O 

•H 

2 

>1 


CO 

2 

2 


00 


3 

4 -S 

3 



2 


3 

a 

3 

> 

2 

2 

3 

2 

*H 


3 

•H 

O 

O 

3 

3 


3 

O 

00 

2 


O 

3 

co 

O 

E 



2 

E 

E 

3 

■H 

3 

2 

in 

3 

3 


U 

2 


3 

3 

CO 


2 

3 

3 

3 

3 

a 

3 

3 

3 

•H 

N 

40 

E 


3 

2 

> 

> 

•H 

2 


o 

2 

2 

O 

o 

2 


u 

2 

CJ 


00 

2 

3 

4*3 

3 


3 

2 


a 

3 

3 

> 

co 

•H 

3 

1—1 


a 

•H 

O 

2 


•H 

3 


3 

2 

CSC 

3 

3 

2 

2 

2 


2 


3 

2 

2 

3 

o 

t2 


i —4 

E 



2 

3 

3 

3 

3 

2 

2 

3 

3 

1 

3 

O 

2 

2 

p—l 

O 

2 

3 

O 


3 

3 

3 

2 


> 

•pH 


TO 

> 

O 


3 

3 

00 

o 

3 

O 

2 

O 

2 

2 

3 

Q 

2 

00 

co 

Q 

2 

*- 

2 

cn 




'S’ 





03 

0 ) 

co 

CO 

3 

S-I 

u 

3 

00 


TO 

0 ) 

CO 

co 

OJ 

S-i 


OJ 


3 


>1 

S~ 

CO 


S-I 

a 

OJ 

2 

2 

3 

> 

CO 

2 


O 

• > 



a 




>1 

• 3 



2 





43 



2 


OO 

CN 

O 

E 



PQ 


(N 

< 3 - 

2 

3 2 



2 





2 2 



a 




CO 

2 3 







3 

O 







3 

2 si 







2 

O CO 







3 

2 

2 







2 

2 ; 






3 

00 3 

CiJ 


CJ 




CO 

3 -H 

2 ^ 


2 




3 

•H 2 

CO 


2 

s-s 

o 


43 

>1 2 

CO 

Pm 

a 


CO 

m 

2 

3 

a 

O 

a 





a 3 

CO 


a 




2 

o 

CO 






O 

2 >4 

< 







O 







3 

O 







E 

>, 2 







O 

2 







CO 

•H O 


hd 

>4 





2 2 


< 

2 




2 

•H [5 


>—i 

2 




O 

2 


CJ 

2 





•H 



a 


r^ 

CN 

2 

W 


■<* 

s 


r-H 

CN 

co 

C rH 


2 

2 




•H 

O <5 


HH 

o 




pH 

a d 


Cm 

a 





CO 


TO 00 
(—1 CO 
03 OJ 
O U 


CO 03 
OJ 2 
U > 


>o| 
S-I 
CO 

E 

w o -H 


o 

-C 

3 

2 

3 

O 

40 


c 

o 

•H 

CO 

CO 

3 

U 

CO 


00 

C 


OJ 


S-i 

Oi 2 


O O 


CO 4-S 
OJ 3 
TO co 

4 -> 

2 £ 
« g, 

OJ S' 

^ -I 
00 H 

- <11 

CO 

c 

OJ 4 _) 
OJ T 3 
44 JO 

4- S 

co 

CO CO 
40 co 
OJ 

5- i 
CO 


3 - 

O 4-S 

>> a 

OJ 


o 

40 

3 


ea 

3 

u 

OJ 

> 

o 

oo 


}-l 

o 


OO 

3 


OJ 
S-I 
OJ , 

40 2 

2 O 

■> 3 

If 

3 C 

^ 3 
O 

3 1+4 

E 

° OO 
co — 


2 co 

3 -H 
CO O 
3 C 

•H -H 
S-4 > >> 

O o 3 

2 + a 
a , 

*> +4 

3 3 O 

" 5 >> 

—I 4-> 


4 

+J 


4-1 

3 


O £ 


CO 


S-i 

3 hj 

O > 

>> c 
00 
3 

4*3 r —4 

•H 3 
1 — 1 S-i 
3 

T 3 TO -H 
- 3 S-i 

t-H +4 a 


O 

a 

co 

3 

3 

>1 

S-I 

3 

c 


S-i >s TO 3 3 
O 3 3 40 40 

f *4 04 3 4 -> 4-1 


'O’ ■'O' CO 
CN 


<f -O’ <f 


in 


OJU 

3 ^ 

S-i 


2 


2 ^3 

o 


2 


O 4J 



a 


00 

S -4 

H 

2 


W 3 

o 

2 

2 


= > 

2 

a 

a 



U 

a 

a 


3 

3 


> 

a 


CO 

§ 

o 

o 

2 


a 

o 

2 

i2 

3 

> 


a 

3 

2 

o 

a 

a 

• O 

3 

o 

< 

co 

>. 2 

> 


i—i 


2 CO 

•H 

a 

CJ 

a 

3 

H 

< 

2 

2 

•H 2 

O-i 

a 

i—i 

< 

3 3 


a 

> 

> 

O 2 

CD 

a 

O 

2 

co 2 


a 

a 

a 

2 

M> 

a 

a 

a 


C” 

3 

O 

• pH 
4-S 
3 
3 
3 
TO 
3 

3 

00 

3 


O 

U 

TO 

3 

3 

>. 

4-1 

•H 

co 

3 

3 

> 


m 


INDIVIDUAL 
NO OPINION (VOL) 















2 

PH 

2 

2 


2 



LD M3 CM lO O O CM 

co ld 


co on O LO CO 
CO CO CM 


in n vfi + sf 

<r + 


ON O' H N o 'J 
CO CM CM 


H 



co 2 oa 

WOP 
CO 2 
CO 
< 


N O' O O M O M 
LiO CO 


h c> <3* 
»~h 


'O vO O O CO <f 

LO CN] r-H 


L-T) P"- CO O O <t 

\sQ rH H 


2 

< 


2 



2; s 

PH O 

2 cp 


4Mnoi't<t o 

r-H CO 


P N N O M 
CM I-H LO 


O' O M M O' O 
Mf CO 


in CM ON O M o 
MO CM 


. 


E-i 




2 



2 





E-h 






E-h 





• 


2 




< 



2 





2 






2 





• 


2 




2 








2 






2 






H 

2 



2 

Q 


E-h 

s 




2 

2 



2 


E-h 

§ 



2 


Pi 

2 

2 



< 

i — i 


2 

2 




2 

2 



< 


2 

2 



< 


O 

2 

2 



i — i 

> 

/—s 

2 

2 



r—■> 

2 

2 



PH 

/—V 


2 



PH 

/—\ 

PH 

*5T| 

2 



2 

PH 

2 

§ 

2 



2 

2 

2 



2 

2 

§ 

2 



2 

2 


Z 

> 

2 


2 

Q 

O 

z 

> 

2 


O 

z 

> 

2 


2 

O 

2 

> 

2 


2 

O 

OP 

s 

O 

O 


i — i 

2 

> 

2 

O 

O 


> 

2 

O 

O 


PH 

> 

2 

O 

O 


PH 

> 

G 

2 

CP 

H 


> 

i—i 

p—-- 

2 

2 

E-I 



2 

2 

E-h 


> 

v— y 

2 

CP 

E-h 


> 

p^- 

•rH 

> 


CP 


O 

—. 


> 


2 



> 


2 


o 


> 


2 


o 


>l 

o 

2 

2 

2 

2 

2 

2 

o 

2 

2 

2 

2 

O 

2 

2 

2 

2 

2 

O 

2 

2 

2 

2 

2 

G 

o 

< 

2 

< 

2 

< 

O 

2 

< 

2 

< 

O 

2 

< 

2 

< 

2 

O 

2 

< 

2 

< 

2 

O 

2 


i—H 


2 


i — i 

PH 


PH 


2 

PH 


PH 


a 


PH 


i — i 


a 


PH 


2 

CP 

2 

Q 

2 

2 

2 

2 

2 

2 

a 

2 

2 

2 

2 

a 

2 

2 

2 

CP 

2 

a 

2 

2 

P-l 

< 

2 

H 

PH 

< 

2 

PH 

<C 

2 

E-h 

PH 

PH 

<£ 

2 

E-h 

PH 

< 

PH 

<c 

2 

E-h 

PH 

< 

PH 

o 

2 

i—i 

< 

> 

2 

PH 

2 

2 

PH 

< 

> 

2 

2 

PH 

< 

> 

2 

2 

2 

i — i 

< 

> 

2 

2 


2 

> 

> 

1—1 

2 

> 

O 

2 

> 

> 

PH 

O 

2 

> 

> 

PH 

2 

O 

2 

> 

> 

PH 

2 

O 

>. 

Q 

O 

i—i 

Q 

Q 

o 


Q 

o 

PH 

a 


a 

o 

1 — 1 

a 

a 


a 

O 

PH 

a 

a 


4-> 

2 

2 

2 

2 

2 

2 

o 

2 

2 

2 

2 

O 

2 

2 

2 

2 

2 

O 

2 

2 

2 

2 

2 

O 

•H 

2 

2 

2 

PH 

2 

2 

2 

2 

2 

2 

PH 

2 

2 

2 

2 

i—i 

2 

2 

2 

2 

2 

PH 

2 

2 


•H 


a 


•H 


W 


G 


O 


2 


w 


<D 


in 




Sh 


G 


E 


•H 


Sh 


2 


CJ 


a 


a 


<u 

er¬ 

> 

as 

g 

Sh 

a 

G 


U 

TO 


r-H 

i—1 

G 

G 

O 

U 

a 

•H 

w 

2 


as 

2 

E 

G 


•1—1 

TO 

a 

G 

a 

G 

G 

r* 

O 

a 

>i 

r-H 


CO 

o 

CJ 

TO 


O 


a 

• 

2r 

VhO 


Cr- 

OP 

G 

-rH 

G 

-H 


G 

Pi 

4-> 

0) 

Sh 


pa 

o 




C~ 

0) 

Pi 

G 

4h 

r-H 

OJ 

5 


o- 

w 

w 

<u 

G 

-rH 

w 

G 

-G 

P 

O 

2 

w 

0) 

> 

•1—1 

4-1 

c 

<D 

CJ 

G 

■H 


00 


ON 









X 

Z 


x 

I—I 

X 

« 

x 

z 


x 


CO X 

co x m 
WOP 
CO Pm 

CO 

< 


r-' 

m cm 


X x 00 O CO 


<T 

O' 


co o o os 


M3 

CM 


r-'' as as 

m 


PS 

o 

X 

X 


c^- 

M 

3 

3 

r*H 

Z 

a 

o 

'H 

M 

3 

CJ 

Z 


E-i 



x 


. 


Eh 





Eh 






X 




Z 



CO 


• 


z 





z 






z 




Cl} 





• 


x 





X 






X 




s 



x 



Eh 

s 




Eh 

5H 



X 


Eh 

s 



X 




Eh 


p 

Z 

z 




z 

z 



< 


Z 

z 



X 

p 



< 


o 

PlS 

z 



/■—s 

g 

z 



X 

/-V 

X 

z 



< 

x 



> 

X 

X 

§ 

x 



X 

x 



X 

X 


X 



> 

> 

z 


X 

O 



> 

z 


o 

> 

z 


z 

o 

Z 

> 

z 


1—1 

o 

o 


Z 

> 

00 

z 

o 

o 


> 

z 

o 

o 


X 

> 

Z 

o 

o 


z 

o 

X 


PL, 

s —s 

3 

x 

o 

X 


s- ' 

Cx3 

X 

X 


> 

s-✓ 

x 

X 

X 


z 


CJ 




•H 

> 


X 





X 


o 


> 


X 



X 

x 


X 

Z 

in 

o 

X 

X 

X 

z 

o 

X 

X 


z 

z 

o 

X 

X 


X 

< 

CO 

< 

< 

O 

3 

a 

< 

CO 

< 

o 

X 

< 

X 

< 

z 

o 

o 

< 

X 

< 

< 

X 


p 

X 

X 

Z 


X 


X 

X 


X 



--s* 

X 


X 


£3 

X 

CJ 

x 

a 

a 

Z 


X 

X 

x 

Q 

z 

X 

X 

X 

Q 

X 

Z 

X 

X 

X 

Q 

X 

z 

Eh 

i—i 

X 

X 

X 

<; 

z 

Eh 

X 

X 

<£ 

z 

Eh 

)—i 

< 

X 

< 

z 

X 

l-H 

z 

X 

< 

> 

> 

Z 

O 

z 

X 

< 

> 

Z 

z 

X 

< 

> 

z 

Z 

z 

X 

< 

> 

1—1 

> 

> 

i—i 

X 

O 


x 

> 

> 

X 

o 

x 

> 

> 

HH 

X 

o 

X 

> 

> 

hH 

> 

o 

X 

Q 

Q 


>s 

a 

o 

X 

Q 


Q 

o 

X 

Q 

Q 


a 

o 

X 

p 

o 

z 

Z 


§ 

O 

X 

x 

z 

Z 

Z 

o 

w 

z 

Z 


X 

o 

X 

z 

z 


z 

Cu 

z 

HH 

X 

z 

•H 

x 

z 

z 

X 

z 

X 

z 

Z 

HH 

X 

z 

X 

z 

z 

hH 

z 


•H 

X 

■ I—I 
M 
C 

o 

z 

w 

<d 

p 

Jo 

p 

3 

e 

•H 

P 

a 

CD 

x 

X 

<u 

> 

CJ 

X! 


3 

O 

x 

w 

X 

c 

•H 

X! 

X 

3 

O 

to 

o 

t: 


x 

)—i 

XJ 

CQ 

33 

PU 


X CJ 
ISO i -1 
CO XI 
CO Cm CQ 
X O 33 
CO z 
CO 
< 


XI t* 
< X 
x i—t 
CJ Z 


30 S 
oh O 
X cj 


r—i r^- -O' o- 
M3 CM 


CM 


NM O O W 

O0 


00 ^ 'f 


OS o 


0" 

OJ 

u 

3 

3 

Pi 

3 

W 

c 

•H 

X 

3 

3 


PW 

E 

CD 

3 

3 


00 O CO M3 O CO 

<1- -a 


x o o <r o <r 

X) CO 


m <f c^ w <f o 
co <r 


o- 

in 

r-H 

p 

CD 

33 

iX 

<d 

CD 

XI 

X 

p 

o 

X 

00 

3 

•H 

P 

3 

CJ 


CM 


oo x in x in 
<r co 


CM 

X 


CM 

CM 


o as 


x o o mj- o- 

CM CO CO 


3 

CD 

£ 

3 

O 

P 

*H 

> 

3 

3 

CD 

X 

X 

00 

3 

-H 

X 

CJ 

3 

X 

O 

P 

Z 


CO 













CJ 
I—( 

P 3^ <N lD <N1 
CQ r-- cm 

p 

p 


i—i p cm i—i <j\ o 
CM i—' i—i 


o 1/1 in w rt 
CM MO 


H 

Z 

Eg 


CJ 

— 

cm 


p 

cm 


CG 

w 

o 

P 

cm 


Oh 

cm 

< 


p 

>h 


< 

H 


1—1 

i—i 


CJ 

z 

< 

Z 


z 

z 


1—1 

o 


cp 

CJ 




O O O 

o 


n tn o 

00 T—* 


M cm ^ M cn 

CM CM CM 1—1 


N in W ^5 4 ON 
CM CO CM 


ON N \0 O ON 

uo cm 


O 00 CM O O 
O' CM 


CP 

CJ 

z 

w 

K 

w 




p 


fp 


P 






/ - N 

< 


p 


< 

/— s 



/—s 


P 

w 


1—1 


W 

p 


cm 

p 


o 

Q 


Q 


Q 

O 

cm 

p 

o 


> 


H 


E-i 


> 

w 

CJ 

'-'1 > 



H 

< 

>< 

< 

H 


X 

i—i 

p ^ 



< 

jx: 

z 

bq 

< 


< 

> 

o 

w 

z 

w 


< 


W 

z 

E-i 

p 

c-v > Z 

> 

o 

2 

W 


w 

P 

o 


P 

P w o 

o 

t—1 

o 

T" 

w 

z 

o 

1—1 

P 

cm 

O P 

p 

z 


o 

2 

o 


z 

cm 


> P z 

p 

1—1 

< 

C/3 

< 

cm 

< 

l—l 

< 

p 

w P 1—1 

p 

p 



z 



p 

w 

CJ 

p p 

< 

o 

p 

P-4 


H 

H 

o 

p 

P 

P P o 

cm 


p 

HP 

H 

PS 

P 


CJ 

P 

H i—i 

i—i 

o 

w 


C 

P 

CJ 

o 

z 

p 

O P O 

a 

z 

3 

X 

z 

P 

P 

z 

1—1 

P 

CO Z Z 


oo 


d 

u 

d 

d 


00 



P 

p 


d 














P 

p 

d 





-X 

. 


d 

TJ 


d 

d 


P 




ip 










d 

d 

d 

S'- 


1 


d 

• 


•H 

0) 

P 

d 

d 

d 

d 




p 




d 






d 

<p 

p 

CO 

P 

p 


•H 

• 


i—! 

o 

O 

> 

CJ 

o 

"O 


d3 


d 

p 



•H 


u 




P 

p 

oo 

p 

d 

u 

co 

P 

CO 


d 

d 


P 


x 

d 


d 


d 

d 


p 


• 

•H 

d 

p 


oo 

•r4 


•H 

d 


d 

P 

P 


d 

TO 

d 

d 

p 


p 


d 


CJ 

p 

X 

d 

CO 

d 

E 

p 

•p 



p 

d 

(J 

p 

*H 

(j 


■H 


’V 

o 

N 

CO 

d 

-d 





d 

p 

d 

p 

P 

•i—i 

O 

p 


p 

d 



•i—! 

d 


•rH 


U 



p 

•pH 


P 

i—i 

d 

d 

X 


p 


p 

p 

P 

d 

d 


<p 

d 


x 

X 

P 

p 


> 

O 

•rH 


p 

p 

CO 

O 


d 

p 

oo 

•rH 



p 

P 


CJ 

oo 

o 

d 

•i—i 

o 

Q- 

d 

E 

d 

p 

p 

P 


P 


o 

d 


•rH 

d 

o 

p 

d 

CO 


d 

u 


>> 

•H 

d 

CJ 

p 


X 

r—4 

d 

o 

T3 


• 

d 


d 

O' 


•rH 

d 

rH 

> 

3 


3 



o 

d 

d 

d 

P 


d 

d 

d 


d 


d 


"O 

• 

CO 

o 

’V 

co 

X 


P 

P 

•rH 


p 


O 


•H 

p 

d 

co 

d 

00 


r —1 

d 

d 

p 

d 

o 

P 


• 


XS 


d 

d 

P 

P 

d 

P 

•- 

o 

00 

P 


CO 


p 


d3 

d 

p 

3 

> 

i—i 


P 

u 

•H 

o 

co 

d 


p 

CJ 


0) 


P 


oo 


d 



CO 

d 

p 

d 


*i—i 

oo 

E 

d 


d3 

d 

d 

d 

3 

P 

P 

X 

•rH 

•rH 


00 

P 


dJ 

d 

d 

•H 

CO 


d 

P 


rp 

p 

3 

d 

•H 



i—i 

E 


P 

P 

•rH 

P 

d 

d 

> 

CLI 

-o 

•rH 

ip 

d 

•H 

> 

4-) 

p 


CJ 

p 

X 

Cd 

•rH 

o 

o 

P 


. 

d 


d 

P 


u 


3 

p 

p 

d 

d 

E 

d 

d 

P 

o 


d 


CO 


d 

o 

d 

p 

p 

CO 

p 

CO 

o 

P 

p 


oo 

•rH 

00 


p 

d 

o 

p 

•rH 

p 


d 

P 

E 

d 


■H 


d 

d 

p 

00 



p 

p 

3 

o 

p 

d 

d 

P 

d 

P 


CO 



pp 

d 

P 

d 

P 

'H 

> 


dO 

CO 

o 

CU 

p 


d 

d3 

3 

•H 


d 


CJ 

•rH 

d 

•rH 

U 

d 


CO 

d 


T3 

d 

P 

P 

P 

p 



p 

E 



>1 

*H 

Ip 

o 

CJ 

P 


p 

3 

P 

"O 

u 

•rH 

u 

d 

<8 

d 

T3 

0) 

d 

CO 

d 


d 


p 

•l—i 


d 

d 

E 


3 

p 

•H 

•H 

— 

p 

-o 

d 


r-* 



CJ 


d 

ip 

<p 

X 


co 

d 

CO 


o 

CJ 

d 

E 

u 

o 

p 

o 

oo 

<P 


p 

3 

d 

X3 

P 

"O 


"C 

3 

•> 

P 

d 



X 

•H 

E 




•H 

P 

o 

3 

d 

d 

p 


d 

p 

d 

P 

P 

•rH 

•rH 

d 


d 

TJ 

5 


o 

P 

co 

I—1 

TO 

E 

CJ 


p 

P 

O 

cm 

'd 

o 

P 

CO 

d 

'3 

d 

p 



CO 

d 

P 


p 

d 

o 

0/ 

2 

o 

P 

p 


d 

'rH 


o 

d 

E 


d 

CJ 

P 


00 


•H 

3 

P 

co 

d 

P 


• 


p 

d 

p 


p 

P 

d 

P 

P 

rp 


Ip 

d3 



p 

d 


CO 

d 

d 

p 

d 

o 

P 

o 

P 

>. 

d 



_x 

p 

p 


P 

S-i 

O 

00 

P 




do 

• 



»• 

p 

p 

00 

p 

E 


d 

CJ 


P 

’■a 

P 

p 



u 

CO 


d 


O 

d 


d 

P 

d 

CO 

P 

d 

X 

d 

d 

p 


O 

p 

d 


oo 


•H 

d 

o 

o 

— 

*H 

d 

P 

d 

d 

P 

P 



d 

d 

d 

co 

p 

d 

d 

o 

d 

d 

CO 

d 

E 

d 

d 

P 

> 

o 


CO 

d 

P 

co 

d 

d 

> 

P 


c~ 

d 

d 

Ph 

X 

3 

p 

co 

E 

CJ 

p 

o 


p 

d 

p 

■rH 

o 

O 

p 

CO 

rH 

o 

3 

d 

d 

a 

O 


p 

p 

d 

E 

CO 

d 

E 



d 

d 


X 

p 

3 

p 

d 

CJ 


P 

CO 

d 

03 

•rH 


d 

u 


p 

co 

o 

d 

d 

d 


p 


p 

d 

p 

d 

oo 


•H 

d 

d 


d 

CO 

a, 


X 


P 

d 

p 



p 

- 

p 

d 

P 

p 

co 


CO 

d 

i-— i 

d 


d 

o 


E 

> 

CO 

T3 

d 


co 

d 

x 

d 

£ 

u 

p 

• 

p 

p 


CJ 


d 

p 

d 

p 

00 

cu 

> 

o 

•H 

83 

CU 

o 

o 

p 

d 

X 

X 

P 

p 

d 

rp 

£ 

d 

d 

p 

d 

d 

CO 


co 

> 

d 

•H 

d 


o 

c 

p 

> 


i—H 

CO 

00 

d 

p 

d 

d 

d 


£ 

P 

d 

•H 

P 

d 


d 

d 

p 

d 

o 

d 


d 

o 

d 

oo 


d 

rv 

d 



d 


P 

p 

d 

d 


d 

P 



d 

d 

E 

CO 

d 

P 

00 

E 

p 

£ 

p 

Oh 


>> 

P 

d 

p 

P 

p 

E 

p 


P 

E 

CO 

d 

•rH 

00 


X 

X 

o 

d 

d 

p 



d 

u 

2 



— 

E 


o 


•rH 

•rH 

d 

o 

oo 


d 

d 

o 


o 

P 

•H 


>1 

p 

d 


d 

p 

p 

d 

p 

p 

p 

CO 

o 

co 

•H 




p 


d 

P 

p 

d 

>> 

P 

p 

d 

W 

p 


d 

p 

d 

p 

d 

d 

I—1 

d 

d 

d 

d 

d 

d 

d 

ip 

P 

iP 

d 

co 

•H 

CJ 

d 

p 


P 

p 

I—I 


X 

X 

> 

CJ 

> 

d 

o 

> 

r—1 

> 

d3 

p 

E 

o 

d 

•rH 

d 

P 

d 

> 

>. 

W 

•H 

> 

CJ 

cO 

•rH 


03 

o 

d 

d 

o 

d 

•H 

P 

p 

o 

o 

o 

P 

p 

• H 

CJ 

d 

Cu 

d 

3 

d 

o 

d 

03 

p 

o 

d 

< 

5 

d 

w 

4-> 

d 

CO 

00 

•H 

p 

E-i 

d 

00 

CJ 

oo 

o 

o 

4-J 

d 

E 

o 

■o 

P 

T3 

o 

3 

CJ 

3 

00 

•rH 

<3- 










LO 














p 


















Now, I'm going to read you a list of some of the ways in which governments could cut their spending in order to reduce 
their deficits and I'd like you to tell me whether you would approve or disapprove of each one, if making that spending 
cut would significantly help to reduce government deficits. How about ... Would you approve or disapprove of 
government doing that in order to reduce deficits? 


CJ 

i—i 

01 

PQ 

Oh 


cm m m 

CM 


E-i 

55 



co 04 pq 
w o p 

CO 0-i 


% 


CO CM '3' 

m 


r-~ co o 
m <f 


i-O 

o 

> 


O HH 

u K 2; 

1> D-t i—i 
O 04 0-1 

05 < O 
0-1 CO 
0-1 H o 
< Q 2 


c- 
0) 
r—I 

4-» 34 
0 } 3 
-3 O 

4-1 TO 
O T3 

CO r—l 

3 

co o 
3 5 
•H 

T3 3 
•H 3 
CO CM 
pQ 

3 >> 
CO 5-1 
3 

T3 T3 

C 3 
3 O 
O 

T3 3 
•h co 
3 I 

4- > 
4-> CO 

3 O 
3 a 
TS 

3 3 
+-> CxO 
CO 3 

5- 1 

on 3 

3 > 

•r-l 3 
4-5 

4-5 3 

3 -3 

3 4—5 


r-~ 



o 

MT 

34 

m 

CM 

CO 

CM 

CO 

CM 

r— 

CD 00 


m 

<r 


m 

'3- 


I—H 

00 


co 

m 


H 

m 


co 

00 


CO 

00 

CTc 

<r 

cm <r 


co 

co 


H 

r-~ 


1 — 1 

r-- 



m 


m 

m 

O 

<r 

r-'- 

o 

o 

o 

O 

m 

m o 


cd 

CO 



m 


CO 

r- 


co 

CD 












y — \ 




i-3 



1-3 



m5 


hJ 




O 



O 



O 


o 




> 



> 



> 


> 







h_/ 





w 




2 


w 

53 


W 

2 


W 2 



> 

O 


> 

O 


> 

o 


> o 



o 

3-1 


o 

i—i 


o 

5—1 


O 5-1 


w 

05 

2; 

w 

05 

2 

w 

05 

2 

pj 

05 2 


> 

04 

34 

> 

O-i 

i—i 

> 

Oh 

5—5 


0-1 5—5 


o 

04 

04 

o 

04 

Oh 

o 

Oh 

Oh 

O 

Oh O-i 


05 

< 

O 

05 

< 

o 

05 

< 

o 

05 

< O 


CL, 

CO 


Oh 

CO 


Oh 

CO 


Oh 

CO 


04 

I-H 

o 

O-i 

5—1 

o 

Oh 

5—1 

o 

Oh 

5-5 O 


< 

Q 

2 

< 

a 

52 

< 

Q 

2 

< 

Q 2 












C" 


3 










>c 


r-H 










3 


a 



3 







4-> 


o 

4-> 









co 


3 

3 


CO 







3 


CL 

3 


Jo 






-o 

TD 



E 


3 






3 

3 


cm 

>» 


T3 






3 

•H 


O 

o 












i—H 


3 






35 

3 


5-1 

3- 


3 



C-- 



3 

4-5 


3 

£ 


3 



3 



3 

3 


-Q 

3 


35 



3 



3 

> 


E 

3 


4-5 



3 



3 

•M 


3 

3 





3 



co 

3 


3 



O 






3 

CL 



3 


4-1 



-3 



3 



3 

O 





4-4 




to 


35 

CH 


3 



!“1 



3 

35 


4-1 



3 



3 



O 




3 


•H 



3 



CM 

T3 


on .—I 


> 



35 




3 


3 

-a 

(>• 

3 






4-> 

4-5 


•H 


CO 

3 



O 



3 

3 


3 

on 

4-5 

CO 



4-4 



O 

3 


3 

•H 

•H 







Oh 

T3 


TO 

rH 

cm 

r—l 



co 



Oh 

3 


3 

3 

3 

3 



co 



3 

O 


3 


3 

4-4 



3 



co 

3 



3 

3 

CO 



3 






to 

3> 

30 

o 



3 



4-> 

4-> 


i—1 



Oh 



3 



3 

3 

. 

i—t 

TO 

3 







o 

3 

. 

3 

,—1 

3 

on 



on 




E 

• 

3 

3 

3 

3 

e- 


3 



on 

Oh 


•H 

o 

3 

•M 

35 


•H 



3 

o 

4-4 

4-1 

5 

3 

3 

3 


3 



•H 

i-H 

3 

CO 


3 

3 

3 


3 



4-1 

3 

O 

3 

o 

CO 

T5 

3 


T5 



4-4 

> 

34 

3 

35 

3 

3 



3 



3 

3 

3 

T5 

3 

•M 

3 

3 


3 



3 

TD 

£ 

. 



. 



. 



. 


O 

00 






O 



r-H 


w 

r—l 



r—t 



CM 



eg 



i-h m 
cm r~ 


-j - h 
ON 


r-> O 

m 


22. cancelling all family allowance payments? APPROVE 

DISAPPROVE 
NO OPINION (VOL) 















CJ 


2 

pa 

O 

a, 


o lD lT> 
CO 2 


H 

2 



co 2 pa 
w o 2 

CO 2 

CO 

< 


r" <f 

co 


2 >H 

< H 


c_> 


2 


2 



H ^ o 

2 CO 


2 

o 

> 


£ 

o 

2 

2 

2 

< 


2 2 
> O 
O I—I 
2 2 
2 2 
2 2 
< O 
CO 


2 o 

Q 2 


G 

o 


G 

<U 

u 

u 

2 

o 


o 

>» 


G 

U 

•H 

2 

CO 

G 

G 

TO 

cc 


T3 

G 

<U c~- 
2 CO 
CO QJ 

E 

00 £ 
C G 
•H 

2 00 

2 O 

3 G 
U 2 


CO 

CN 


O uO uO 

CO 'G 00 

O CM 2 

-G lD 

03 2 

U0 -G 

2 O <t 

CO CO G 

CM -G <f 

CM I-" 

O' 2 

LD -G 

<t O' o 

O G O 

2 a> O 

vG m 


2 CO 


/—\ 


2 

2 

2 

O 

O 

O 

> 

> 

> 

v—' 


w 

2 2 

2 2 

2 2 

> O 

> O 

> O 

O 21 

O 2 

O 2 

2 2 2 

2 2 2 

2 2 2 

>22 

>22 

>22 

0 2 2 

0 2 2 

0 2 2 

2 < O 

2 < O 

2 < O 

2 CO 

2 CO 

2 CO 

2 i—i O 

2 2 0 

2 2 0 

< Q 2 

< Q 2 

<C Q 2 

c- 



co 



<u 

S-i 

QJ 

u 

o 

2 

' rH 

2 

2 

> 



5-1 

00 

5-1 

<U 

c 

O 

CO 

*H 

2 


2 


>> 

5-1 

2 

S-i 

O 

5-1 

U 

3 

O 

QJ 


2 

2 

<u 

2 


r-H 

G 

XJ 

2 

co 

C 

O 


G 

CD 

r-H 


2 

G 

t2 


■H 

•H 

2 

U 

G 

O 

G 

5-1 


G 


5-1 

G 

5-1 

QJ 

*H 

<U 

2 

2 

00 

£ 


G 

3 C— 

(—H 

qj 

C 2 

i—1 

co 

G 

G 

co 

QJ QJ 


G 

2 £ 

00 

2 

2 G 

C 


5-1 

*r"i 

00 

00 QJ 

2 

G 

c > 

G 

■H 

2 O 

C 

CJ 

v oo 

• M 

G 

G 

E 0" 

X5 

T3 0) 

•H CJ 

0) 

<U 2 

2 CC 

5-1 

5-1 2 

0) CJ 

<G 

lO 

2 

CM 

CM 

CN 


2 

G 

QJ 

U 

S-l 

CD 

2 

O 

o 


E 

O T3 

G 

O QJ 

to 

LO XJ 


- G 

2 

2 O 

O 

II U 

G 

C QJ 


5-1 

> 

U 

G 

•H 2 

E 

2 O 


2 G 

co 

G 

QJ 

2 QJ 

CO 

5-1 

c 

- QJ 

o 

co 3 

2 CM 

CO 

II co 

QJ 

G QJ 

5-> 

CO 


>. G 


2 O 

oo 

•H G, 

G 

G co 

■ H 

3 OJ 

xj 

£ S-f 

G 

£ 

G 

O CO 

O 

U 0) 

5-1 

2 


2 G 

O 

G U 

2 

•rH *H 


CJ 2 

00 

G G 

G 

G *H 

•H 

G 


■H + 

o 

2 + 


co 

w 

H 

O 

2 












Fiscal policy ‘crowding-out’ of private 
investment in an open economy: 
the case of Canada 

R. G. Wirick* 


Analysis of possible channels for fiscal policy 'crowding-out' of private 
investment in a small, open economy is significantly different from, and in 
many ways substantially simpler than, the analogous task for a closed (or 
large, open) economy. Both the difference and the relative simplicity 
arise from the fact that in a creditworthy country such as Canada, gaps 
between domestic savings and investment can be financed by virtually 
unlimited international capital flows at an exogenous world interest rate. 
Indeed, the standard textbook model of a small, open economy, drawing on 
the work of Fleming (1962) and Mundell (1963), indicates that when inter¬ 
national capital markets are well-integrated, stimulative fiscal policy will 
have no adverse effect at all on domestic investment . 1 

In a number of ways, the Fleming-Mundell model is simplistic. For 
example, in keeping with its Keynesian roots, it incorporates the assump¬ 
tion of a fixed aggregate price level. Moreover, it is a (timeless) compara¬ 
tive static model that ignores all dynamic effects stemming either from 
price, output, or exchange-rate expectations or from asset accumulations. 
Yet despite these limitations, the Fleming-Mundell analysis is an exceeding¬ 
ly useful point of departure for an examination of the effect of fiscal 
policy on investment. Its conclusions on this particular issue are relative¬ 
ly independent of the restrictiveness of its macroeconomic assumptions; 
they rest principally upon the very basic idea that profit-maximizing 
arbitrage behaviour will assure egual returns for equivalent financial 
instruments. 

The methodological approach of the present paper is to examine this 
arbitrage behaviour in some detail, with particular emphasis on the specific 

* Assistant Professor, School of Business Administration, University of 

Western Ontario 


215 




circumstances facing Canada. The bulk of the analysis assumes that fiscal 
policy crowding-out of private investment occurs if (and only if) fiscal 
policy results in an increase in borrowing costs to Canadian corporations. 2 
Consideration is given to possible influences on both long-term and short¬ 
term borrowing costs. Throughout the paper, United States inflation and 
interest rates are considered to be unaffected by any change in Canadian 
fiscal policy. 

LONG-TERM BORROWING COSTS: AN ANALYTICAL PERPSECTIVE 


Consider the following definitional indentity: 




X CG^ 


(i 


CG 


l CG^ + ^CG " ^SG^ + ^SG' 


( 1 ) 


where 


: CC 

: CG 

* 

1 CG 

* 

X USG 


the interest rate on long-term Canadian corporate bonds, 
issued in Canadian dollars; 

the interest rate on long-term Government of Canada bonds 
issued in Canadian dollars; 

the interest rate on long-term Government of Canada bonds 
issued in United States dollars; 

the interest rate on long-term United States Treasury 
bonds, issued in United States dollars. 


The question of whether stimulative fiscal policy raises long-term nominal 

corporate borrowing costs, Iqq, then depends on whether such action 

increases one or more of the interest rate differentials of equation (1). 

* 

(Historical movements in iygQ and i CG are given in Figure 1.) The use¬ 
fulness of such a taxonomic approach lies in the fact that each of these 
differentials is determined by quite separate economic influences. 

Alternatively, it could be argued that the appropriate method of 
measuring borrowing costs is to analyze real, not nominal, interest costs. 
Equation (1) can be rewritten to facilitate such an analysis. Specifically, 


^ r CC + n C^ ^CC 


* * 

: CG^ + ^CG ' 1 CG^ + ^CG 


X USG^ + ^ r USG + ^US^ 


or 


216 



Figure 1 

Long-term bond yields: domestic Canadian corporate versus United States 
Treasury bonds, 1948-82 


Per cent 



McLeod, Young, Weir 
10 Industrials 


U.S. Treasury 
long-term bonds 


Canadian corporate 
minus U.S. Treasury 
differential 


SOURCE: CANSIM. 


r cc (l cc 

*CG^ + ^CG *CG^ ( "C "US^ + ^CG 

‘USG 5 + 

* 

r USG' 





(2) 

where 





A A 

"C' "us 

= the average expected rates of inflation 

in Canada and 


the United States, 

respectively, over the bond 

matur- 


ity period; 




r cc 

= the real interest 

rate on long-term 

United 

States 

* 

government bonds 

*CC " "c ); 



r USG 

= the real interest 

rate on long-term 
* 

United 

States 


government bonds 

( = *USG " "US^’ 




Equation (2) states that real Canadian corporate borrowing costs are equal 
to the (exogenous) real United States government bond rate plus, again, 
three interest-differential terms. The first and third terms are identical 
to those given in equation (1), while the second term is altered only by 
subtracting the difference in (long-term) expected Canadian and American 
inflation rates. 

It is clear that in a world of perfect information and frictionless mar- 


217 




kets real, not nominal, interest rates should be the focus of attention. 
However, nominal interest rates may also affect capital expenditure deci¬ 
sions under 'real world' conditions - for example, by creating corporate 
cash-flow problems or because the tax system is not neutral with respect 
to changes in the inflation rate. 3 All subsequent analysis considers both 
real and nominal interest-rate channels, and these discussions will make it 
clear why equation (2) is written in the particular form used. 

The Canadian corporate interest premium 


Consider the first interest-differential of equations (1) and (2) - a dif¬ 
ferential that could be labelled the Canadian corporate premium. The fact 
that such a premium exists indicates that capital market lenders prefer 
Government of Canada bonds to corporate bonds. One reason for this 
preference, and almost certainly the most important, is the perceived 
greater default risk for corporate debt. Obviously, the extent of this 
added risk varies, from a very small increment for top-rated corporate 
debt issues to a rather sizable premium for corporations under serious 
financial pressure. 

In addition to a premium for default risk, corporate interest rates 
may include a compensation for higher transactions costs. These higher 
costs reflect, in part, a thinner market (and therefore lower liquidity) for 
debt issues of any specific company than for Canadian government bonds 
and therefore somewhat higher bid/ask spreads. The transactions costs 
may also include higher 'monitoring' costs - for example, the costs of 
assessing whether the default risk of a particular bond has changed. 

Another factor affecting the corporate interest premium is investors' 
desire for portfolio diversification. In general, risk-averse investors will 
decrease the chance of catastrophic loss by diversifying their portfolio 
investments across a wide range of income-earning assets. To the extent 
that holdings of any particular asset decrease the overall income variability 
of the portfolio, that asset can be an attractive risk-reduction vehicle, 
even if its own inherent return is quite uncertain. 4 This argument implies 
that the corporate interest premium can alter if the relative portfolio-diver¬ 
sification characteristics of government and corporate bonds alter, or if 
the relative supplies of these instruments change. For example, an in¬ 
crease in the supply of corporate bonds might raise the interest premium, 
even if the inherent default risk were no greater than before, because this 


218 



risk could not be as easily minimized through diversification (since on 
average corporate bonds would now constitute a higher proportion of the 
average investor's portfolio). 

In the present instance, however, this diversification effect is not 
likely to be of very great importance. For one thing, since the residual 
holder of both Canadian corporate and government debt is the foreign 
sector, and since the Canadian economy is 'small', the proportion of either 
Canadian private or public bonds in foreign investors' portfolios is also 
likely to be small. Therefore, changes in relative Canadian supplies 
should not significantly shift portfolio preferences. Moreover, diversifica¬ 
tion is only important to the extent that the two bond risks are imperfectly 
correlated - in particular, to the extent that holding corporate bonds may 
help to guard against the risk of government bond default. 5 Yet it is 
difficult to imagine circumstances in which the Canadian government would 
default on its obligations but Canadian corporate bonds would remain 
default free. For both of these reasons, changes in relative bond supplies 
are unlikely to explain much, if any, of the movement in the corporate 
interest premium. 6 

Therefore, if expansionary fiscal policy does affect this premium, it is 
likely to do so by influencing either differential transactions costs or 
relative default risks. The first of these possibilities would appear to be 
relatively remote. Increases in the government debt might conceivably 
'thicken' the government bond market and therefore lower costs of govern¬ 
ment bond transactions. But such an occurrence is unlikely to be of any 
major significance, and in any case it would raise the premium by causing 
a fall in Iqq, relative to Iqq- 

The major consideration must be the effect of stimulative fiscal policy 
on relative default risk. There are two possible arguments here. The 
first is that the spectre of rising fiscal deficits could kindle fears of 
government default and hence raise Iqq- However, whether in such a 
situation the corporate premium would fall by an equivalent amount is a 
debatable matter. The economic disruption that would accompany such a 
situation, raising fears of escalating future tax rates, severed international 
capital markets, and possible domestic political crisis, would almost certain¬ 
ly have an adverse effect on confidence in corporate, as well as govern¬ 
ment, solvency. As a result, the corporate interest rate could rise as 
much or more than the government rate. Although this is a harrowing 
possibility, it should be emphasized that at no time in the historical past, 


219 


nor under any of the scenarios currently envisioned, has the risk of 

government default been considered to have more than a small probability. 

In any case, if this probability did start to grow, the effect would appear 

first in international capital markets. Consequently, further discussion of 

this point is deferred to the analysis of Canadian international interest- 

* * 

rate premium (Iqq - i-^g). 

Stimulative fiscal policy can affect the corporate premium in another, 
more likely fashion. Although a major portion of private default risk is 
related to company-specific circumstances, default risk for the private 
sector as a whole varies strongly with overall cyclical movements in the 
economy. A major cyclical downturn not only raises the number of actual 
bankruptcies, but also increases the immediate chance of insolvency for 
virtually all firms. The extent to which this risk gets reflected in a 
higher interest premium will depend not only on the severity of the reces¬ 
sion, but also on the degree to which this severity was unanticipated and 
on the prospect for future recovery. For example, if cyclical contractions 
are mild and relatively regular, then the corporate (default) premium on 
long-term bonds should show only modest cyclical patterns. 7 On the other 
hand, a recession that is unexpectedly strong, and/or one in which the 
prospects for recovery are more uncertain, may markedly raise market 
fears of default. 

Therefore, fiscal policy (and for that matter monetary policy) could, 
in principle, reduce the magnitude of the corporate interest premium in at 
least two ways. First, activist stabilization during any particularly severe 
economic downturn should moderate default fears over that time period, 
with resulting downward pressure on corporate bond rates. Second, to 
the extent that government policies act to smooth cyclical fluctuations in 
general, there should be less chance of unfortunate cyclical suprises. 
Both the average default premium and the cyclical sensitivity of this 
premium should be correspondingly reduced. All this leaves open the 
question of whether 'real world' fiscal and monetary policy can be expected 
to operate on aggregate demand in a counter-cyclical fashion. 8 But the 
essential point is, nonetheless, a challenging one, for it implies that dur¬ 
ing a steep recession a deficit-financed fiscal stimulus may actually 
decrease Canadian corporate interest costs, by lowering default risk. 

In summary, the discussion to this point indicates that the corporate 
interest premium may vary systematically over the economic cycle, but that 
in 'normal' circumstances it is unlikely to be very large. Furthermore, 


220 



Figure 2 

The Canadian corporate long-term bond premium, 1948-82 


Per cent 



McLeod, Young, Weir 
10 industrials 

Government of Canada 
long-term bonds 


Corporate premium 


SOURCE: CANSIM. 

fiscal deficits could cause investment crowding-in, rather than crowding- 
out. 

Evidence on actual behaviour of the Canadian corporate premium is 
given in Figure 2. As can be seen, the corporate and government bond 
yields follow each other's movements very closely; most of the variation in 
the corporate rate reflects a movement in the underlying government rate, 
with the gap between the two generally ranging between 40 and 80 basic 
points. There is also some evidence of cyclical variability in the corporate 
premium. For example, in the current steep recession (from mid-1981 to 
the present), the interest premium has averaged well over 100 basic 
points, peaking at close to two full percentage points in the spring and 
summer of 1982. There are similar high interest spreads during the 1974-5 
and 1970-1 recessions, and to a lesser extent during the milder 1957-8 
slowdown. 9 

Additional and more spectacular evidence of business-cycle effects on 
the corporate bond premium can be found in the experience of the Great 
Depression. As is shown in Figure 3, the corporate/government interest 
spread averaged less than 1 percentage point during the 1920's. During 
the terrible economic plunge of 1929-33, when real Canadian GNP fell by a 


221 





Figure 3 

The Canadian corporate premium, December rates 1919-40 


Percent 



SOURCE: Urquhart and Buckley, Historical Statistics of Canada. 
(Toronto: MacMillan Co. of Canada, 1965). 


third, federal long-term bond yields also dropped - supporting the 
standard wisdom that the Depression was a time of low (nominal) interest 
rates. However, during the same period corporate bond yields rose dra¬ 
matically, peaking at close to 10 per cent at the end of 1932; the cor¬ 
responding corporate premium was almost 5 percentage points. Only as 
the economy started its long and painful recovery did this premium drop 
back to modest values. There is little doubt that a deficit-financed fiscal 
stimulus in the early 1930s would have helped to lower , not raise , corpo¬ 
rate interest rates. 

The currency-related differential 


The second term of equation (1) is the differential between the interest 
rate paid by the federal government on bonds issued in Canadian currency 
and that paid on U.S. dollar issues. Since the issuer is identical in both 


222 









cases, a well-functioning capital market should distinguish between the two 
types of bonds only on the basis of the .probability distribution of future 
exchange values between Canadian and American currencies. At one 
extreme, if it were known with certainty that the exchange rate would 
remain fixed at its present level indefinitely into the future, then the two 
bond yields should be identical. Since this is obviously not the case, both 
the expected future movement in the exchange rate, and the risk associ¬ 
ated with the variance of actual future exchange rates around this expec¬ 
ted value, will create a finite currency-related interest differential. For 
example, if the Canadian dollar is expected to depreciate against the U.S. 
dollar by an average of 1 per cent a year, then by itself this expectation 
would cause a Canadian-dollar denominated bond to carry (approximately) a 
1 per cent higher interest rate than a U.S. currency issue of the same 
maturity period. On the other hand, anticipated appreciation of the 
Canadian dollar would lead, other things being equal, to a negative dif¬ 
ferential between the yields of the Canadian and American currency issues. 

Further insights into the role of exchange-rate expectations can be 
obtained by examining the corresponding currency-related term in equation 
(2) for the real interest cost of borrowing. Specifically, this term is: 

^CG " *CG^ " (n C " n US^‘ 

If the currency-related differential in nominal interest rates increases 
because of an expected future depreciation of the Canadian dollar, then 
this term implies that there will be resulting upward pressure on Canadian 
real interest rates only to the extent that the expected depreciation 
exceeds the expected excess of the Canadian over the U.S. inflation rate. 
The phrase real exchange rate is sometimes used to mean the measured 
exchange rate (i.e., the price of the Canadian dollar in terms of the U.S. 
dollar) multiplied by the ratio of Canadian to American domestic price 
levels. A drop (rise) in this real exchange rate roughly represents a 
deterioration (improvement) in the Canadian terms of trade with the United 
States. 16 The previous observation can, therefore, be rephrased to say 
that an expected exchange-rate depreciation will cause Canadian real 
interest rates to rise only if, and to the extent that, the real exchange 
rate also is expected to depreciate. In general, over long periods of time, 
purchasing-power-parity pressures cause exchange-rate movements that 
offset any significant difference in inflation rates. 11 Therefore, the 


223 





chance of substantive long-term movements in the terms of trade is un¬ 
likely, and hence any induced effect of currency depreciation (or apprecia¬ 
tion) on real interest costs is almost certain to be quite a bit less than the 
corresponding impact on nominal interest rates. 

In addition to the impact of expected exchange-rate movements, the 
currency-related interest differential (both nominal and real) may be 
affected by exchange-rate risk . The nature of this risk should be con¬ 
sidered from the viewpoint of an American owner of Canadian government 
bonds, since it is the foreign portfolio investor (and usually the U.S. 
investor) who is the residual holder of Canadian debt issues, both those 
issued in Canadian currency and those issued in American currency. 12 In 
other words, the currency-related differential, whether calculated in 
nominal or real terms, must reflect the relative risks to the U.S. investor 
of holding Canadian-dollar-denominated versus U. S.-dollar-demoninated 
assets. It is useful to note that this risk comparison differs depending on 
whether the calculation is made in nominal or real terms. In nominal 
values, the U.S.-dollar denominated bond is clearly less risky to the U.S. 
investor, since exchange-rate fluctuations have no impact on the bond's 
effective, U.S. dollar yield-to-maturity. Canadian-dollar bonds have to 
carry an interest rate that not only reflects an equivalent expected yield 
(inclusive of anticipated exchange rate appreciation or depreciation), but 
also includes a premium to compensate for the risk of unforeseen exchange 
rate movements. 

However, the picture is less clear (at least in principle) when risks 
are considered from the viewpoint of unexpected shifts in the real purchas¬ 
ing power of bond yields. The U.S.-dollar bond is no longer automatically 
a less risky asset, since its real yield will be affected by unanticipated 
shifts in U.S. inflation rates. On the other hand, the real yield on 
Canadian-dollar denominated bonds is influenced by both exchange-rate 
movements and American inflation rates. If these two variables moved 
completely independently of one another, then clearly the Canadian-dollar 
bond would remain more risky than the U.S.-dollar bond. However, if 
unexpected appreciation (depreciation) of the exchange-rate value of the 
Canadian dollar is associated wtith unanticipated increases (decreases) in 
the U.S. domestic price level, then it is possible that the Canadian-dollar 
bond will be less risky than its U.S.-dollar equivalent. For example, if 
purchasing-power-parity kept the real exchange constant at all times, and 
if the Canadian inflation rate were very steady and predictable, then any 


224 




fluctuations in U.S. prices would be virtually completely offset by induced 
exchange-rate movements. A Canadian-dollar bond would offer the U.S. 
investor a kind of inflation-indexed asset, and, therefore, could be sold at 
a lower expected yield than its U.S.-dollar analogue. 

In practice, such conditions are unlikely to occur. Real exchange 
rates do fluctuate, and Canadian inflation rates are probably at least as 
volatile as American inflation rates. Therefore, Canadian-dollar bonds will 
probably have to incorporate an (exchange-rate) risk premium in compari¬ 
son with U.S.-dollar bonds. The magnitude of this premium will depend 
on the perceived variance of the exchange rate, the degree of investor 
risk aversion, and the extent to which this risk is diversifiable. Since 
Canadian government bonds constitute a very small fraction of total out¬ 
standing bonds and other portfolio assets held in the United States, it 
seems likely that diversification would assure that much of the unique ex¬ 
change-rate risk would have no impact on the interest differential. Only 
to the extent that unanticipated exchange-rate movements were correlated 
with fluctuations in overall portfolio returns would the market dictate a 
positive risk premium. This reasoning also implies that variations in the 
relative mix of Canadian-dollar and U.S.-dollar federal debt should have 
virtually no impact on the magnitude of the risk premium. 

All of these arguments suggest that fiscal (and monetary) policy 
changes will affect the currency-related interest differential only if these 
actions create expectations about future differential movements in Canadian/ 
American inflation rates and/or real exchange rates, or alter the (non- 
diversifiable) risk associated with unanticipated future variations in these 
factors. It is worth emphasizing that the interest differential responds to 
future exchange-rate movements , not to the current exchange-rate value. 
This circumstance has important implications. For example, a matter of 
recent, and recurring, debate in Canada has been the question of whether 
or not it would be a wise economic strategy to seek a (temporarily) lower 
domestic interest rate accompanied by a depreciated value of the Canadian 
dollar. The previous conclusion implies that the very nature of this 
debate is erroneous. An immediate depreciation of the dollar is consistent 
with lower interest rates only if the dollar's value is expected to rebound 
and appreciate again in the future; domestic interest rates are related to 
expected future changes in the exchange value, not to current levels. 13 

In general, fiscal policy can create upward pressure on the currency- 
related interest differential in three different ways. First, it can raise 


225 




the nominal interest differential if it increases the expected future inflation 
rate, perhaps because greater deficit financing strengthens the belief that 
the monetary authorities will accelerate the expansion of the domestic 
money supply. Of course, if nominal interest rates rise because of higher 
expected inflation, then there is no change in real interest rates. To the 
extent that investment is constrained by real rather than nominal interest 
costs, there will be no resulting crowding-out effect. Second, expansion¬ 
ary fiscal policy may cause the real exchange rate to temporarily appre¬ 
ciate above its equilibrium value. During the period in which the real 
exchange rate depreciated back toward equilibrium, domestic interest rates 
would lie above foreign interest rates. Third, a fiscal stimulus could 
actually decrease the equilibrium real exchange rate. Once again, as the 
actual rate depreciated toward the new equilibrium, a positive gap between 
domestic and foreign interest rates would occur. In both these latter 
cases, real as well as nominal interest rates would rise. 14 

In short, the currency-related interest differential does, in principle, 
provide an avenue through which bond-financed fiscal expansion can crowd 
out domestic investment. Nevertheless, the practical importance of this 
effect can be questioned on at least two grounds. First, fiscal deficits are 
unlikely to lead to higher expected inflation if the deficits are largely 
self-eliminating and/or illusory - that is, if they reflect cyclically induced 
revenue/expenditure effects and/or inflation-related measurement errors. 
Hopefully, market expectations will be formulated on this basis. 15 There 
is, in any case, little empirical evidence in Canada of any connection 
between the size of the current fiscal deficit and the future growth rates 
of either the money supply or aggregate prices. 16 

Second, although fiscal policy may induce both short-term and endur¬ 
ing movements in the real exchange rate, the empirical effect of these 
movements on the long-term interest-rate gap is likely to be small. For 
example, consider two twenty-year Canadian government bonds, one issued 
in Canadian dollars and the other in U.S. dollars. Assume that both 
carry the same initial coupon rate. Now, in order to open up just a 1 per 
cent gap in the respective yields to maturity, it is necessary that the real 
exchange rate be expected to depreciate by an equivalent of 1 per cent 
per year compounded. Over the twenty-year maturity period, this repre¬ 
sents a cumulative 22 per cent deterioration in Canada's terms of trade 
with the United States. Such a movement is completely unprecedented in 
Canadian history. In short, expected movements in the real exchange rate 


226 





probably cause significantly less than a 1 per cent swing in interest rates, 

and the fiscally-induced component of such movements is undoubtedly 

smaller still. The necessary corollary is that any substantial change in 

the two yields amost certainly reflects corresponding shifts in expected 

Canadian versus American inflation rates. 

Unfortunately, exploration of the actual empirical evidence on these 

* 

issues is difficult because there is no readily accessible data on i CG , the 
maturity yield of long-term Canadian bonds issued in U.S. dollars. Data 
eventually were obtained for the rather limited time period of 1978-82. 17 
However, for reasons that it is hoped will become clear, analysis of these 
data is delayed until after the discussion of the third interest differential, 
the sovereign-risk premium. 

The sovereign-risk premium 


The differential i^ ' ^usg * R e T ua tions (1) and (2) represents essentially 
two factors. First, the market for Canadian government securities is 
thinner than that for U.S. Treasury bonds, so the transactions costs of 
dealing in the former are larger than those of dealing in the latter. This 
cost difference is relatively small and is unlikely to be affected at all by 
Canadian fiscal policy changes. Second, the interest spread represents 
the differential (to the U.S. investor) in the risk of default between the 
two securities. Historically, both American and Canadian government 
bonds have been considered extremely safe financial assets, with U.S. 
Treasury bonds evaluated as only slightly more 'riskless' than their 
Canadian counterparts. In theory, however, this channel of differential 
sovereign risk represents an important way in which Canadian fiscal policy 
could cause upward pressure on (both real and nominal) domestic interest 
rates. For if international credit markets consider the magnitude of bond- 
financed fiscal deficits to be inappropriately large, then the sovereign-risk 
premium will rise, with corresponding upward pressure on Canadian cor¬ 
porate borrowing rates, other things being equal. 18 

Ultimately, the question is an empirical one. Have the burgeoning 
Canadian deficits of the last few years and/or the prospect of even larger 
deficits in the immediate future caused an increase in the sovereign-risk 
premium? Figure 4 provides rather impressive evidence on this issue for 
the period 1978-82. The Canadian government rate, i CG , represents the 
average yield to maturity on 9 \ per cent bonds maturing in 1994. The 


227 



U.S. government rate, i^gQ, is the average yield on long-term Treasury 
bonds. The two interest rates remain very close together as they move, 
with the Canadian rate averaging about 1 percentage point higher than its 
U.S. counterpart. Actually, the correspondence between the two borrow¬ 
ing costs is probably even closer than these data show, since the two 
series are not matched exactly for either coupon rate or maturity period. 19 

In any case, the message of Figure 4 is quite clear: there is abso¬ 
lutely no indication of any recent increase in the sovereign-risk premium, 
despite the ballooning deficits of the Canadian federal government. 

It is perhaps of some interest to ask whether, in principle, deficit 
spending can trigger an increase in the sovereign-risk premium. In fact, 
there are at least two questions involved. First, there is the existence 
issue - are there major international variations in sovereign-risk premiums? 
Even a very casual survey of the evidence reveals that quite significant 
variations occur. For example, consider the following sample of U.S.- 
dollar bond yields of different countries as given by trades on the New 
York Bond Exchange: 20 


Coupon rate and Yield to 


Issuer 

maturity date 

maturity 


U.S. Treasury 

9s of 1994 

10.3 


Australia 

9 l/8s of 1996 

11.3 


Sweden 

9s of 1997 

13.1 


Mexico 

8 l/8s of 1997 

18.1 


The Australian premium is about the same as 

the Canadian. 

Swedish 

bonds, however, carry a 

risk premium of more than double this 

amount. 

Almost certainly, this 

higher premium reflects 

market concern over 

Sweden's growing public 

and international debt. 

Finally, Mexican bonds 


carry a very large premium of over 1 \ percentage points. Clearly, the 
market considers Mexican default risk to be a major worry. 21 

The second question, or rather set of questions, has to do with the 
determinants and dynamics of these sovereign-risk premia. What indicators 
does the market use to help assess default risk? How do these premia 
vary over time, and, in particular, is the variation relatively gradual or 
can the premia shift markedly in a short while? If the determinants are 
uncertain or highly variable and/or the premia respond in a very non¬ 
linear fashion, then the current level of the risk premium may give only a 
weak indication as to whether the present direction of fiscal policy could 
lead to future upward pressure on interest rates. There is substantial 


228 





Figure 4 

The Canadian sovereign-risk premium, 1978-82 



Year and month 


SOURCES: Wood Gundy, Ltd.; CANSIM. 


scope for careful work in this area, although for the time being the empir¬ 
ical record strongly indicates that there is little reason to be concerned 
about induced changes in the Canadian sovereign-risk premium. 22 


Empirical evidence on the currency-related interest differential 


Because of the limited number of observations on i CG , the currency- 
related differential Oqq - i CG ) can be calculated only for the years 
1978-82. Figure 5 displays this information. As can be seen, the cur¬ 
rency-related differential is quite a bit more volatile then either the 
corporate-risk or sovereign-risk premia. Even over this relatively short 
period of time the differential ranges as high as percentage points, 
while on the low side it is actually negative in several months. While still 
small in comparison with the overall movement in the domestic government 


229 






Figure 5 

The currency-related differential, 1978-82 


Per cent 



Year and month 


Per cent 



SOURCES: Wood Gundy, Ltd.; CANSIM. 


230 






bond yield (which ranges from 9H to 17 per cent), these fluctuations are 
nonetheless quite pronounced. Most of the volatility occurred during 1981, 
when the differential first rose from about 0 in the winter of 1980-1 to 2 \ 
percentage points in the summer of 1981. By the following winter it had 
dropped down to about 1 percentage point, and most recently it has fallen 
to well under 50 basis points. 

Analysis of the determinants of the currency-related differential 
suggests that any substantial variation in the differential could probably 
be traced to expectations regarding future exchange-rate movements, 
expectations based upon anticipated (long-term) differences in Canadian 
and American inflation rates. Although this thesis cannot be proved, it 
does offer a consistent explanation of the 1981-2 experience. During 1981, 
Canadian inflation surged well above American rates of price increase. 
Furthermore, there were growing indications that this higher inflation 
might be more than a temporary phenomenon, since the gap between 
Canadian and American wage increases was even larger than the difference 
in inflation rates. It is not surprising that market participants reacted 
strongly to these circumstances, especially since it was not immediately 
clear what the response of the Bank of Canada would be. From the end of 
1980 through the first quarter of 1981, the narrowly defined money stock. 
Ml, was at or above the upper bound of its target growth range. Only in 
the second half of 1981 was the extent of monetary policy restrictiveness 
revealed, and this Bank of Canada stance may have led to the subsequent 
narrowing of the interest gap during the second half of 1981. The remain¬ 
ing differential of 1 percentage point may have reflected residual concern 
over the stubbornness with which Canadian wage and price increases 
resisted the restraining influence of rapidly weakening aggregate demand 
conditions. Only during 1983 has it become clear that domestic inflation 
rates are finally falling toward U.S. levels. The further drop of the 
differential to a more normal, presumably risk-related, level is almost 
certainly not a coincidence. 

Further evidence of the importance of inflationary expectations to the 

determination of the currency-related differential can be obtained indirectly 

by observing the spread between domestic Canadian and American govern- 

* 

ment bond yields, i CG - i USG • This interest gap is just the sum of the 
currency-related differential and the sovereign-risk premium. Since both 
a priori reasoning and empirical evidence suggest that the latter is a 
relatively steady number, any major variations in Canadian/U.S. bond 


231 



yields are probably caused by exchange-rate effects. 

The time series for these bond yields, displayed in Figure 6, reveals 
two interesting facts. First, the most steady period for the differential 
bond yield was 1963-7, a period encompassing most of the Canadian fixed 
exchange-rate experience. Of course, if market participants were com¬ 
pletely confident that a fixed Canadian/American exchange rate would be 
maintained indefinitely, then there would be no currency-related differen¬ 
tial. Variation in domestic government bond yields would be a result of 
changes in the sovereign-risk premium alone. Perhaps this was roughly 
true of the 1963-7 period. In any case, the exchange crisis of 1968 seems 
to have ushered in a new era of interest-rate volatility, a volatility that 
intensified under the floating exchange-rate system of the 1970s and 1980s. 

Second, there have been two periods in which the government bond 
differential reached 2% percentage points, 1974-6 and 1981-2. It has 
already been argued that the latter experience reflected fears that Cana¬ 
dian inflation would exceed U.S. rates of price increase, and roughly the 
same argument can be made for 1974-6. Principally because of a much 
more restrictive central bank policy, U.S. inflation rates were substantially 
below Canadian rates throughout■ the middle 1970s. For example, the GNE 
deflator rose in Canada by a cumulative 39 per cent during this three year 
span; the corresponding U.S. increase was 26 per cent. In essence, 
market participants inferred from this differential price behaviour that 
there would be a substantial future depreciation of the Canadian dollar 
(which during most of this period was trading at or above par). This 
expectation was presumably based in part on the continued discrepancy be¬ 
tween the two inflation rates and in part on the belief that the real value 
of the Canadian dollar was above its long-run level. Of course, the ex¬ 
pectation was correct: in the nine years since 1974, the dollar has drop¬ 
ped by a total of about 20 per cent. Even over the life of a 10-20 year 
bond, this drop represents a significant annual loss in purchasing power. 

In summary, the empirical evidence provides support for the previous 
analytical claim that most of the fluctuation between Canadian and American 
bond yields stems from differing expected inflation rates. This conclusion 
has two corollaries. First, there has been substantially less variation in 
the real than in the nominal differential between the two yields. Second, 
there is very little reason to believe that bond-financed fiscal deficits of 
the magnitude presently contemplated will exert any upward pressure on 


232 


Figure 6 

Differential Canadian government minus U.S. government long-term 
bond yields, 1948-82 


Per cent 



Year 


SOURCE: CANSIM. 

domestic real interest costs. Indeed, unless fiscal policy triggers fears of 
accelerating inflation, there also should be no significant change in nominal 
interest rates. 

As a final check on this second set of implications, movements in the 
differential government bond yield were plotted against two measures of 
fiscal 'pressure' on bond markets. Figure 7 shows the interest differential 
against the actual federal deficit as a percentage of GNP. Figure 8 de¬ 
picts the same differential against the total unmatured federal debt, again 
as a pecentage of GNP. As can be seen, there is absolutely no relation¬ 
ship between either fiscal measure and the interest-rate spread. 

SHORT-TERM BORROWING COSTS 

Corporate borrowing costs are affected by short-term as well as long-term 
interest rates. Indeed, during the last several years uncertainty about 
(long-term) inflation rates led to an increasing business reliance on finan¬ 
cing arrangements with short-term maturities. Essentially the same inter¬ 
national arbitrage conditions apply to the short-term money market as hold 
for the long-term bond market. In principle, relationships perfectly 
analogous to equations (1) and (2) could be defined for the short-term 
market, and indeed, that is the general procedure followed in this section. 

However, the particular nature of the short-term money market sug¬ 
gested two modifications. First, no direct use was made of yield data on 


233 




Figure 7 

Differential Canadian/American government bond yield versus Canadian 
federal budgetary balance as a percentage of GNP, 1954-82 



SOURCES: CANSIM; Department of Finance Economic Review; author’s estimate for 1982. 


short-term government securities - chiefly because it was feared that 
various regulatory and institutional constraints on the holding of Canadian 
and American treasury bills (such as the secondary reserve requirements 
imposed on Canadian chartered banks) might prevent, at times, full market 
arbitrage of interest yields. Instead, the comparison focuses on the 
differential rates for Canadian and American prime corporate paper. 
Second, there exists an explicit forward market for hedging interest yields 
against the effects of short-run exchange-rate movements. This forward 
exchange-rate premium (or discount) can be explicitly incorporated as one 
of the explanatory components of short-term interest differentials. 

In view of these considerations, the following two identities were 
specified: 

i CCP = *USCP + FWD + *GAP ^ 

r CCP = r USP + t FWD " ^CP " \fSP^ + *GAP' ^ 


234 



































Figure 8 

Differential Canadian/American bond yield versus Canadian 
federal debt as a percentage of GNP, 1957-82 


Per cent 



NOTE: The government debt ratio is calculated as the debt outstanding as of 
March 31, 1983 divided by the GNP level of the previous year. 

SOURCES: CANSIM; Department of Finance Economic Review. 


where 

nominal interest rate on Canadian 90-day corporate 
paper, issued in Canadian dollars; 

nominal interest rate on U.S. 90-day corporate paper, 
issued in U.S. dollars; 


X CCP " 
* 

Hjscp e 


235 






FWD = 90-day forward premium (+) or discount (-) on the 

Canadian dollar cost of a U.S. dollar, expressed as a 
percentage of the current spot price of the U.S. 
dollar; 

* 

*GAP E ^CCP ’ *USCP ~ FWD E the (exchange-rate) covered 

Canadian/American interest differential. 

Equation (3) states that the short-term Canadian nominal interest rate 
can rise relative to its (assumedly exogenous) U.S. counterpart only to 
the extent that either the forward premium or the covered differential 
rises. Equation (4) implies that Canadian and American real interest rates 
diverge only when the forward premium is not equal to the difference 
between the two inflation rates or when there is a non-zero covered dif¬ 
ferential . 

First consider the possible determinants of the covered differential. 
As exchange-rate movements have no impact on this interest gap, it is 
quite analogous to the first and third terms in equations (1) and (2). 
The covered differential could increase, for example, if Canadian corporate 
default risk rose relative to U.S. default risk. Or changes in taxation 
rules could affect the covered interest gap. Finally, shifts in relative 
transaction costs could change the differential. Under normal circum¬ 
stances, none of these possibilities are likely to be especially important, 
and so the covered differential should be relatively small. It is at least 
possible, though, that a significant Canadian economic slowdown, relative 
to the U.S., could raise Canadian default risk and drive up the differen¬ 
tial. To the extent that this occurred, there again could be opportunity 
for a counter-cyclical fiscal stimulus to create downward movement in 
Canadian interest rates. 

Shifts in the forward premium will occur as a result of changes in the 
probability distribution of the Canadian/American exchange rate. In 
essence, the forward premium term is virtually identical to the second term 
of equations (1) and (2). The only difference is that since the markets 
for exchange-rate variation on the one hand and Canadian financial assets 
on the other are 'unbundled' in the money market, it might be expected 
that these short-term markets would work even more efficiently than the 
capital markets. In any case, the basic determinants of differential inter¬ 
est movements are the same in both cases. In particular, the forward 


236 


premium reflects a combination of the expected (short-run) depreciation/ 
appreciation of the Canadian dollar and a further increment to compensate 
for the risk of unanticipated fluctuations in the Canadian/U.S.-dollar 
exchange rates. It is unlikely that this latter risk premium will be very 
large, or particularly variable. As previous reasoning indicates, portfolio 
diversification of Canadian corporate-debt issues across North America 
should leave any specific U.S. creditor with very little unique exchange- 
rate risk. Therefore, the market should require compensation only for the 
non-diversifiable component of this risk. 

Expected exchange-rate movements, however, can create a relatively 
large positive or negative forward premium. Again it is useful to make a 
distinction between anticipated exchange-rate changes that simply reflect 
differential expected Canadian/American inflation rates and changes that 
represent movements in the real exchange rate. As equations (3) and (4) 
indicate, the former, inflation-induced factor changes nominal but not real, 
domestic interest rates. An increase (decrease) in real short-term rates, 
relative to U.S. levels, can occur only when there is an anticipated depre¬ 
ciation (appreciation) of the real exchange rate. It was argued above, 
with respect to long-term rates, that most expected exchange-rate move¬ 
ments are likely to reflect the purchasing-power-parity influences gener¬ 
ated by differential anticipated inflation rates. This conclusion was based 
on the argument that over, say, a twenty-year time span even a relatively 
large shift in the terms of trade can be accomplished by a relatively small 
annual percentage change in the (real) exchange rate. No such easy claim 
can be made with respect to short-term interest movements: any substan¬ 
tial anticipated short-run movements in real exchange rates will have 
significant effects upon the corresponding interest-rate differential. Of 
course, it can be argued that over short periods equilibrium real exchange 
rates are unlikely to alter, at least to any great extent, so the interest- 
rate effect will remain small. This is probably true, as far as it goes. 
But even if anticipated shifts in the equilibrium exchange rate are uni¬ 
formly small, fairly substantial deviations from this equilibrium rate 
definitely can occur - for example, as a result of temporary differences 
between Canadian and American inflation rates that have not been fully 
reflected in offsetting (nominal) exchange rate movements, or because of 
agriculture, energy, or other commodity price shocks that temporarily shift 
the terms of trade. 23 If the market expects a return toward equilibrium 
in the short run, then there will be an anticipated exchange-rate movement 


237 






and a corresponding effect on real short-term interest rates. 

The possible influence of fiscal policy on all this is potentially com¬ 
plex. As indicated previously, rather simple variations on the standard 
Fleming-Mundell model can generate fiscal-policy effects on real exchange 
rates. For example, any impact on (real) short-term interest rates as a 
result of a temporary fiscal stimulus would almost certainly have to occur 
by causing the Canadian dollar value to rise above its equilibrium level, 
creating expectations of future exchange-rate depreciation. The magnitude 
of the resulting effect on short-term interest rates would depend on a 
variety of factors, including the size of the initial appreciation and the 
time horizon over which readjustment was expected to occur. 

It is useful at this point to turn to the empirical evidence. Figure 9 
gives the Canadian and American average interest rates paid on 90-day 
prime corporate paper for the period 1962-82. Also shown is the corre¬ 
sponding interest differential plotted against the forward premium or 
discount on the U.S. dollar. There are two general points worth making 
about these data. First, as with the long-term interest rate series, move¬ 
ments in short-term Canadian rates in general very closely track the U.S. 
figures. This is true not only with respect to the general two-decade 
pattern; it also holds for many of the short-lived interest-rate variations. 
Indeed, over most of the two decades, the Canadian rate has been within 
about 1 percentage point of its U.S. counterpart. Second, when devia¬ 
tions between the two rates do occur, they are almost completely explained 
by movements in the forward premium. This would suggest that all factors 
(differential default risk, transaction costs, etc.) are of little importance. 
(However, this conclusion must be tempered somewhat by noting that the 
data are not well-suited for revealing either the existence of or changes in 
any default premium, since only prime corporate paper is included in the 
sample.) 

Now consider the two major periods in which Canadian and American 
money-market rates diverged, 1975-7 and 1981-2. These are virtually the 
same periods in which Canadian government bond rates exceeded U.S. 
bond rates by the largest amounts. In other words, both the money and 
capital markets indicated that the Canadian dollar was expected to depreci¬ 
ate with respect to the U.S. dollar during these years. As was argued 
above, it is very likely that much of the bond-yield divergence can be 
explained by differences between expected U.S. and Canadian inflation 
rates. This is probably true for the money market also. 


238 



Figure 9 

Canadian and U.S. interest rates on 90-day prime corporate paper versus the 
forward exchange rate premium on the U.S. Dollar, 1962-82 



Yield on Canadian 
prime corporate 
paper 

♦ Yield on U.S. 
prime corporate 
paper 


Uncovered Canadian/ 
American interest 
differential 

Forward premium (+) 
or discount (-) on 
U.S. dollar 


SOURCE: CANSIM. 


However, there is one interesting variation between the experiences 
of the short-term and long-term markets. In the bond market, the yield 
spread was about the same in the two periods, while the interest differen¬ 
tial for 90-day paper was much greater in both absolute and percentage 
terms in 1975-7 than it was in 1981-2. During 1976, in fact, the money- 
market spread averaged close to 4 per cent, since Canadian rates were 
almost double U.S. rates. While it is hazardous to make behavioural 
inferences on the basis of such casual empirical evidence, it is nonetheless 


239 












interesting to note that the discrepancy in experience is consistent with 
the previous argument that anticipated real exchange-rate movements are 
likely to be more important in explaining short-term than long-term yield 
differentials. Therefore, a portion of the short-term interest differential 
in 1975-7 probably represented a rise in Canadian real interest rates. 

In summary, examination of the short-term interest rate experience 
generally supports the previous conclusion that (real) Canadian interest 
rates cannot deviate by very much or for very long from U.S. rates. 
However, there is probably more scope for real interest rate variation in 
the short-term money market than there is in the long-term bond market. 

OTHER CHANNELS FOR FISCAL POLICY EFFECTS ON INVESTMENT 

The previous discussion has assumed that a (bond-financed) fiscal stimulus 
alters domestic investment expenditures only to the extent that it produces 
some change in corporate borrowing costs. Although this is probably the 
most important, and certainly the most discussed, channel through which 
investment crowding-out (or crowding-in) can occur, it is not the only 
one. The present section discusses - in an admittedly rudimentary fash¬ 
ion - a number of other possible methods of influence. 

Even if fiscal policy actions have no impact on borrowing costs, they 
may affect the cost of equity financing by altering either the expected 
value or the variance in the market return on investment. There are a 
number of ways in which such an alteration could occur. As has already 
been mentioned, if a specific fiscal stimulus acts in a truly counter-cyclical 
way on aggregate demand, then the recessionary drop in output will be 
moderated, bringing a corresponding improvement in investment return. 
More generally, if fiscal policy usually acts to dampen cyclical fluctuations, 
and the market perceives this to be the case, then the variance of invest¬ 
ment returns will be reduced, a result that effectively implies a lowering 
of equity financing costs. 

Two issues are involved here. The first is whether discretionary 
fiscal policy can and will be (output) stabilizing in its influence. The 
second question, closely related to the first, is how market participants 
perceive this fiscal policy impact. For example, if individuals are pes¬ 
simistic about the effect of fiscal policy initiatives, there may be an ad¬ 
verse impact on investment even if these expectations are wrong and fiscal 
policy is indeed operating in a counter-cyclical fashion. 


240 


There is heated debate among economists on both these issues. At 
the risk of oversimplification, a middle or consensus opinion might be 
summarized as follows. Although efforts to 'fine tune' aggregate demand 
run great dangers of either immediately backfiring or creating cumulative 
future instabilities, there is scope for activist fiscal and monetary policy to 
offset or at least ameliorate strong and sustained economic fluctuations. 
Furthermore, if prudent stabilization policy is followed, it is most unlikely 
that private sector individuals will misinterpret these actions. Misjudg- 
ments may be made about the impact of particular intervention episodes, 
but systematic errors will almost certainly be avoided. 

These issues relate to the more general question of business con¬ 
fidence. Fiscal actions that strengthen private-sector belief in the appro¬ 
priateness and consistency of government policy-setting, and that improve 
overall macroeconomic performance, are likely to encourage a high level of 
business investment. And, of course, the opposite is also true. 

A second way that fiscal policy could affect investment is through the 
tax system. Fiscal initiatives that either raise after-tax returns on invest¬ 
ment (e.g., accelerated depreciation allowances, investment tax credits, 
etc.) or lower the after-tax cost of borrowing (e.g., interest subsidization 
or loan guarantee provisions) could provide a significant investment stimu¬ 
lus. Indeed, many of the current advocates of fiscal stimulus argue for 
specific fiscal measures that would encourage investment expenditures. 

Finally, in addition to any economy-wide effects it may have on 
capital formation, fiscal policy may change the mix of investment expendi¬ 
tures. This could occur either because of the direct nature of the fiscal 
stimulus (e.g., interest subsidies for housing construction but not for 
corporate fixed investment) or because of the indirect impact of the initia¬ 
tive on relative prices or relative demand. For example, the fiscal stimu¬ 
lus may alter real exchange rates or domestic energy prices. 

Any attempt to estimate the possible magnitude - or even the direct¬ 
ion - of the net impact of these different fiscal influences on demand is far 
beyond the scope of the present paper. In any case, such an estimation 
would depend on the exact nature of the fiscal actions taken. Two con¬ 
clusions only will be emphasized. First, it is quite possible for a (bond- 
financed) fiscal stimulus to lead to an increased level of domestic invest¬ 
ment. Second, the nature of the investment effects is far broader than 
the simple idea of investment crowding-out through higher interest rates. 


241 


THE CROWDING-OUT OF NET EXPORTS 


The principal focus of this paper is on the effect of Canadian fiscal policy 
on investment. The argument has been made that because Canada is a 
small, open economy functioning in an integrated world capital market, 
there should be little, if any, investment crowding-out, at least through 
the usual interest-rate channel. However, the very openness of the 
economy, which precludes any significant interest-rate squeeze on invest¬ 
ment, also creates the strong possibility that fiscal stimulation will crowd 
out another component of aggregate expenditures: net exports. An 
increase in government expenditures (or a decrease in taxes) can put 
upward pressure on the Canadian exchange rate (because of induced 
capital inflows stemming from the higher financing needs), and the result¬ 
ing appreciation will increase Canadian imports and decrease Canadian 
exports. In the Fleming-Mundell model, the drop in net exports completely 
offsets the rise in the government deficit - i.e., there is 100 per cent 
crowding-out through the foreign trade sector. For more complex models, 
there will generally be less than complete crowding-out. Nonetheless, the 
potential exists for a major decrease in net exports if fiscal policy causes a 
significant exchange-rate appreciation. 

All of this stresses the importance of the monetary policy decisions 
made during the time of the fiscal stimulus. For example, if monetary 
policy is set with the goal of preventing, or at least moderating, any 
changes in the exchange rate, then the adverse effects on net exports will 
be reduced. In the limit, if expansionary monetary policy is used to 
prevent any exchange rate appreciation, then there will be no crowding- 
out. What the actual response of the Bank of Canada would be under the 
present circumstances is another of those debatable questions. It is 
certainly clear that, for the last couple of years, exchange-rate considera¬ 
tions have been an important determinant of Bank of Canada policy. 
However, both the Bank's statements and its actions have largely reflected 
concern about the effects of too steep and/or too rapid a fall in the value 
of the exchange rate. Excessive depreciation has been seen as raising the 
domestic cost of imports and import-competing goods, and hence accentuat¬ 
ing the inflation-adjustment problem that has been the Bank's principal 
focus since 1976. Whether the Bank would resist any significant currency 
appreciation is much more of an open question. Bank officials have indi¬ 
cated that they do not have a specific exchange-rate value, or even a 


242 




range of values, that they consider appropriate and would be willing to 
defend. On the other hand, it is hard to believe that the Bank would sit 
back idly and watch the dollar appreciate to, say, 90 cents U.S. Effect¬ 
ively, therefore, Canada may well have a kind of 'wobbly peg' exchange- 
rate system. 

If the Bank of Canada did prevent any substantial movement in the 
exchange rate, it would in the long run lose the independent power to 
control the domestic money supply. Some observers might argue that this 
has already occurred, and is reflected in the November 1982 decision of 
the Bank to suspend the setting and use of explicit Ml targets. The 
present writer, however, accepts the official explanation for this action (as 
a response to a short-run instability in the demand for money function) 
and believes that the Bank's principal focus remains on the reduction of 
Canadian inflation through continued monetary restraint. Pegging the 
exchange rate, therefore, would represent a significant departure for Bank 
policy, since in the longer term such an action would essentially set the 
Canadian inflation rate equal to the U.S. rate and hence pass the problem 
of inflation control into the hands of the U.S. Federal Reserve Board. Of 
course, with the current year-over-year American inflation rate running at 
less than 4 per cent, such a prospect may be viewed without much alarm. 
Indeed, a substantial economic case can be made that, given the pervasive 
interconnections between the economies of the two countries, efforts to run 
a substantially different Canadian monetary policy may be too costly to be 
worthwhile. On the other hand, even a casual glance at the swings in 
Federal Reserve Board actions during the last decade would suggest that it 
would at least be desirable to be able to insulate ourselves from some of 
the effects of these policy shifts. Whether what is desirable is also 
possible is still another subject for debate. 

In any case, the fundamental point of this section remains true. 
There is a potential fiscal crowding-out problem affecting net exports, 
rather than investment. Therefore, the actual demand effect of any fiscal 
policy initiative will depend to a significant extent on the nature of the 
coincident monetary policy. 

CONCLUSIONS 

The basic points of the present paper stem from a fact well known to 
Canadians: there are advantages and disadvantages to being a small econ- 


243 


omy well integrated with a much larger neighbour. In this particular 
case, the disadvantage is that any increase in American interest rates will 
have a major and direct effect on Canadian borrowing costs, and hence on 
business investment. The advantage is that shifts in Canadian fiscal 
policy are unlikely to cause any significant additional upward pressure on 
domestic interest rates, particularly long-term rates. This is especially 
true if the fiscal stimulus is designed to be (and is perceived to be) 
self-eliminating as the economy recovers from its present severe recession. 
Indeed, there is some real chance that an anti-cyclical stimulus of this sort 
could actually lower real corporate borrowing costs by decreasing the risk 
of default. Consideration of other channels through which fiscal policy 
may affect investment obscures the situation somewhat. There is the 
possibility that a particular fiscal action could be intepreted as raising 
economic uncertainty and business risk in the future and hence could 
discourage investment even if there were no change in interest costs. 
However, it is also possible that investment stimulation could occur if 
business expected the fiscal stimulus to improve economic performance or if 
the specific fiscal initiatives favoured investment expenditures. The 
nature and determinants of business perceptions about economic policy are 
clearly issues of great importance here. 

Despite these last few caveats, there is little doubt in the author's 
mind that a package of fiscal stimulation could be designed, even one of 
substantial size, that would have no adverse effect on domestic interest 
rates or investment. Furthermore, if the Bank of Canada cooperated by 
preventing any significant appreciation of the Canadian dollar, there 
should not be any major crowding-out of the net export sector. In short, 
it should be possible for the fiscal and monetary authorities to provide a 
large aggregate demand stimulus to the Canadian economy, if this were 
deemed desirable . 

The question of desirability, unfortunately, is not an easy one to 
resolve. On one level the answer would appear obvious. The tragically 
high unemployment rate, the idled plant capacity, and the resulting lost 
output all scream for a concentrated program of economic stimulus. Yet 
there is a counter-argument, though it does not lie in apocalyptic visions 
of Canadians being drowned in a sea of government red ink. The real 
concern is whether the current economic recovery, strengthened and 
accelerated by government policy will leave the decade-long problem of 
persistent inflation unresolved and potentially resurgent. It must be re- 


244 




cognized that while the past recession has been more severe in Canada 
than in the United States, Canadian inflation rates until recently have 
remained well above American levels. This divergence strongly suggests 
either that inflationary expectations are more deeply embedded in Canada 
or that our wage and price setting mechanisms are less sensitive to market 
conditions - or both. Most recently there have been encouraging signs 
that deceleration of wage and price increases is finally bringing Canadian 
inflation rates down toward the low U.S. levels. It would be a terrible 
disaster if a buoyant economic recovery reversed these favourable trends 
and led to still another inflationary burst and a subsequent, possibly even 
worse economic recession. 

Judging the benefits and costs of economic stimulation is no easy 
task. Ultimately the choice between stimulus and restraint may have to 
rest upon the weighing of two risks, each of which concerns the formation 
of economic expectations. If restraint is maintained, there is the risk of 
being trapped in a downward spiral of fading business confidence, plum¬ 
meting investment and output, individual and firm bankruptcies, and 
retaliatory trade restrictions. Our monetary and fiscal institutions (with 
considerable help, one hopes, from a recovering world economy) might well 
be able to pull the economy out of such a slide, but it would be nice not 
to have to put them to the test. On the other hand, a strong recovery 
could well trigger inflationary expectations - expectations that would 
presumably be even more painful to reverse in the future than they have 
been in the recent past. 

Avoiding these twin risks - a downward spiral of plunging business 
and consumer confidence and an upward spiral of escalating price expecta¬ 
tions - will require knowledge, skill, public debate over issues, and pos¬ 
sibly a little luck. This is the really important issue in the deficit debate. 
It must be made the focus of both public and professional attention. 


NOTES 


1 However, crowding-out of net exports can take place, unless expansion¬ 
ary monetary policy prevents an appreciation of the exchange rate that 
could otherwise occur. This issue is discussed in more detail later 
in this paper. 

2 This assumption is relaxed later in the paper when other channels 
whereby government deficit spending may affect business investment 
decisions are discussed briefly. 


245 


3 For a recent summary of some of these taxation issues, see the 
readings in Conklin, ed., 1982. 

4 An example may clarify this point. Consider two hypothetical bonds, A 
and B, both carrying the same yield to maturity. Assume also that a 
particular investor considers that the chance of default is 3 per cent 
in the case of bond A and only 1 per cent in the case of bond B. At 
first glance, it may appear that bond B dominates bond A as an invest¬ 
ment for this individual - i.e., that his holdings of bond A will be 
zero. However, this need not be the case if the investor is risk 
averse in the specific sense of being willing to accept a somewhat 
lower expected yield in order to minimize the chance of a catastrophic 
(default) loss. The investor may then choose to hold some of bond A 
if the chance of bond A default depends on very different (ideally, 
opposite ) circumstances than the chance of bond B default. This might 
be the case, for example, if bond A were issued by a high energy-using 
company and bond B by an energy-producing company. By diversifying 
his holding between the two, the investor would raise the chance that 
some default would occur (and hence lower his expected yield), but he 
would minimize the chance that his whole bond investment would become 
worthless. 

5 Or to hedge against some other income-reducing event - e.g., an in¬ 
crease in effective tax rates on interest income from government 
bonds. 

6 One possible counter-argument would rest on the assumption that, 
rather than being widely disbursed, holdings of Canadian corporate or 
government bonds will be concentrated among the relatively small 
number of domestic and foreign traders (perhaps because of informa¬ 
tional costs). In such circumstances relative supplies might impact 
on the differential yield. Such a possibility could be empirically 
tested. 

7 Some responsiveness will remain because the expected discounted 
returns to bondholders will vary with the time pattern of default 
risk. 

8 The results that Douglas Auld has presented to this conference suggest 

that, in Ontario at least, discretionary fiscal policy has had a 
counter-cyclical orientation. It should also be noted that the argu¬ 
ment presented so far implicitly assumes that stabilizing aggregate 
demand fluctuations will also smooth aggregate output fluctuations. 
The advocates of the strong form rational expectations school would 
challenge this (e.g., Lucas 1972, 1973; Sargent and Wallace 1975). 

9 The magnitude of the cyclical sensitivity is almost certainly under¬ 
stated by the data. The corporate bond index excludes any issue that 
drops below a BB rating. Also, the lower-rated bonds are much more 
cyclically variable than their higher-rated counterparts. Indeed, the 
market for lower-rated bonds thins substantially in economic down¬ 
turns. I am indebted to David Adamo at McLeod, Young, Weir for these 
points. Roderick Macgillivray, Dominion Securities Ames, made similar 
comments to me. 

10 The equivalency is not exact, since the terms of trade are usually 
calculated using export and import price indices rather than some 


246 






aggregate domestic price measure such as the GNE price deflators, and 
since terms-of-trade figures are usually given for all trade trans¬ 
actions - not just those with one other country. Both of these dis¬ 
tinctions are ignored in the remainder of the paper. 

11 Purchasing-power-parity is a reflection of goods-market arbitrage. In 
the sense used here, it implies that prices of similar commodities in 
the United States and Canada cannot experience continually different 
growth rates, when calculated in the same currency units . When gen¬ 
eralized across all prices in the two economies, purchasing-power- 
parity implies that, over longer periods, any excess (shortfall) of 
the Canadian over (below) the American yearly inflation rate must be 
offset by an equal annual depreciation (appreciation) of the Canadian 
dollar. The relevance of the goods-market arbitrage condition is 
quite apparent for nearly homogeneous commodities such as agricul¬ 
tural, forestry, fishing, and mining products. But even for hetero¬ 
geneous products such as manufactured goods it should be clear that 
continual increases in Canadian prices versus U.S. prices (valued in 
Canadian dollar terms) are unsustainable in the long term. 

12 At the end of 1981, non-residents held $9 billion worth of Canadian 
government securities out of a total of $77 billion outstanding (ex¬ 
cluding holdings of the Bank of Canada). Only $3.5 billion of the 
debt was payable in foreign currencies. 

13 These conclusions have been derived with respect to long-term interest 
rates. However, it will become clear later in the paper that the same 
reasoning applies with respect to short-term rates. 

14 The nature of the theoretically possible instances in which these two 

effects can occur depends upon the type of analytic model used. In 
the simple Fleming-Mundell model, neither effect can occur - because 
there is no dynamic mechanism for generating movements in exchange- 
rate expectations. However, other assumptions can yield different 
conclusions. For example, consider the impact of a stimulative fiscal 
policy that is known by everyone to be short-run (a temporary tax cut, 
for example). Even incorporating the remaining Fleming-Mundell as¬ 
sumptions (e.g., fixed prices, no asset accumulations), this situation 
can raise Canadian (real) interest rates by increasing the currency- 
related differential. Specifically, the capital inflows triggered by 
the increased financing needs will cause the spot exchange rate to 
appreciate. However, when the stimulus is abandoned, the exchange 
rate will return to its original value. Therefore, this expected 

depreciation will drive Canadian interest rates up by some (probably 
small) amount during the interim period. 

Fiscal policy can also alter the equilibrium real exchange rate. 
Indeed, this will occur in the previous example if asset accumulations 
are incorporated into the analysis. During the stimulus period, the 
appreciated value of the Canadian dollar causes a decline in the 
current account balance of the Canadian balance of payments. This is 
offset by a corresponding capital inflow. In other words, Canada's 
net foreign indebtedness will rise, and therefore interest payments to 
service this debt will also rise. After the stimulus is removed, a 
higher trade surplus will be required, which means that the Canadian 
dollar will eventually have to depreciate below its original value to 
restore long-run equilibrium. As a result, the interest-rate effect 
will be somewhat greater and/or more prolonged than would be predicted 


247 





if asset accumulation effects were ignored. 

15 Nonetheless the possibility of systematic misperceptions of market 
participants does exist, especially during a transitional period. 

16 Chart 5 in John Grant's 1983 paper provides some supporting evidence 

of this lack of correlation. This also was the conclusion of Crozier 

(1976a, 1976b, 1977), though see Pattison (1977), Richardson (1977), 

and Christofides (1977) for differing views. 

17 Through the most helpful' cooperation of John Grant at Wood Gundy. 

18 Conceivably, other things may differ. In particular, there may be a 

(partially or fully) offsetting downward movement in the corporate 
premium, i^ - i^, if the market decides that the government's fiscal 
problems imply little or nothing about the default risk of private 
Canadian borrowers. Because this line of inquiry is so hypothetical 
as to make even an economist uncomfortable, the author will leave 
further speculations about possible market behaviour to the interested 
reader. 

19 Bonds with equal maturity periods but differing coupon rates can carry 
different maturity yields both because holding period yields will vary 
between the two and because capital gains on appreciation of principal 
are taxed at a lower rate than interest income. Differing maturity 
periods will cause variation in yields if the term structure is not 
flat. It should also be noted that the Canadian series is end-of- 
month data, while the American series is the average yield over the 
month. 

20 As reported in the Wall Street Journal , 23 February 1982. 

21 More recent data indicate that these fears may have lessened as the 
Mexican premium, though still large, dropped by about one half during 
the spring and summer of 1983, perhaps reflecting confidence in the 
IMF-imposed austerity program. Whether this confidence is well- 
founded remains to be seen. 

22 To give some quick perspective - the ratio of external federal debt to 
total exports is probably about 20-25 times higher for Mexico than it 
is for Canada. 

23 The question of whether such relative price movements cause short-run 
(disequilibrium) deviations from a long-run equilibrium or simply 
generate random fluctuations in the equilibrium rate itself is pro¬ 
bably of interest only to the economic semanticist. 


REFERENCES 


Christofides, L.N. (1977) 'The federal government's budget constraint 
1955-75.' Canadian Public Policy 3, 291-8 
Conklin, D., ed. (1982) Inflation and the Taxation of Personal Investment 
Income: An Analysis of the Canadian 1982 Reform Proposals (Toronto: 


248 






Ontario Economic Council) 

Crozier, R.B. (1976a) 'Deficit financing and inflation: facts and fictions.' 
The Conference Board of Canada, Occasional Paper No. 3, Ottawa 

- (1976b) 'Inflation, government financing, the money supply and the 

fiscal setting: a review of evidence.' The Conference Board of 
Canada, Occasional Paper No. 5, Ottawa 

- (1977) 'Deficit financing and inflation: a review of the evidence.' 

Canadian Public Policy 3, 270-7 

Fleming, M. (1962) 'Domestic policies under fixed and under floating 
exchange rates.' International Monetary Fund Staff Papers 369-79 
Grant, J. (1983) 'Deficits and capital markets' in this volume. 

Lucas, R.E., Jr. (1972) 'Expectations and the neutrality of money.' 
Journal of Economic Theory 11, 103-24 

- (1973) 'Some international evidence on output-inflation trade-offs.' 

American Economic Review 63, 326-34 
Mundell, R. (1963) 'Capital mobility and stabilization policy under fixed 
and flexiable exchange rates.' Canadian Journal of Economics and 
Political Science 29, 475-85 

Pattison, J.C. (1977) 'Government deficits and inflation: the evidence 
reconsidered.' Canadian Public Policy 3, 285-90 
Richardson, R.M. (1977) 'Deficit financing and inflation: a reply to the 
Crozier Report and the Department of Finance.' Canadian Public 
Policy 3, 278-84 

Sargent, T.J. and N. Wallace (1975) 'Rational expectations, the optimal 
monetary instrument, and the optimal money-supply rule.' Journal 
of Political Economy 83, 245-54 


Peter Andersen* 

Professor Wirick has provided us with an interesting and thought- 
provoking analysis of a subject that must be a primary concern to policy¬ 
makers, the private sector, and academicians alike. The paper relies 
heavily on the Fleming-Mundell model, which assumes infinitely elastic 
capital flows and static expectations. While this model is an important 
theoretical construct, allowing the economist to analyze the myriad compli- 

* Chief Economist, Burns Fry Limited. 


249 













cated relationships affecting the behaviour of a small, open economy in a 
partial equilibrium framework, the model's usefulness is limited in direct 
application to the real world. It is important for us to note the ways in 
which the Canadian and world economies differ from the model's assump¬ 
tions and the degree to which these differences critically affect the 
author's conclusions. It would be a much easier task for our policymakers 
if in fact an infinitesimal increase in Canadian interest rates could bring in 
waves of foreign capital. It would also be wonderful if Mr. Lalonde could 
stimulate the economy back to full employment and not aggravate adverse 
inflationary and exchange-rate expectations. 

The problem is that the real world doesn't operate that way today and 
probably never did. 

Professor Wirick's paper seems to completely miss the point that 
budgetary deficits in Canada appear to have a large structural component 
that will remain long after the recession is over. This is when the 
crowding-out will take place - when private sector borrowing requirements 
have revived and must compete with large government requirements for a 
limited suppply of net domestic saving. Professor Wirick attempts to allay 
our fears by assuring us that gaps between domestic saving and invest¬ 
ment can be financed by virtually unlimited international capital flows. 
This is the familiar argument that Canadians never have to worry about 
capital availability because foreign saving will always be available - an 
assumption, I submit, that is highly unrealistic. In current circum¬ 
stances, any major increase in our reliance on foreign saving would involve 
a major deterioration in our current account balance, significantly higher 
interest rates, lowered rates of capital formation, and an ultimate deterior¬ 
ation in productivity performance. 

There is also the possibility that foreign saving might not give us all 
the capital we want, at least at prices that we are willing to pay. In 
recent testimony before Congress, U.S. Federal Reserve Chairman Paul 
Volcker repeatedly warned that a serious capital availability shortage could 
develop in the United States as private borrowing needs recover and 
collide with large structural deficits. President Reagan admitted in his 
budget presentation to Congress that the Treasury's borrowing needs in 
fiscal 1983-4 would consume 94 per cent of total net saving in the U.S. (in 
1970, the figure was 26 per cent). This would leave very little left over 
for Canada. 

Another weakness in the paper is that it completely ignores crowding- 


250 


out that may be caused by an increase in inflation. To the extent that 
inflation causes government revenues to rise at the expense of private 
incomes, due to the incomplete indexing of the tax system, private invest¬ 
ment may be crowded out in periods of accelerating prices. 

Moreover, the analysis for the most part ignores the existence of 
monetary policy and assumes that fiscal policy can act completely indepen¬ 
dently. It is often the case, however, that central banks attempt to 
soften the collision between private sector and public sector borrowing 
demands by creating substantial amounts of additional credit. 

Interest rates will not necessarily rise as a result of higher inflation, 
but crowding-out will nonetheless occur. Real incomes will be lower, and 
consequently the private sector will be crowded out by the public sector in 
the competition for real resources. 

I found the paper to be somewhat longer than necessary. For 
example, it contains a needlessly drawn out narrative on the identity 
relating corporate rates to the corporate default risk, the currency dif¬ 
ferential, and the sovereign-risk premium. 

If I were addressing the issue of crowding-out, I would focus on the 
spread between Canadian corporate bonds and government bonds and the 
level of Canadian government bonds. Professor Wirick spends too much 
time explaining why risk premia exist. The issue at hand is how they 
change in response to discretionary fiscal policy. 

The discussion of portfolio diversification could easily be shortened. 
The author admits that diversification is only important in the unlikely 

event that the two bond risks are imperfectly correlated. I agree, and I 

feel that too much time is spent in discussing this issue. 

I also consider the discussion of government default risk to be need¬ 
lessly long. The point that activist stabilization policy could moderate 
corporate default fears is a good one, but this policy would produce down¬ 
ward pressure on risk premia only and would not necessarily produce 

lower absolute corporate interest rates. Corporate spreads could conceiv¬ 

ably narrow, relative to government spreads, but I am not at all convinced 
that the actual level of corporate rates would decline as the author says 
they would do. 

Professor Wirick's statement that in normal circumstances the cor¬ 
porate risk premium is unlikely to be large reflects his treatment of all 
corporate instruments as prime triple A credits. In reality, of course, 
crowding-out is most likely to affect the less creditworthy borrowers. I 


251 


suggest that the author examine the corporate risk premium by quality 
rating. He would find that there are marked cyclical swings and that wide 
spreads open up between prime credits and lower-rated issues as the 
economy deteriorates. These spreads are likely inversely related to the 
counter-cyclical policy measures. 

I am not happy with the discussion of the currency-related interest 
differential. It reintroduces the unrealistic Fleming-Mundell model, which 
produces the odd result that aggressive fiscal stimulation would strengthen 
the Canadian dollar. I believe that in present circumstances expectations 
would produce exactly the opposite effect. 

Professor Wirick attempts to downplay the potential impact of a bond- 
financed program of fiscal expansion by claiming that the resulting deficits 
would be self-eliminating, or illusory. He states the hope that market 
expectations would be formulated on this basis. This is where I feel a few 
basic points need to be made. 

First, we are supposed to be talking about discretionary fiscal ac¬ 
tions, not just the cyclical growth of the deficit as a result of the reces¬ 
sion. Professor Wirick does not recognize the difference between a struc¬ 
tural and a cyclical deficit problem. I think the whole point of the discus¬ 
sion on deficits is that a large part of the deficit is no longer self- 
eliminating. If Professor Wirick could design a fiscal plan with a deficit 
that eliminated itself by the mid-1980s and convince Ottawa to adopt it, I 
would be much less concerned about this issue. Given the current fiscal 
policy setting, however, the current federal deficit will remain very large 
well after the recession ends. 

The discussion of the sovereign risk premium completely ignores the 
existence of monetary policy. Spreads on U.S. pay Canadas and U.S. 
Treasury bonds have definitely moved as a result of discretionary mone¬ 
tary policy. Fiscal and monetary policy must not be treated as if they 
were independent of one another. Furthermore, the evidence of a wider 
spread would be found at a time of recovery in private sector credit 
demand rather than in the depth of a recession. 

Finally, the last section which deals with the crowding-out of net 
exports, makes no sense to me. The author, using the Fleming-Mundell 
assumptions, implies that discretionary fiscal stimulus would raise the value 
of the Canadian dollar. As I said, I hold the opposite view. A $5 billion 
to $10 billion additional stimulus on top of the $29 billion initially estimated 
for fiscal 1983-4 would, I believe, produce strong expectations of a weaker 


252 


Canadian dollar and a need for the Bank of Canada to respond with higher 
rather than lower interest rates. 

To wrap up, I do not believe that the paper supports the conclusion 
that, in current circumstances, larger Canadian government deficits would 
be unlikely to cause any significant additional upward pressure on interest 
rates, and that as a result, little crowding-out would be forthcoming. 


Discussion 


MODERATOR: If ever there was an example of the two solitudes, I think 
we've got it here. I am going to give Professor Wirick just a couple of 
minutes for rebuttal. Then we will go to questions from the floor. 

R.G. WIRICK: First of all, I will agree about the rambling nature of the 
paper. I learned long ago that under time pressure it is much easier to 
write long than it is to write short. I hope that future revisions will 
tighten the paper up a bit. 

On the other hand there's very little else that Peter Andersen said 
that I do agree with. Indeed, the basis for some of his arguments eludes 
me completely. 

I went through the identities because they are identities . If you 
believe that there is going to be crowding-out through the interest-rate 
mechanism, you must specify a way in which those interest differentials 
widen. You cannot just wave your arms in the air, invoke some statement 
that we will not have enough domestic savings which will upset markets in 
some unknown way, and claim that domestic investment will be crowded 
out. You must identify specifically one or more of those interest com¬ 
ponents that is supposed to increase, and you have to say how this will 
occur. The point is a simple and obvious one, but, given the nature of 
some of Dr. Andersen's comments, perhaps it needs to be re-emphasized. 

For example, Dr. Andersen accepts that 'activist stabilization policy 
could moderate corporate default fears,' thus narrowing corporate-govern¬ 
ment interest rate spreads. Indeed, he seems to argue that this impact 
might be even stronger than my data would suggest (since the cyclical 
sensitivity of lower-rated bonds is greater than the actual interest rate 
series used in my paper). Despite these admissions, he is 'not at all con- 


253 




vinced that the actual level of corporate rates would decline.' Yet for 
Dr. Andersen to maintain such a conclusion he must believe that either the 
sovereign-risk or the currency-related differential will rise by at least as 
great an amount as the corporate-default premium will fall. Which does he 
expect to be the culprit? Why should such a rise occur? What evidence is 
there for such a position in the historical data? On all these points Dr. 
Andersen is silent. 

Regarding another, though closely related issue, it is clearly quite 
true, as Dr. Andersen asserts, that the critical questions are how and 
why the three interest differentials change over time. Yet unless you 
discuss what determines these differentials in the first place, it's very 
difficult to talk about what is going to cause them to change. Consider, 
for example, the sovereign-risk premium. It's clearly true that if Canada 
runs an imprudent fiscal policy, the international financial markets will 
start demanding a very stiff risk premium. I think one focus of future 
work must be an effort to determine what signals the market provides 
when our fiscal policy starts to get us into danger. My paper gives a 
snapshot in time of the significant variation in sovereign-risk premia among 
four countries. Perhaps a more interesting question, which I only allude 
to in my paper, is what the dynamics of those premia have been. For 
example, take a case such as Sweden.- or an even more extreme instance 
such as Mexico. Did the market gradually raise the risk premium to 
Mexico, or did it remain fairly steady for a long period of time and then 
shoot way up because of a new perception about world oil markets? If the 
latter alternative describes the case, then we must be especially careful in 
Canada and, if anything, err on the side of caution when we tap interna¬ 
tional markets. If we get into heavy debt, corporate or public, the inter¬ 
national market might suddenly re-examine what's happening in Canada (or 
what it fears may happen through monetary or fiscal policy in the future) 
and abruptly demand a much bigger risk premium. We could then be in 
real trouble, trying to service a large debt at much higher real interest 
rates than we initially anticipated. 

Let me turn to a final point. Dr. Andersen is skeptical about 
Canada's future access to capital markets, apparently quite apart from a 
concern about increases in the sovereign-risk premium. He cites testimony 
by Federal Reserve Chairman Paul Volcker that U.S. public and private 
borrowing needs could cause a 'capital availability shortage' in the future. 
Dr. Andersen concludes that there may be 'very little [capital] left over 
for Canada.' 


254 


I find this conclusion perplexing. 

I am confident that Dr. Andersen realizes that the response of U.S. 
capital markets in such a situation would be to increase U.S. real interest 
rates. And this result would, of course, be completely consistent with my 
paper. Increased U.S. rates would cause a direct and immediate rise in 
Canadian rates, and as a result Canadian investment, as well as U.S. 
investment, would decline. What I am arguing in my paper is that dif¬ 
ferential changes in our own government's financing needs are not going to 
put a noticeable increment on top of the U.S. rate. In any case, interna¬ 
tional capital is not allocated like tickets to a football game - which may 
simply not be available if you get to the stadium late. Capital funds are 
distributed through changes in interest rates. If we are willing to pay 
the cost, we will get the funds. 

PETER ANDERSEN: I would like to add two additional comments before the 
session is opened to questions from the floor. First, Professor Wirick's 
examination of historical interest rate spreads, while interesting, does not 
constitute sufficient evidence on which to base his conclusion. Prior to 
1974, the federal deficit in Canada did in fact tend to be self-eliminating 
and did move towards surplus in recovery years. We must remember that 
chronic structural deficits have only developed since the mid-1970s and 
that private sector growth since then has been modest. It is, therefore, 
not surprising that Professor Wirick's historical analysis finds little evid¬ 
ence of crowding-out. However, the risk is definitely there for the mid- 
1980s, when we will need a strong spurt in private sector output and 
investment just in order to regain our previous standards of living. If 
deficits were still self-eliminating, then I would be much more at ease on 
this issue. However, it is generally accepted that large structural deficits 
will remain with us long after the recession ends. 

Second, a brief word on inflationary expectations. Professor Wirick 
argues that if nominal interest rates are pushed up because of inflationary 
expectations associated with larger deficits, real interest rates don't in 
fact change and that investment spending won't be affected. This is an 
overly simplistic view of expectations. Larger deficits that destabilize 
bond market expectations are unlikely to have the same impact on the 
expectations of corporate management. The expectations of financial 
market participants (i.e., bond traders and investment managers) and 
corporate management are unlikely to be uniform, and they are based on 


255 



different factors. Financial market participants tend to be much more 
sensitive to inflationary expectations, and higher nominal interest rates 
from this source could definitely mean an increase in real interest rates for 
corporations. The present situation is a good example. Currently, bond 
participants and corporate management show widely divergent expectations 
of inflation. 

QUESTION: My question for Professor Wirick relates to the corporate 

differential shown in Figure 2. It declines to a very low level in the early 
sixties and then it increases. What is the explanation for this? Is the 
corporate differential really the measure of crowding-out, and if it is, is it 
related to the share of government in GNP, the ratio of government debt 
to income, or something else? 

R.G. WIRICK: I don't know. I would like to explore some of this a lot 
more systematically that I have. On the basis of casual empiricism, the 
corporate premium appears to be related to economic cycles. You are quite 
right, particularly about the sixties. In my paper I talked about the high 
points for the premium, but look at the period, say, from '61 or '62 up to 
about '65 or '66. That was the old golden age that I sometimes talk about 
in my classes - when even economists were looked at with respect. We 
probably ill-deserved the respect then as much as we ill-deserve the 
disapprobation now. But it was a time when economic behaviour was 
relatively stable. Unemployment rates were low and falling during most of 
the decade. Inflation rates started to rise towards the end of the 1960s, 
but during most of the period they were low. I think it was this basic 
stability of economic circumstances that allowed the corporate sector to pay 
only a small extra risk premium above the government yield. In essence, 
there simply wasn't much risk of corporate default. So I would maintain 
that most of the pattern here is very definitely a cyclical one. If I were 
to try to test this in a formal sense I would use explanatory variables that 
were in essence cyclical variables - the gap between the actual and the 
natural unemployment rate, or other variables of that sort. To that extent 
I wouldn't expect to find much of a relationship with respect to fiscal 
measures such as the ratio of government expenditures to GNP - unless it 
were in essence an indication that discretionary or automatic stabilization 
policy was being used to help ameliorate cyclical output swings. 


256 


QUESTION: Ron, the thing that stands out very strongly in your chart is 
the importance of uncertainty with respect to the inflation rate - in 1974-6 
and again in 1981-2. I would like to ask both you and Peter if one can 
assume that it's not the deficit per se that affects the differential between 
the U.S. government bond rate and the Canadian government bond rate, 
both expressed in their own currencies. If it's not the deficit per se , but 
rather a perception of uncertainty with respect to the future rate of 
inflation, then under present circumstances of substantial economic slack in 
Canada, would either of you see much potential for a major change in 
uncertainty about inflation as a result of an increase in the deficit that 
was clearly temporary? 

PETER ANDERSEN: I'm not so sure that an increase in the deficit is 
purely temporary. I think this is one of those big assumptions - arguing 
about structural deficits that are still there when the economy is moving 
into a recovery. I don't think you are going to see a 7 per cent or 6 per 
cent unemployment rate at any time within the next five years, and I think 
large structural deficits are going to compete for funds with the re- 
emerging needs of the private sector. 

QUESTION: That may be, but let's just confine it to change in the 

deficit. Say there's a change in the deficit over the next year or two: 
what would be the marginal impact of that? 

PETER ANDERSEN: All right. We are talking about a base figure now of 
at least $29 billion for the federal deficit in 1983-4, the year that begins 
April 1. If we have a $5 billion to $10 billion additional stimulus on top of 
that, I would expect inflationary expectations to remain very volatile, even 
though the capacity utilization level is at an all-time low. 

R.G. WIRICK: I am quite concerned about inflationary expectations - 
indeed, that was the point I was trying to end with in my initial presenta¬ 
tion. Perhaps I gave the wrong impression in my paper, Peter, because I 
was not arguing that we should undertake a fiscal stimulus. I remain 
quite uncertain and would like to have more discussion about it. What I 
was arguing is that it's quite feasible to undertake a fiscal stimulus and 
that such action would indeed stimulate aggregate demand overall. The 
question of whether a stimulus is desirable or not is quite another matter. 


257 






We don't know enough about how inflationary expectations are generated; 
we simply don't know enough. And I don't think any model builder, large 
scale, small scale, or intermediate, would really quarrel with that point. 
What we do know is this: that when inflationary expectations take off and 
become deeply embedded, they're very, very costly to roll back, and as a 
result I would suggest that there is a real danger of trying to stimulate 
the economy too fast or too far. Of course, we are a long way from any 
kind of measure of capacity output and the costs of being that far away 
are also very great, so I'm really on the horns of a dilemma here. 

One final point with respect to inflationary expectations, as far as the 
narrow focus of this paper is concerned: if you increase inflationary 
expectations, you will raise Canadian nominal interest rates versus U.S. 
nominal interest rates, but you still haven't opened up a gap in real 
interest rates. And it's real interest rates, predominantly, that crowd out 
domestic investment, not nominal interest rates. 

QUESTION: My question is for Dr. Andersen. I'm not sure I heard him 
correctly, but I thought he said at one point that he could foresee circum¬ 
stances in which Canadian borrowers could not get the capital they needed 
on international markets, even at higher rates, whereas Professor Wirick 
said that if borrowers are prepared to pay the real rate demanded they 
can always get capital. I tend to lean to Professor Wirick's view and I 
find it difficult to understand Dr. Andersen's view. 

PETER ANDERSEN: My point is that it is highly unrealistic to assume 
there is an infinitely elastic international capital flow available to Canada, 
and that if we increase interest rates by one or two basis points we can 
have all the foreign financing we require to finance a deficit of $35 billion 
to $40 billion at a time when the economy is expanding at a rate of 6 to 8 
per cent. 

COMMENT: I would say, that if Canadian borrowers, corporate or public, 
are prepared to pay the rate demanded on the foreign markets, they can 
always get as much as they want. 

PETER ANDERSEN: My feeling is this: if there is an imbalance between 
domestic saving and domestic borrowing requirements, and if the United 
States is in a capital shortage position at that time, then it will take a 


258 





significant increase in interest rates to attract the additional funds to 
Canada. Provinces with overly large deficits will face reduced credit 
ratings. Of course, funds will always be available at a price, but will the 
borrowers be willing to pay the price or will they cut back their spending 
plans? 

COMMENT: Capital shortage means, presumably, higher rates. 

PETER ANDERSEN: Capital shortage means a U.S. budgetary deficit 
that's taking up 88 to 94 per cent of total net savings in the United States 
at a time when private sector borrowing needs are recovering at a fairly 
high rate. 

COMMENT: Which would drive U.S. rates up? If Canadians were willing 
to pay those rates, or those rates plus the premiums, then Canadians 
would get the capital they needed. 

PETER ANDERSEN: I am saying that the alternative of foreign saving is 
not a guaranteed fact of life and that it would be pretty foolish to think 
we could rely on it at all times. Because fiscal deficits have taken a 
quantum leap only just recently and are generally expected to stay very 
large even after the recession ends, I am very worried that we are head¬ 
ing into new ground and into potential trouble. 

QUESTION: I have one question for you, Ron, and it is this. The 

Fleming-Mundell model says that if we expand fiscal expenditure and create 
a larger deficit you do in fact get full crowing-out on your assumptions of 
elastic capital mobility. It happens to occur through the trade balance and 
not through investment savings, so if you were to stimulate in your world 
by fiscal expansion you would also have to stimulate your monetary expan¬ 
sion, and therein lies one of the dangers. 

R.G. WIRICK: The Fleming-Mundell model, as Dr. Andersen pointed out, 
has a lot of unrealistic assumptions in it. What I am really trying to 
demonstrate in this paper is that the conclusion with respect to the invest¬ 
ment effect is pretty much independent of those unrealistic assumptions 
but instead rests on the international arbitrage of returns on comparable 
financial assets. But you're obviously quite right. In that model, or 


259 


indeed I think in virtually any more realistic variant, you are certainly 
going to get crowding-out of net export expenditures unless monetary 
policy is accomodating. What happens, basically, is that the capital in¬ 
flows triggered by the financing needs of the fiscal stimulus can create an 
instantaneous exchange rate appreciation, which in turn cuts net export 
demand. 

All of this, however, is dependent on the nature of monetary policy. 
The critical question is: what would the Bank of Canada do in such a 
circumstance? We have people here who can, and perhaps will, comment 
on this better than I can. It is clear that the Bank of Canada doesn't 
want the Canadian dollar to depreciate too much, because of its concern 
about subsequent effects on import costs and inflation. But it is not clear 
what the Bank would do if the exchange rate started appreciating. My 
guess is that under current economic circumstances it wouldn't let it 
appreciate too much, which in essence means (a) you would have monetary 
accommodation of any fiscal stimulus and (b) you wouldn't have much, if 
any, adverse effect on net exports. Whether such accomodation would 
create a resurgence of inflationary expectations is, again, a question for 
which I have no certain answer, but it is one that does cause me concern. 
John Bossons probably would respond that as long as the fiscal stimulus is 
temporary and as long as people can agree that there isn't any structural 
deficit, then there should not be any problem with inflationary expecta¬ 
tions, since all we have is a temporary fiscal stimulus accompanied by 
temporary monetary stimulus. But I'm not entirely sure that the stimulus 
would be temporary, nor am I certain that Canadians in general would 
perceive it to be temporary. 


260 


Deficits and capital markets 

John Grant* 


It does not take much acquaintance with the financial press or with the 
business community generally to realize that government deficits, especially 
federal government deficits, are a bogey. There is a palpable protest, 
voiced in hundreds of different ways, that sings out loudly and clearly in 
many, if not most, market commentaries on government economic policy. 
Financial market participants appear to feel that government deficits repre¬ 
sent a serious threat to the stability and productivity of the economy and 
the financial superstructure. Whatever may be the conclusions from 
learned analysis of structural versus cyclical components, inflation adjust¬ 
ments, distinctions between current expenditure and capital expenditure- 
related components, and distinctions between deficits at the federal level 
and those at provincial or municipal levels, these conclusions appear at 
first sight to be pretty much irrelevant to the instinctive distaste for 
deficits of whatever description. 

We may be dealing with a partly irrational phenomenon. But there 
are good reasons in principle for concern about deficits. The first and, 
for capital markets, still the most important concern is related to inflation. 
Throughout history, government outlays in excess of what the taxpayer 
can be persuaded to finance have often been financed instead through the 
creation of money, money lent by bankers to the government. Whenever 
the supply of money has grown faster than society's ability to produce real 
goods and services, it has become the basis for an increase in the price 
level. In the past, such escapades were usually associated with wars, 
which exacerbated the inflationary consequences on the supply side 
because many of the able-bodied young peasants were pressed into military 
service, leaving the economy of the day even less capable than usual of 

* Director and Chief Economist, Wood Gundy Limited. 


261 


expanding physical output in response to increases in monetary demand. 
Even after normal conditions had been restored, the price level might 
remain permanently higher, depending on the ratio of the money supply to 
the size of the physical economy. 

Financial markets today still have every reason to worry about this 
pattern. Elected and authoritarian governments alike continue to follow 
policies of monetary expansion, often with the deliberate intent of debasing 
their currencies. It is of course necessary to keep in mind that a debase¬ 
ment of the currency that is widely anticipated will not provide the govern¬ 
ment with cheap financing. If everyone expects a high inflation rate, then 
prices will tend to rise at a rate that thwarts the government's efforts. 
Currency debasement will only provide cheap financing if the currency is 
pressed on those who are unaware of the rate of debasement. 

Price stability is one thing. Expectations of price stability are 
another. A government that has once destroyed social trust in price 
stability must work painfully and hard to restore it. Thus monetary 
debasement is by no means a simple game to play. It is also evident that 
there is no necessary connection between monetary debasement and govern¬ 
ment deficits. Nevertheless, the temptation to finance ministers to try to 
engage in cheap finance, especially in times of war, recession, and social 
strain, is obviously very great. Figure 1 illustrates for Canada the size 
of the federal debt relative to GNP and the cost of interest on the public 
debt, also as a percentage of GNP, from 1951 to the present. In the early 
post-war years, even though the country was still living with the public 
debt accumulated during the Second World War, the burden of interest 
payments on the debt was for taxpayers not an onerous one. By compar¬ 
ison, the heavy compensation now levied by lenders for the expected rate 
of inflation has exploded the cost to taxpayers of the present round of 
expansion of the national debt. This cost, together with the burden of 
interest both individuals and corporations carry because of their own 
borrowing activities (see Figure 2), has resulted in a huge ongoing trans¬ 
fer of real resources from borrowers to lenders - a transfer that is felt to 
be oppressive and that creates a significant temptation for borrowers, 
private and public, to try to find ways to ease the strain. 

A second legitimate bogey is the threat of crowding-out. Let us 
assume for the moment that deficits do not spawn excessive money crea¬ 
tion. Nonetheless, the real resources used up by government outlays must 
be either borrowed or financed through taxes. Here the question is one 


262 



Figure 1 

Interest on the federal debt as percentage of gnp 
and federal debt as percentage of gnp 



6.0 

5.5 
5.0 

4.5 
4.0 

3.5 
3.0 

2.5 
2.0 

1.5 
1.0 


SOURCES: Statistics Canada; Bank of Canada; 
forecast from 82: IV by Wood Gundy Limited. 


not of inflation but simply of deciding who of all the individuals in society 
will be permitted access to the available real resources and on what terms. 
Even given that monetary policy is appropriate, in the sense that it tends 
to generate just enough monetary demand to employ society's physical 
resources without destabilizing the price level, the superior credit of the 
government could freeze out some would-be private spenders just as effec¬ 
tively as an outright increase in taxes. 

On the other hand, if policy is appropriate in the sense I have just 
mentioned, one can certainly argue that government will not physically 
crowd out private borrowers as long as there are under-used resources 
available. Despite whatever level of borrowing governments undertake in 
the capital market, the monetary authority can manage matters in such a 


263 






Figure 2 

Canadian industrial corporations’ interest coverage 

Corporate Income 
Interest Costs 



SOURCE: Statistics Canada, Industrial Corporations Financial Statistics. 


way as to ensure a level of total demand, including private demand, just 
sufficient to employ all those who wish to work at the going wage rate. 
Thus, as at present, interest rates can fall to encourage private borrow¬ 
ing, despite the heavy financing requirements of government. 

This point stresses the overriding importance of monetary policy, not 
just with respect to the determination of the future of inflation and the 
rate at which the national debt is eroded by monetary debasement, but 
also with respect to the crowding-out issue itself. The monetary authority 
effectively takes on itself to determine the relative price of money and 
bonds; that is, the rate of interest. Can an overambitious fiscal stimulus 
lead to physical crowding-out? Yes it can, but only if the monetary 
authority permits it. In this sense, financial crowding-out is not at all 
the same as physical crowding-out. For instance, if the Federal Reserve 


264 







forces interest rates to rise next year in the context of a huge Reaganite 
deficit, the effect will be to prevent the sort of physical jostling for real 
resources that might otherwise encourage and permit a return to inflation¬ 
ary behaviour. 

However, this is certainly not the end of the story. Canada, as 
Professor Wirick pointed out in the previous paper, is an extremely open 
economy with respect both to trade and to capital flows. Even if Canadian 
domestic physical resources were fully utilized, Canadian borrowers would 
still have access to foreigners' savings, and thus to their real resources 
as well. There is nothing to prevent Canadians from borrowing foreign 
resources and using them, as long as they are willing to pay the approp¬ 
riate rate of interest on the international capital markets. In fact, the 
Canadian economy is so small relative to the world economy that it is 
probably impossible for the Canadian government to physically crowd 
private Canadian borrowers out. It can crowd them offshore, but it 
cannot - at least not through fiscal policy alone - deny them access to real 
resources. On the other hand, as Professor Wirick has noted, 'irrespon¬ 
sible' behaviour by the Canadian government, if perceived by foreign 
lenders, can raise the cost of funds to the government and to private 
Canadian borrowers in foreign markets. Professor Wirick's effort to esti¬ 
mate the risk premiums involved is interesting. The threat of behaviour 
that could lead, for instance, to devaluation of the Canadian dollar is 
constantly and vigilantly assessed on financial markets. 

One sometimes hears the argument that, by forcing Canadian borrow¬ 
ers offshore, the government effectively denies some of them access to 
funds because the cost of financing in offshore capital markets can be 
prohibitively high. We are considering here costs such as those associated 
with marketing new securities, introducing the borrower's name, and so 
on. Indeed, many Canadians' funding needs are simply not large enough 
to warrant such efforts. However, there are many Canadian financial 
intermediaries engaged in minimizing these costs, including my own indus¬ 
try and the Canadian banks. Canadians are certainly frequent and heavy 
users of non-resident savings. Recent experience demonstrates very 
strikingly how, at times when Canadian domestic saving falls short of total 
investment requirements, the country simply expands its offshore bor¬ 
rowing. Canadians have generally been considered excellent risks by 
foreign lenders, who are impressed by our affluence, our economic diversi¬ 
fication, and our relative political stability. The extent to which the 


265 



foreign capital market can be a safety valve was demonstrated during 1981, 
when the Canadian chartered banks expanded their net foreign liabilities 
from $3 billion to $5.2 billion and their gross foreign liabilities by a stag¬ 
gering $40 billion, from $113 billion at the start of the year to $152.6 
billion at the end. Through Wood Gundy and its fellow investment 
dealers, Canadian borrowers offshore raised a further $9.5 billion net in 
bonds, shares, and short-term commercial paper in 1981, denominated in 
Canadian dollars as well as in foreign currencies. 

Figure 3 illustrates how the Canadian chartered banks were able, at a 
time of exploding demand for large business loans, to maintain their loans 
to small business at the same time. Much of this lending was funded 
offshore. 

Of course, foreign borrowing does carry an exchange risk. But I 
find it very difficult to believe that such problems account for the general 
antipathy felt in financial markets toward deficits. It may be that the 
aversion expressed by market participants to deficits is really the squawk 
of the overburdened taxpayer. If this is the case, then complaints about 
deficits are simply a reminder to ministers that there is no free lunch. 
Postponement of taxes through borrowing carries a heavy real interest cost 
these days, and it simply shifts the burden of paying for government 
outlays from today to tomorrow. At least some taxpayers are well aware 
that their tax burden will be augmented in the future by the full amount 
of the deferral plus the real interest rate, whatever that turns out to be. 

However, this explanation does not entirely satisfy. 

We don't really know very much about how the burden of the accru¬ 
ing real interest on the national debt is perceived by capital market partici¬ 
pants. Some of them, conceivably, might prefer to bear tax now, rather 
than wait and pay the real rate of interest on the accruing national debt. 
On the other hand, capital market participants may simply represent the 
group of taxpayers who feel that they will receive less than an adequate 
return on their tax dollar, whether present or future, and who are simply 
and sourly pointing out to the minister that they cannot be appeased by 
postponement of their tax bill. 

It can certainly be argued that participants in the financial markets 
are likely to be relatively free of debt illusion. That is, they are more 
likely than most people to recognize that tax deferral carries a cost, 
namely the real cost of interest. However, it is interesting to consider 
the possibility that the people in financial markets who strongly oppose 


266 



Figure 3 

Business loans as percentage of gnp 



SOURCE: Bank of Canada (series discontinued in 1981). 


deficits also represent the class of taxpayers who strongly prefer current 
to deferred taxation. In other words, the tax burden associated with 
government outlays and transfers may excite less concern among market 
participants than the decision to postpone this burden. This possibility 
may seem rather unlikely, but we must consider the fact that the interest 
burden of the national debt, relative to the GNP, is currently at an all- 
time high (see Figure 1). We must recognize that there are groups of 
taxpayers who might well wish to reduce this interest burden even if such 
a reduction involved paying higher current taxes. Who might these people 
be? 

For a start, those who expect their average tax rates to rise in the 
future might have a preference for paying now. This group includes many 
young, upwardly mobile individuals. 


267 







Second, people who have few attractive alternative investment oppor¬ 
tunities at the present time might be relatively happy to pay tax now and 
save the cost of future interest on the public debt; on the other hand, 
those who expect to do relatively well in the stock or real estate markets 
would probably rather defer the tax bite, since they would expect to be 
able to pay the interest on the national debt out of their superior private 
gains. 

A third group of taxpayers might simply be relatively indifferent 
between purchasing power today and purchasing power tomorrow. Such 
individuals would presumably rather save the real interest cost associated 
with a mounting public debt by paying taxes now. 

However, I think we would most likely find that the antipathy of 
financial market participants to deficits is largely an expression of the 
opinion of highly independent, highly enterprising, so-called 'self-made' 
individuals, people who expect to receive less government services than 
they pay for, people who also happen to be adept at piercing financial 
veils and at understanding the code language of financial statements, 
people who are expert at assessing the burden of debt, private or public. 
A most fascinating study could be made of the characteristics of financial 
market participants - a study that would help us more fully understand 
the phenomena we are exploring today. 

What we have seen so far is that there is indeed a legitimate reason 
for continuing concern by bondholders and their representatives about the 
threat of future unanticipated debasement of the currency. Bondholders 
are not interested in paper returns; they are interested in the purchasing 
power of their returns, and the greatest risk they face from governments 
is the repudiation, overt or covert, of the real interest burden of the 
debt. Crowding-out, on the other hand, is unlikely to be the real issue. 
Financial crowding-out is not the same thing as physical crowding-out, and 
anyway Canadian fiscal policy, unless it is highly irresponsible, is not able 
to deny creditworthy Canadian borrowers access to real resources as long 
as they are prepared to pay the going international rate of interest for 
credits of a given class. We have seen that complaints about deficits 
emanating from financial markets might stem simply from the antipathy to 
additional taxation of individuals who are particularly free from illusion 
about the real cost of tax deferral. They could even stem from individuals 
who have particular reasons to prefer being taxed today to being taxed 
tomorrow, given their anticipation of the real interest cost of adding 


268 


further to the national debt. 

The inflation theme is the important one. Unanticipated reduction in 
real interest returns due to unexpected monetary debasement is the real 
threat for bondholders. We can think of the bargaining between bond 
buyers and the government as a struggle in which the latter holds an 
excess of power. Potential bond buyers will make their choices on the 
basis of the alternative opportunities open to them, including the real rate 
of interest they can expect on the international capital markets, with all 
the uncertainties associated with those alternatives. The Canadian govern¬ 
ment must compete with those alternatives in offering an expectation of 
real return, but it always has an opportunity, after the bonds have been 
placed, to change the real rate of return from the rate that had been 
anticipated. Specifically, the government can unexpectedly debase the 
currency. Thus, in the negotiations surrounding the issue of government 
debt, both parties engage in a complex assessment of each other's respon¬ 
ses. The government will wish to provide as credible a facade as possible 
for its anti-inflationary posture, hoping to induce bond purchasers to 
reduce their insurance premiums against future debasement. The bond 
buyers, on the other hand, will take past experience into account in 
setting their inflation premiums, and if that experience includes a series of 
anticipated or unanticipated episodes of monetary debasement and resulting 
inflation, they will charge higher inflation premiums. 

This description of the strategies on both sides is, of course, incom¬ 
plete and overly simplified. For one thing, in the real world the parties 
are not independent of each other. Most of the people to whom the gov¬ 
ernment wishes to sell bonds are also taxpayers. The 'game' is therefore 
horribly complicated. However, I think that here and now, in 1983, the 
market's suspicions about unexpected future monetary debasement are 
extraordinarily high, greater in fact than they need to be. As far as 
short-term securities are concerned, at least, the probability is very low 
that inflation will be so high that an investor today would earn only a low 
real return on them. In fact, at this point the government has little hope 
of cheating the financial markets, because it can only reduce the inflation 
insurance premium to low levels by pursuing a policy that amounts to 
pleasantly surprising the market over an extended period of time. Past 
monetary debasement has caught up with the government, and it is still 
exacting a compensatory risk premium in the interest rate on new bor¬ 
rowing . 


269 


If the capital markets are enforcing monetary discipline on govern¬ 
ments now, it is a discipline that, paradoxically, is strengthened, not 
weakened, by the existence of a huge ongoing deficit. After all, if the 
government were now running a balanced budget or a surplus, and ex¬ 
pected to continue to do so, it could be relatively indifferent to the 
market's insurance premium. In fact, it has no option but to submit itself 
to the market's judgment, and it can only reduce the investors' premium 
by persistently surprising them with the effects of monetary restraint. 

This suggests the proposition, which some may find amusing, that the 
best time for a government to adopt unanticipated monetary expansion is 
not when its deficit is high, but later, when through a succession of 
surpluses it has reduced the potential costs to itself of exciting a new 
increase in investors' inflation premiums. As long as the government 
requires a great deal of borrowed money, it faces a financial market that 
may punish it for previous bad behaviour. Once it has reached a state of 
budget balance, once it has locked-in the holders of the outstanding debt, 
the debt holders are at its mercy. A bout of unexpected debasement at 
that point would reduce the real cost of servicing the outstanding debt, 
and if the budget remained in balance - so goes the argument - the insur¬ 
ance premium would never have to be paid. 

Therefore, we can argue that participants in Canadian capital markets 
are unnecessarily paranoid at the moment. Their very wariness will en¬ 
force good behaviour. 

This assessment is reinforced by the fact that the existing federal 
debt is for the most part relatively short-term debt. A minister of finance 
who ordered his central banker to pump up the money supply would 
quickly find that a sizeable chunk of the outstanding debt would have to 
face refinancing at high inflation-insurance premiums. 

Figure 4 illustrates the evolution since 1955 of the maturity structure 
of the national debt. Financial market participants have protected them¬ 
selves against unanticipated inflation by shortening term. It is interesting 
to observe that the average term of the federal debt in the hands of the 
general public, not including Canada Savings Bonds, reached a modern 
minimum of 5 years and 9 months in mid-1975 after a stiff bout of infla¬ 
tion, then slowly wound its way back up to 10 years and 4 months in late 
1979 as inflation fell, and since then again has been falling, to 6 years 
and 6 months at the end of November 1982. 

Actually, even the government, not just the holders of the debt, has 


270 


Figure 4 

Average term to maturity of federal marketable securities held by the general 
public, and Savings Bonds and Treasury bills as percentage of federal loans 
and securities held outside the Bank of Canada 



SOURCE: Bank of Canada. 


an interest (from a different perspective) in reacting to unanticipated 
inflation by shortening the term structure. Given the lag between mone¬ 
tary expansion and the resulting inflation, the holders of debt maturing 
in, say, up to two or three years can be reasonably certain that they will 
not lose much of their anticipated real return as a result of unexpected 
currency debasement; thus the inflation insurance premiums on short-term 
debt should be lower, and thus less onerous for the government, than 
those on long-term securities. 

The thin line in Figure 4 shows the behaviour of the average term to 
maturity of the government's unmatured direct and guaranteed securities, 
including Treasury bills but not Canada Savings Bonds and perpetuals, 
held by the general public (that is, outside the Bank of Canada, the 


271 





chartered banks, and government accounts). The solid line shows, 
for the total of federal government loans and securities outstanding not 
held by the Bank of Canada, the percentage represented by Treasury bills 
and Canada Savings Bonds. 

Both measures tell a similar story. As inflation climbed during the 
1960s and early 1970s, the government permitted its debt to shorten in 
term. After inflation began to fall, in 1975, a successful campaign was 
carried out to lengthen the term of the debt and put less reliance on 
short-term instruments. However, since the re-ignition of inflation in 
1979, the average term of the debt has once again shortened quite drama¬ 
tically, and there has even been a major return to the Canada Savings 
Bond market. 

All this suggests to me that the government would not be a big 
winner from a covert or unanticipated effort to reflate. In fact, I would 
argue that the government's past behaviour has built up such paranoid 
suspicions in the marketplace that its only hope for cheaper borrowing 
costs is to pursue a credible and sustained anti-inflationary monetary 
policy over the years. This will not prevent it from having to pay a real 
interest rate competitive with other borrowers, but it will tend at least to 
reduce whatever insurance premiums paranoid investors impose, and it will 
allow the government greater freedom to lengthen the term of the debt 
structure if that seems to be appropriate. 

Let me turn to a different question. Have deficits over the last 
twenty-five years actually led governments in Canada to accelerate mone¬ 
tary expansion? I do not propose to examine this carefully, but Figure 5 
provides us with a quick appreciation of the facts in the case. We see 
here, quarter by quarter, the federal deficit (on a National Accounts 
Basis) as a percentage of the GNP, seasonally adjusted at annual rates, 
and the four-quarter rate of expansion of narrow money, M-l. It is 
immediately clear that there has been no simple relationship between defi¬ 
cits and the rate of monetary expansion. Certainly, one cannot pin the 
blame for the massive monetary explosion of 1970-2 on the deficits of that 
day. In fact, during the early part of that explosion, the federal govern¬ 
ment was more or less continuously in surplus! On the other hand, some 
in this audience will remember that James Coyne, the Governor of the 
Bank of Canada during the Diefenbaker years, was dismissed in effect 
because he wished to pursue a hard money policy, which the government 
of the day could not stomach. It is evident from the chart that the swing 


272 


Figure 5 

Federal deficit as a percentage of gnp vs. four-quarter growth rate of M-1 
(shift adjusted from 81:1) 



SOURCES: Statistics Canada; Bank of Canada; forecast and shift-adjustment by 
Wood Gundy Limited. 


from a generous surplus in 1956 to a large deficit - for those days - in 
1958 was accompanied by a dramatic leap in monetary expansion. However, 
this leap had minimal inflationary consequences, probably because Mr. 
Coyne quickly took action to reverse the process and actually reduced the 
money supply during late 1959. Mr. Coyne's independence was, of course, 
intolerable to the government of the day. Canadian central bankers have 
learned a lesson from this experience, and now the Governor would be 
obliged to offer his resignation should he feel unable to carry out the 
monetary policy of the government. 

The last five years certainly demonstrate an association of deficits 
and inflation. However, it is not easy to argue in this case that the 
government's deficits have tempted the Bank of Canada to accelerate the 


273 















pace of monetary expansion. It is much easier to argue the converse: 
namely, that stalwart monetary deceleration, given the stubbornness and 
entrenchment of previously-awakened inflationary pressures and OPEC oil 
prices, has given us a major recession and a huge cyclical deficit, a 
deficit worsened by the exploding cost of interest in both real and nominal 
terms. In this view, monetary restraint causes deficits! However, the 
way out of the dilemma is not to re-inflate the money supply, but instead, 
if necessary, to use supporting mechanisms to hasten the decline of infla¬ 
tion itself. Given the rate of monetary expansion, this approach permits 
greater real activity, growth in tax revenues, and reduced unemployment 
insurance costs. In fact, the benefits of years of increasing restraint are 
finally beginning to appear. Inflation is falling rapidly, real economic 
activity is beginning to rise, workers and businessmen alike are taking 
steps to ensure that they will survive and prosper in a low-inflation 
world. But the capital markets demonstrate that heavy weight is still 
given to the possibility of re-inflation. In retrospect, we would probably 
have made a quicker adjustment to reality had the Bank of Canada adopted 
a severe rather than a gradual approach to monetary deceleration in 1975. 
Since inflation follows cumulatively upon the cumulative rate of growth of 
money over time, it would have been possible, and arguably less painful, 
for Canada to have eliminated inflation relatively quickly by the mid-1970s, 
when we were not so habituated to it. (Possibly, the experience of Mr. 
Coyne was in the minds of the government and the Bank when the decision 
to embrace gradualism was made in 1975. If so, it is not at all clear that 
they drew the right conclusion from his experience.) 

With the assurance that hindsight provides, I can argue that the 
Canadian experience of both inflation and deficits in the last decade has 
been fundamentally nothing but a symptom of our unrealistic ambitions. 
We failed to realize in time, in the early 1970s, that our productivity 
potential had begun to stagnate. Monetary stimulus in the early 1970s, 
the self-indulgent gradualism that followed it, and the mounting federal 
deficits that followed the gradualism can in retrospect all be viewed as 
symptoms of our unwillingness as a nation - or our inability as a nation - 
to confront realistically the stagnation of our physical productivity poten¬ 
tial (see Figure 6). 

To refer again to Figure 1, note that I have extended the historical 
data on the debt and the percentage of GNP required to service the 
interest on it through 1985, using our current economic forecast. In the 


274 


forecast I have assumed that the Bank of Canada will maintain the equi¬ 
valent of a stable growth of M-l this year, followed by renewed deceler¬ 
ation in 1984 and 1985 (see also Figure 7). We have reached a point in 
the evolution of monetary and fiscal policy where we are making very 
substantial progress against the rate of inflation. Wage pressures, the 
last of the major components in the inflation cycle, are now decelerating 
very rapidly. As we have seen, however, inflation insurance is still very 
much in evidence in the financial markets. If government has the courage 
and sense to maintain steadfast monetary policies, consistent with the low 
inflation rates we should achieve in the next few years (see Figure 8), we 
will manage to lock in those low rates of inflation and, at an increasing 
rate, find workers and companies pricing themselves effectively into work. 
Given appropriate monetary discipline, the deficits of the federal govern¬ 
ment will, in my view, also finally tend to look after themselves. As 
inflation falls, and as monetary discipline persists, the inflation insurance 
premium built into interest rates will decline, while the acceleration of real 
economic activity will generate strong real growth in personal and cor¬ 
porate income tax revenues for the government. I believe that in the 
decade to come voters will prefer to see the role of government in the 
economy continue to decline (see Figure 9). Despite some social priorities, 
such as the expansion of the western rail network and oil self-sufficiency, 
that may continue to require heavy infusions of public funds at the federal 
level, it seems very likely that with our low complement of dependent 
young and old, and our heavy complement of 25- to 40-year-old strivers, 
we will see a reduced demand for government services relative to the size 
of the economy as a whole. These assumptions taken together suggest to 
me that it may be relatively easy for the government to reduce its deficit 
in the late 1980s. However, the capital markets will only slowly reduce 
their inflation insurance premiums. They will remain vigilant against the 
threat of unexpected monetary debasement long after the behaviour that 
gave rise to their vigilance has vanished. 

To sum up, I think we would be wrong to place great emphasis on 
the temptation that deficits, or the costs of servicing them, represent for 
governments. In my opinion, a strong and reliable consensus has grown 
up in the last decade, not just in Canada, in support of the belief that 
the continuing costs to society associated with high and variable inflation 
rates are more onerous than the once-and-for-all costs associated with 
returning to a reasonably stable price level. The difficulty for the future 


275 


Figure 6 

Real gnp per employed person 



SOURCE: Statistics Canada to 82:111; forecast by Wood Gundy Limited. 


is not that Canadians will be unwilling to complete the transition to stable 
price conditions, but that the Bank of Canada may have some forecasting 
problems in trying to decide what rate of money growth to aim at in order 
to assure reasonable price stability. We have very inadequate information 
about our stock of capital, about its rate of obsolescence, and about its 
ability to meet the changing mix of product demands as the 1980s continue. 
We are uneasily aware that many carefully-developed labour skills will 
require substantial reinvestment or may even have to be abandoned as 
microtechnology substitutes across the board for many heretofore human 
functions. It will be difficult, therefore, for governments and central 
bankers to gauge the capacity of our physical and human capital, the 
'supply side', to respond to any given growth rate of nominal demand 


276 




Figure 7 

Growth in nominal gnevs. growth in ‘shift-adjusted’ M-1 
(four-year growth rates) 



NOTE: Policy assumption is that, after adjusting for shifts in demand for M-1 since 1980, 
the annual 3rd quarter to 3rd quarter growth rate of M-1 will be held to 6 per cent in 1983, 
5 per cent in 1984, and 4 per cent in 1985. 

SOURCES: Statistics Canada; Bank of Canada; forecast and shift-adjustment by 
Wood Gundy Limited. 


without departing from reasonable price stability. 

On the whole, I expect that central bankers and governments will try 
to reflect this supply-side uncertainty by maintaining generally cautious 
monetary and fiscal regimes. The message that the capital markets are 
trying to deliver is, in my opinion, one that Ottawa has accepted for some 
time, despite appearances to the contrary. Of course, it is impossible for 
government to eliminate its deficit in the next few years. The very effort 
to do so would thwart itself. There is, however, room to begin taking 
fiscal actions that would shift the tax burden from future taxpayers back 


277 




Figure 8 

Various measures of Canadian inflation 



SOURCES: Statistics Canada; forecast by Wood Gundy Limited 


to the present, but such actions would have to be directed specifically at 
those most capable of bearing their weight. As far as the capital markets 
are concerned, the most important treatment for deficits is likely to be 
steadfast monetary restraint: a slow-acting cure, but a certain one. 


278 





Figure 9 

Government expenditure on goods and services 
(not including transfer payments) as percentage of gne 



SOURCES: Statistics Canada to 82:111; forecast by Wood Gundy Limited. 


Roger Keane* 

In assessing John Grant's work on the issue of deficits and capital 
markets, one can quickly identify several winning points. First, it is true 
that differentiating in the strict sense between private and public deficits, 
especially federal government deficits, has led to unnecessarily worried 
markets. Nevertheless, the paramountcy of market interpretations cannot 
be ignored. Second, the crucial difference between financial crowding-out 

* Vice-President and Chief Economist, Midland Doherty. 


279 




and physical or real resource substitution has been correctly stressed. 
Third, the appropriate policy prescription for markets and indeed the 
economy, the implementation of a non-debasement monetary policy, may yet 
be initiated. However, like most good analyses of complex questions, Dr. 
Grant's paper may have opened more doors than it has closed. I would 
like to pursue and expand on some of these themes in the interest of both 
adding perspective and emphasizing the importance of the questions raised. 

I propose to comment on four areas: those rigidities within markets 
that determine the so-called 'paranoia premium' in interest rates, the 
impact of irresponsible behaviour, the types of crowding-out, and finally 
the appropriate policy description. 

As a starting point, I would like to emphasize the difference between 
the theorist's world and the real world, at least the real world of capital 
markets. For example, in the theoretical world, supply curves are 
assumed to be continuous (ending only at the margin of Samuelson's text). 
In practice, supply curves do not go on forever. They do stop. Con¬ 
sider the example of a hypothetical Argentine financing, which I propose 
to offer this audience at the exceedingly attractive rate of 'Libor plus 7 
per cent'. How many buyers would there be? None. 

The point is that because economic theory is highly aggregated it 
tends to assume away such problems. Indeed, in the real world there is 
no price for our hypothetical issue. Whether this relates to the 'paranoia' 
of the game players (or whatever) is irrelevant, since the rigidities are 
there. More likely, the experience of the participants (expectations) has 
led to such conclusions. This topic represents a key element in the dis¬ 
cussion. 

RIGIDITIES 

Rigidities in capital markets develop on the basis of a mix of expectations 
and perception. They commonly take the form of quotas, credit limits, or 
other reflections of attitude and can regularly disrupt what ought to be 
(by theory) an efficient and unbiased allocation of capital. 

An examination of offshore borrowing potential will clarify this point. 
Whereas Canadian business and governments have had free and easy access 
to foreign markets in the past, it is no longer clear (a) that those markets 
will have unused capital or (b) that real-world risks and rigidities may not 
now discourage Canadian borrowers, effectively crowding them out. 


280 


There are four specific concerns: 


1. The size of government financial needs in the United States and 
Europe may have reduced the capital available for foreigners. 

2. Besides the obvious danger of volatile international exchange rates, 
twenty-year securities cannot be hedged in futures markets (one year 
is the maximum protection). 

3. Medium-sized Canadian companies are not known in external markets 
(e.g.. Dominion Stores). 

4. The international banking system limits the credits available to any 
one country. If governments utilize the available credit lines, busi¬ 
ness will have no access. 

Each of these considerations raises the spectre of greatly limited credibility 
and capacity for Canadian borrowers, especially corporations, in foreign 
markets. 

I would offer three other examples of the substantive impact of the 
rigidity phenomenon. While there is no theoretical difference between the 
federal government's and a corporation's, say IMO's, requiring $1 billion of 
bond funds, the market may respond negatively to the perceived presence 
of government needs. Such responses imply that the markets do in fact 
differentiate between the types of financing. This differentiation can 
presumably be traced to the fears of individuals (who compose the market) 
that deficits represent a lack of discipline in that they inherently encour¬ 
age the postponing of costs. Given the assumed desirability of delaying 
costs from a political standpoint, it is logical that this process should be 
thought to lead to higher spending and, inevitably, higher deficits. 
Whether this proves to be the end result or not, the market's judgment, 
as the highest court on these matters, is correct. 

Another rigidity or perception challenges the well-established dictum 
(Stigler) that appropriate transaction prices can be found at all times. At 
what price would Dome Petroleum be able to raise bond monies today, 100 
per cent or 200 per cent? The answer is at no price. This would be the 
case not because the price was not a sensible one, but because people 
would not buy the bonds. Again, the market has made a judgment and, 
uninformed or not, this judgment is an indelible fact. 

The final example of the pervasiveness of rigidities relates to small 
business. Despite the evidence in Dr. Grant's graph depicting bank 


281 


lending at all levels of business, this represents another aggregate 
picture. In recent years, small business in Canada has had access only to 
floating rate bank loans. Given the illiquidity of corporations (62 per cent 
of adjusted cash flow went to interest payments as of the second quarter 
of 1982), such funds, while available, are inappropriate for many small 
businesses. Such discomfort factors (conditions possibly accentuated by 
rising deficits) are capable of crowding out potential borrowers. 

IRRESPONSIBLE BEHAVIOUR 

The size of government deficits is not necessarily the only indication of 
irresponsible behaviour. Surely, a debt management policy that contrived 
the issuance of 19.50 per cent Canada Savings Bonds can be viewed as 
irresponsible. The reality is simple. No private corporation could compete 
with such largesse, not because it was prevented from paying a premium 
interest rate but because it would not contemplate this gesture (let alone 
this maturity structure). This debt management program substantially 
altered the capability for private term fundings. Moreover, it impacted 
expectations on the type of offerings that could be anticipated from gov¬ 
ernment. 

FINANCIAL VERSUS PHYSICAL CROWDING-OUT 

Most commonly, crowding-out is a financial phenomenon. However, it 
seems to this observer that while a difference exists between the two forms 
of crowding-out, financial displacement can produce physical crowding-out. 
Assuming that a lack of funds has inhibited corporate expansion, and 
given the government's ability to expand (fiscal policy) and its unique 
capability to finance (monetary policy), there exists the temptation for 
governments to occupy resources that would otherwise be used privately. 
This may be seen as a moral obligation; nevertheless, an occupation of 
private resources will occur. 

Perhaps, in this regard, it is a positive development that capital 
markets are skeptical about the implicit deficits. 

POLICY PRESCRIPTION 

Despite the desirability of John Grant's recommended 'steadfast monetary 


282 


restraint' the problems associated with its attainment could be consider¬ 
able. Given the presumption of ongoing high risk premiums in interest 
rates, this policy course risks higher interest rates in the short term, 
which may choke off recovery. Avoidance of this politically undesirable 
outcome could mean return to the dangers of the old stimulative religion. 

More to the point, however, problems in capital markets seem to 
relate primarily to the unwillingness of government to alter its rapid deficit 
expansion strategies of the last ten years through either tax increases or 
expenditure curtailment. This policy direction and the related expedient 
debt management policy have justifiably upset capital markets. If not 
addressed, this policy course will not only guarantee the maintenance of 
high risk premiums in interest rates but also risk validating those prem¬ 
iums by rekindling inflation. On this basis, the market may be conscien¬ 
tiously right after all! 


283 


Government deficits: historical analysis 
and present policy alternatives 

John McCallum* 


This paper presents an analysis of Canadian fiscal policy in the past and 
then considers options for the present and future. Recently there has 
been renewed interest in the Great Depression of the 1930s, and several 
commentators have wondered aloud whether that experience might ever be 
repeated. In part, the answer may depend on fiscal policy, and following 
the development of an analytical framework in the first section of the 
paper, the second section considers the role of fiscal policy in the 
thirties. What was the stance of fiscal policy, and how much did it 
contribute to (or alleviate) the Depression? How much less severe would 
the Depression have been if today's automatic stabilizers had been in place 
and/or today's approach to discretionary fiscal policy had then prevailed? 
Or, to turn the question around, how much worse would our present 
recession be if we had followed 1930s-style fiscal policy in 1982? Answers 
to these questions may shed light on the importance of fiscal policy, as 
well as on the likelihood of another depression. 

The third section presents an analysis of fiscal policy since the 
Second World War, with emphasis on the past decade or so. The key 
issues addressed are whether we now have a structural deficit and whether 
fiscal policy over the past decade has stabilized or destabilized the 
economy. To answer this last question, simulations were conducted to 
estimate what would have happened if the federal government had pursued 
a 'fixed fiscal rule' rather than the policies it actually conducted. Based 
in part on the historical analysis, the fourth section considers present 
policy options and presents conclusions. 


* Professor, Department of Economics, Universite du Quebec a Montreal. 


284 


ANALYTICAL FRAMEWORK 


The standard national accounts budget balance may be written as: 

B = T - G - iD, (1) 

where B is the nominal budget balance or surplus, T is nominal taxes 
minus transfer payments, G is nominal government spending on goods and 
services (current and capital), i is the nominal interest rate on govern¬ 
ment debt, and D is the nominal value of the government's net interest 
bearing debt. Taxes-minus-transfers may be divided into two categories: 
the non-cyclical component that would prevail at 'normal' levels of economic 
activity and the cyclical component due to deviations of nominal GNP (y) 
from its normal level (y). These two components may be designated res¬ 
pectively T and -z(Y-Y), where z is the marginal tax-minus-transfer rate. 
A second decomposition is to rewrite iD as (r + n)D, where n is the infla¬ 
tion rate and r is the ex post real interest rate defined as i - n. Using 
these two compositions, the budget balance B may be written in the alter¬ 
native form: 

B = T- G-rD-nD- z(y-y). (la) 

The conventional cyclical adjustment to budget balances is given by 
the last term of equation (la), while the conventional inflation adjustment 
is given by the second to last term. Thus the cyclically- and 
inflation-adjusted budget balance may be written: 

B A = T - G - rD. (2) 

Two questions may now be raised. First, is equation (2) an appropriate 
measure of the 'structural' budget balance? If B = 0, then real govern¬ 
ment debt will be constant over the business cycle if deficits are entirely 
bond-financed. However, if one defines neutrality as a constant ratio of 
government debt to GNP and if one also allows for high powered money 
creation, then the structural budget balance becomes: 

B S = T - G - (r-g)D + (g+n)H, (3) 


285 



where g is the growth rate of real GNP and H is nominal value of high- 

S 

powered money. If B =0, then the ratio of government debt to GNP will 
be constant over the long run, although because of the automatic sta¬ 
bilizers the debt ratio will rise during recessions and fall during booms. 

The second question is whether equation (2) gives an appropriate 
measure of discretionary fiscal policy. Three points may be made on this 

A 

issue. First, changes in the adjusted budget balance B are not neces- 

A 

sarily discretionary: for example,. B is affected by the real interest rate 
and the world price of oil, neither of which is subject to full government 

A 

control. Thus B may be viewed as the 'non-cyclical' component of the 
budget rather than the 'discretionary' component. Second, as a measure 
of fiscal stance the budget balance should be stated in real rather than 
nominal values, and this will make a difference to the extent that there are 
changes in relative prices over time. Finally, each component of the 
budget should be weighted by its multiplier or first-round impact on 
aggregate demand. 

Incorporating these amendments and using lower case letters for real 
variables, the budget balance as a measure of non-cyclical fiscal policy 
may be written: 

b F = (c-m)t - (1-m )g - (c-m)(l-t)(l-m D )rd, (4) 

where c, m, and t are the marginal propensities to consume, import, and 
be taxed, m is the import share of government purchases, and m^ is the 
percentage of government debt that is held by foreigners. It can be seen 
that the fiscal impact of changes in g is likely to be greatest, while the 
impact of changes in real interest payments is likely to be weakest, 
especially if a large proportion of the debt is held by foreigners. The 
variable nD is not included in (4) because this amount is required to 
maintain the real value of interest-bearing assets and hence should be 
fully saved. 

The next two sections apply this analytical framework to the thirties 
and then to the postwar period. A formulation such as (3) will be used 
for questions concerning structural deficits, while a formulation such as 
(4) will be used for issues relating to the fiscal impact of the government 
budget. 


286 


THE THIRTIES 


Between 1929 and 1933, Canada's real GNP and total employment both fell 
by 30 per cent, and it was not until the end of the decade that output 
and employment had recovered to their 1929 levels. Throughout the 1930s 
the unemployment rate varied between 9 per cent and 19 per cent. The 
purpose of this section is to analyse the impact of fiscal policy during this 
period and to compare Canadian policy with that of the United States. We 
begin with a description of the methods used to measure the stance of 
fiscal policy and then turn to a Canada-U.S. comparison and an analysis of 
the impact of fiscal policy. This is followed by an attempt to answer two 
hypothetical questions. First, to what extent would the Depression have 
been moderated if fiscal policy had been conducted in what might today be 
judged to be a 'reasonable' manner? Second, what would have happened if 
the discretionary policies of the thirties had remained unchanged but the 
automatic stabilizers of the 1980s had been in place during the thirties? 
The answers to these two questions permit a reponse to the third and final 
question: how much worse would our present recession be if fiscal policy 
in 1982 had been conducted along the lines of fiscal policy in the thirties? 

The estimates 


The basic data used in the analysis are set out in Table 1, which provides 
estimates of six components of the budget balance for both the federal 
government and the government sector as a whole. Each of these com¬ 
ponents is expressed as a percentage of potential GNP, which was 
estimated by applying a constant growth rate to the actual level of real 
GNP in 1929. Following the same procedure used by Brown (1956) for the 
United States, this growth rate was obtained by drawing a straight line 
between the actual levels of real GNP in 1929 and 1942. This procedure 
gives an annual growth rate of potential output of 3.9 per cent for Canada 
as compared with 3.2 per cent for the United States. 

Referring now to the columns of Table 1, g is real government 
spending on goods and services as a per cent of potential GNP, t is the 
non-cyclical component of taxes-minus-transfers excluding interest in the 
national debt. Interest payments are broken into two components: the 
inflation premium component 7id and the ex post real interest payments 
component rd. The fourth column (-zGAP) gives the cyclical component of 


287 




TABLE 1 

Components of government budget balances, as percentage of potential GNP, 
1929-42 


All governments 



“g 

t 

-rd 

-Ttd 

-zGAP 

RPE 

B 

1929 

-11.5 

13.8 

-2.8 

-.8 

0 

1.5 

.2 

1930 

-12.7 

12.6 

-5.2 

1.5 

-1.2 

1.4 

-3.6 

1931 

-12.0 

13.2 

-8.7 

4.7 

-3.1 

.8 

-5.1 

1932 

-10.5 

14.3 

-12.1 

7.5 

-4.4 

.3 

-4.9 

1933 

-8.3 

15.5 

-6.1 

1.4 

-5.7 

.2 

-3.0 

1934 

-8.6 

14.6 

-3.3 

-1.2 

-4.8 

.2 

-3.1 

1935 

-8.8 

14.7 

-3.8 

-.4 

-4.6 

.2 

-2.7 

1936 

-8.4 

16.3 

-1.1 

-2.9 

-4.7 

.3 

-.5 

1937 

-8.9 

16.3 

-1.6 

-2.1 

-4.3 

.2 

-.4 

1938 

-9.2 

14.9 

-3.4 

0 

-4.5 

.2 

-2.0 

1939 

-9.1 

15.6 

-4.2 

.8 

-4.0 

.3 

-.6 

1940 

-13.5 

19.6 

.3 

-3.4 

-4.2 

.4 

-.8 

1941 

-18.2 

23.1 

2.2 

-5.2 

-2.4 

1.2 

.7 

1942 

-35.5 

23.1 

0 

-2.9 

0 

0 

-15.3 


Federal government 



-g 

t 

-rd 

-TXd 


-zGAP 

RPE 


B 

1929 

-2.9 

5.3 

-1.5 

-.4 


0 

.4 


.9 

1930 

-3.1 

4.1 

-2.6 

.8 


-1.1 

.4 


-1.5 

1931 

-2.7 

4.1 

-4.2 

2.3 


-2.2 

. 1 


-2.6 

1932 

-2.3 

4.9 

-5.8 

3.6 


-3.1 

0 


-2.7 

1933 

-2.1 

6.5 

-3.0 

.7 


-4.1 

0 


-2.0 

1934 

-2.2 

6.2 

-1.6 

-.6 


-3.4 

. 1 


-1.5 

1935 

-2.6 

5.8 

-1.8 

-.2 


-3.1 

0 


-1.9 

1936 

-2.4 

6.9 

-.5 

-1.5 


-3.1 

. 1 


-.5 

1937 

-2.3 

7.1 

-.7 

-1.1 


-3.0 

. 1 


.1 

1938 

-2.5 

5.9 

-1.6 

0 


-3.0 

0 


-1.2 

1939 

-2.9 

7.0 

-2.0 

.4 


-2.6 

. 1 


0 

1940 

-8.1 

10.3 

.2 

-1.8 


-2.5 

.2 


-1.7 

1941 

-13.1 

15.4 

1.3 

-2.9 


-1.8 

.8 


-.3 

1942 

-31.0 

16.0 

. 1 

-1.8 


0 

-. 1 


-16.8 

SOURCES: 

: See appendix. 








taxes-minus-transfers, defined as losses 

of 

government 

business 

enter- 

prises plus the 

product of the output 

gap 

and the actual 

levels 

of 

revenue 

from cyclically 

sensitive 

taxes. The 

latter 

are defined 

as all 

direct and 


indirect taxes other than property tax. Expenditures on direct relief were 
not classified as cyclical in order to preserve comparability with the U.S. 


288 

















estimates. The sum of the first four columns of Table 1 is the National 
Accounts budget balance as a percentage of potential GNP measured in 
constant dollars (1947 prices), while the sixth column (B) gives the 
nominal budget balance as a percentage of nominal potential GNP. The 
relative price effect (RPE) is the difference between B and the constant 
dollar budget balance. The sources of these estimates and further details 
concerning them are given in the appendix. 

Turning to an examination of the data contained in Table 1, it can be 
seen that total real government spending on goods and services rose 
somewhat in 1930 but then fell from 12.7 per cent of potential GNP to 8.3 
per cent in 1933. Government spending did not recover until the war 
years. The cuts in spending were most severe at the provincial and 
municipal levels, where education and road-building were especially badly 
hit (see Royal Commission on Dominion-Provincial Relations). Full employ¬ 
ment taxes-minus-transfers also increased substantially after 1930, mainly 
at the federal level. Federal sales tax was increased from 1 per cent to 8 
per cent, income tax rates rose, and tariffs increased very substantially. 
Relief expenditures, which peaked at 1.6 per cent of potential GNP in 
1934, are treated as non-cyclical (i.e., included in t), and so it could be 
argued that the increase in t up to 1934 is understated by this amount. 
With consumer prices falling by 23 per cent between 1929 and 1933, the 
inflation premium component of interest payments was negative over this 
period and then positive as prices recovered after 1933. As a result, real 
interest payments of government debt were extraordinarily high in the 
early thirties, peaking at 12 per cent of potential GNP (18 per cent of 
actual GNP) in 1932. Finally, because the government expenditure price 
deflator fell by less than consumer prices until 1932, the relative price 
effect shown in Column 6 declined until that year but then remained stable 
for the rest of the decade. 

Fiscal impact and Canada-U.S. comparison 


As indicated in the first section, estimates of the fiscal impact of the 
deficit ought to be based on a weighted average of the various components 
(see equation (4) above). The results of such an exercise are reported in 
Table 2 and Figure 1. The budget balances reported are defined as: 


289 



TABLE 2 

Budget balances as percentages of GNP, selected averages 1929-42 



All 

governments 

Federal 

government 

Other 

governments 


bl 

bl 

b2 

bl 

bl 

b2 

bl 

bl 

b2 


US 

Cda 

Cda 

US 

Cda 

Cda 

US 

Cda 

Cda 

1929 

0 

0 

0 

0 

0 

0 

0 

0 

0 

1930-2 

-1.0 

-1.0 

-3.5 

-1.3 

-.6 

-1.9 

.3 

-.4 

-1.6 

1933-6 

.3 

3.5 

3.9 

-2.1 

1.3 

1.3 

2.5 

2.3 

2.6 

1937-9 

.9 

3.9 

3.7 

-1.3 

1.6 

1.5 

2.1 

2.3 

2.3 

1942 

- 

-17.3 

-16.4 

- - 

18.9 

-18.3 

- 

1.6 

1.9 


NOTE: 1929 set equal to zero. 

SOURCE: Table 1, Brown (1956). 

bl = -g + .8[t - (r+7t)d], 

b2 = -g + .8t - .5rd. 

The bl measure was chosen because it was available for both Canada and 
the United States. The measure implies first round leakages to saving, 
imports, and taxes equal to 20 per cent of disposable income, an amount 
that would be too small today but perhaps not inappropriate for the 
thirties. The measure b2 differs from bl in allowing for an inflation 
adjustment. The weight attached to nd is zero, while the weight applying 
to rd is .5. The latter figure is slightly less than the product of the 
coefficient on t (.8) and the proportion of government debt held by 
Canadian residents during the thirties (about 70 per cent). While b2 
seems theoretically superior to bl, only the latter measure was available 
for the United States. 

Figure 1 provides a Canada-U.S. comparison. It can be seen that 
while non-federal governments behaved in much the same (pro-cyclical) 
way in both countries, there was a marked difference between the two 
countries in terms of federal fiscal policy after 1932. In relation to a 1929 
budget balance set to zero, the federal bl over 1933-9 averaged a surplus 
of 1.4 per cent of potential GNP in Canada as compared with a 1.8 per 
cent deficit in the United States. The difference between the two is a 
rather substantial 3.2 per cent of potential GNP, or 4.7 per cent of actual 
GNP. It should be pointed out, however, that a part of this difference 
may be due to differences in the method of calculating the full employment 


290 








Figure 1 

Fiscal policy in the thirties 


Adjusted 
budget 
balance 
(bl) as a 
percentage 
of potential 
GNP (1929 
set to zero) 



Adjusted 
budget 
balance 
(bl)as a 
percentage 
of potential 
GNP (1929 
set to zero) 



surplus, although efforts were made to achieve comparability. In terms of 
the Canadian government as a whole, in relation to 1929 there was a shift 
to surplus averaging 3.7 per cent of GNP over 1933-9, of which 1.4 per¬ 
centage points were due to the federal government and the remaining 2.3 
points to provinces and municipalities. Finally, the inflation adjustment to 
the deficit gives rise to substantially larger deficits in 1931 and 1932 but 
makes little difference in the other years. 

Alternative policies 


To what extent would the Depression have been mitigated by more sensible 
discretionary fiscal policy and/or the existence of today's automatic stabil¬ 
izers? Rough estimates of answers to these questions are given in Figure 


291 








TABLE 3 

Estimated effects of alternative policies, 1932-9 



(Average 

Actual 

values 
Case A 

1932-9, as percentages 
Case B Case C 

of GNP) 

Case D 

Output gap 

33% 

25% 

23% 

23% 

19% 

Budget balances, 






all governments 






(i) t>2 

1.8 

-.7 

-1.4 

-1.5 

1.8 

(ii) Actual balance 

-3.2 

-4.8 

-5.1 

-5.5 

-7.2 

Current account 

1.2 

0 

0 

-.1 

-.9 


NOTE: The four cases are defined in the text, and the methods used to 

estimate these results are given in the appendix. The output gap and the 
adjusted budget balance b2 are given as a percentage of potential GNP, 
while the actual budget balance and the current account are given as a 
percentage of actual GNP. 

2 and Table 3. In terms of discretionary policy, no changes are consi¬ 
dered for 1930-1 (when most indicators suggest neutral or expansionary 
policy), but the following policy alternatives are considered for later 
years: 


(a) government spending maintained at 1929 levels in relation to potential 
GNP (11.5 per cent) over the period 1932-9; 

(b) federal full-employment budget balance (bl) set equal to U.S. federal 
balance (as per cent of potential GNP) over the period 1932-9; 

(c) the all-government fiscal stance (b2) maintained at its 1929 level over 
the years 1933-9. 

To examine the role of automatic stabilizers, it was supposed that dis¬ 
cretionary policy remained unchanged, but the marginal tax-minus-transfer 
rate was set equal to .44 rather than its actual average value of .14. In 
other words, it was assumed that a one dollar fall in GNP gave rise to a 
44-cent reduction in taxes-minus-transfers (as at the present time; see 
next section) rather than a 14 cent reduction. These four alternatives are 
respectively designated Cases A to D in Table 3. 

For cases A to C, the impact of the alternative policies on the adjus- 


292 







Figure 2 

Output gap under different policies, 1929-42 


Canada actual 
U.S. actual 



Canada actual 
Canada with 1981 
automatic stabili¬ 
zers (Case D) 
Canada with 1981 
automatic stabili- 



ted budget balance (b2) may be obtained directly from the data, and 
this amount times the fiscal multiplier gives the estimated effect on GNP. 
The multiplier would have been much larger in the thirties than today for 
several reasons: a tax-minus-transfer rate of .14 rather than .44, a ratio 
of merchandise imports to GNP of .13 rather than about .25, negative 
personal saving, and very little potential for crowding-out. Based on 
these facts and calculations, given in the appendix, it is suggested that 
an assumed government expenditure multiplier of three is not unreasonable 
for the 1930s. As is shown in Table 3, the three alternatives subtract 2.5 
to 3.3 percentage points from the average value of b2 over 1932-9 and 
hence reduce the average output gap by 8 to 10 percentage points. Since 
the actual output gap averaged 33 per cent of potential GNP, the alterna- 


293 















tive policies would have reduced the impact of the Depression by a quarter 
to a third - not a huge effect but certainly significant. Also, it can be 
seen from Figure 2 that Canada's output gap would have followed that of 
the United States even more closely than it actually did if Canadian federal 
fiscal policy had been as expansionary as U.S. policy. 

It would seem too that there were no overwhelming technical obstacles 
to such a policy. As is indicated in the table, the actual budget deficit 
for all governments combined woulql have increased from about 3 per cent 
of actual GNP to 5 per cent. The current account of the balance of pay¬ 
ments would have been moved from an average surplus of just over 1 per 
cent of GNP to a position of balance. (See the appendix for the methods 
used to obtain these estimates.) It seems clear, then, that the barriers to 
such policies lay with the economic thinking and prevailing ideology of the 
day. 

Had the automatic stabilizers of today existed during the thirties, the 
multiplier and hence the depth of the Depression would have been reduced 
by just over 40 per cent. The adjusted budget balance b2 would not have 
been changed, but the actual government deficit would have increased from 
an average of 3.2 per cent of GNP to an estimated 7.2 per cent over the 
period 1932-9. A final possibility to consider for the thirties is a com¬ 
bination of alternative discretionary policy plus present day automatic 
stabilizers. A combination of today's automatic stabilizers plus approxi¬ 
mately neutral discretionary policy is about the policy combination being 
pursued in 1982, and so the impact of this combination on the Depression 
gives an estimate of what might have happened had fiscal policy as a whole 
been conducted along present-day lines. Allowing for the fact that the 
multiplier would have been smaller, a combination of Cases C and D would 
have reduced the average output gap from 33 per cent to an estimated 13 
per cent of potential GNP. This suggests that the absence of both auto¬ 
matic stabilizers and a neutral discretionary fiscal policy made the Depres¬ 
sion very much worse than it would otherwise have been. The impact 
would be even greater if one also allowed for a moderating effect of U.S. 
automatic stabilizers on the U.S. depression and hence on Canada. 

Finally, we may turn the question around and ask how much worse 
the 1982 recession would have been had we followed 1930s-style fiscal 
policy. Assuming an actual 1982 output gap of 8 per cent of potential 
GNP, a shift to 1930s automatic stabilizers might have increased this to 11 
per cent, while a shift to discretionary fiscal restraint of, say, 2 per cent 


294 



of GNP might have increased the output gap by a further 4 percentage 
points. This implies a 12 per cent drop in real GNP in 1982 instead of the 
actual 5 per cent. The largest annual drop in real GNP recorded in the 
thirties was 13 per cent in 1931. These rough calculations suggest that 
the shock to the Canadian economy in 1982 may have been of the same 
order of magnitude as the shocks of the early thirties and that the dif¬ 
ference in overall fiscal policy has been a key factor in explaining why the 
economy has not this time fallen into major depression. 

THE POST-WAR PERIOD 

This section begins by setting out estimates of government deficits and 
then turns to consider three issues: whether or not there is a 'structural' 
deficit, whether fiscal policy has been counter-cyclical, and what the 
consequences would have been if the government had not conducted dis¬ 
cretionary counter-cyclical policy in the seventies but had instead followed 
a 'fixed fiscal rule.' 

The estimates 


Table 4 sets out estimates of the components of government budget balan¬ 
ces over the period 1971-81 (see the appendix for the same information 
over the period 1954-70). These estimates are given in the same form as 
the information for the thirties set out in Table 1, except that, for rea¬ 
sons to be discussed shortly, the components of the budget are given in 
nominal rather than constant dollar values. Two comments on the estimates 
may be made at this point, the first relating to the cyclical adjustment and 
the second to the relative price effect (RPE). 

Originally, the intention was to use Department of Finance estimates 
of the cyclical adjustment, but for reasons discussed in the appendix the 
Finance estimates of potential GNP seemed unsatisfactory for the years 
after 1976. Consequently, the procedures used to calculate the cyclical 
adjustment were as follows: (i) estimates of the output gap were based on 
an Okun's Law relationship with the unemployment gap, as estimated by 
Fortin and Phaneuf (1979); and (ii) estimates of government revenue and 
unemployment insurance payments that would be forthcoming at potential 
output were based on Department of Finance information. The details of 
these procedures are set out in the appendix, and the estimates of the 


295 



output gap are shown in Figure 3 below. 

The second comment about the estimates has to do with the relative 
price effect shown in Table 4. In principle, government spending and 
revenue ought to be measured in real terms if one wishes to measure the 
impact of fiscal policy on aggregate demand, but in practice this is dif¬ 
ficult to do because of the unsatisfactory nature of existing price indices 
for goods and services purchased by government. Because public sector 
output is difficult to measure, these indices are essentially based on the 
growth rate of public sector nominal wages, implying that there is zero 
productivity growth in the public sector. Thus the measured price of 
government output relative to private sector output has increased con¬ 
sistently over the post-war period by an amount averaging about 2 \ per 
cent per year, or about the trend growth rate of private sector produc¬ 
tivity. If one is considering periods of any length of time, this relative 
price effect tends to dominate the analysis: for example, the all govern¬ 
ment actual budget balance in 1981 was a deficit equal to 1.2 per cent of 
GNP in current dollars, but measured in 1971 dollars it was a surplus of 
3.4 per cent of GNP (see Table 4). Thus the longer-term trend of fiscal 
policy on a constant dollar basis is bound to be in the direction of even 
larger surpluses. 

In light of these considerations, it would seem that the nominal deficit 
provides a better indicator of fiscal stimulus or restriction than the con¬ 
stant dollar deficit. Certainly this would be the case if one were inter¬ 
ested in the impact of policy on employment (as opposed to output). To 
the extent that the relative price effect comes from differences in recorded 
or actual productivity growth, a constant nominal adjusted deficit would 
imply a constant impact on employment, while other things being equal a 
constant real adjusted deficit would imply a steadily rising level of employ¬ 
ment in the public sector. The nominal deficit would also rise if public 
sector wages rose relative to private sector wages (or vice versa), but 
this would not necessarily detract from the nominal deficit as a measure of 
fiscal stimulus, since a higher (lower) relative wage in the public sector 
may be seen as a transfer to (tax on) public sector workers, with the 
usual impact on aggregate demand. Thus, while both real and nominal 
budget balances are presented, it is felt that the latter gives the better 
measure of fiscal stance. In any case, the differences between the two are 
not critical to an evaluation of federal fiscal policy since 1970, which will 
be our primary concern. 


296 



Figure 3 

Was fiscal policy counter-cyciical? 




Is there a structural deficit? 


As was indicated in equation (3) of the first section, the structural budget 
balance is given by the conventional inflation-adjusted budget balance (the 
sum of the first three columns of Table 4) plus an allowance for growth in 
the demand for real government debt as a result of economic growth plus 
an allowance for high-powered money creation. Assuming a long-run 
growth rate of 3 per cent and (for the sake of conservatism) a zero long- 
run inflation rate, these two adjustments increase the structural budget 


297 









TABLE 4 

Components of government budget balances as percentage of potential GNP, 
1971-82 


All governments 



-G 

T 

-rD 

-ttD 

-zGAP 

B 

RPE 

b 

1971 

-23.3 

28.0 

-2.6 

-1.2 

-. 7 

.1 

0 

. 1 

1972 

-23.2 

27.2 

-2.4 

-1.6 

. 1 

.1 

.6 

.7 

1973 

-22.7 

26.6 

-1.3 

-2.7 

1.1 

1.1 

.9 

2.0 

1974 

-23.0 

27.5 

- .1 

-3.6 

1.1 

1.9 

2.1 

4.0 

1975 

-23.8 

26.1 

-1.7 

-2.2 

-.9 

-2.4 

2.8 

.4 

1976 

-23.5 

26.5 

-2.3 

-2.0 

-.4 

-1.7 

3.9 

2.2 

1977 

-23.7 

27.0 

-2.5 

-1.9 

-1.3 

-2.4 

4.2 

1.8 

1978 

-23.5 

26.7 

-2.8 

-2.1 

-1.5 

-3.2 

4.3 

1.1 

1979 

-22.5 

26.1 

-2.4 

-2.7 

-.3 

-1.9 

4.1 

2.2 

1980 

-22.6 

26.6 

-2.6 

-2.8 

-.8 

-2.1 

4.3 

2.2 

1981 

-22.9 

28.8 

-3.0 

-3.2 

-1.0 

-1.2 

4.6 

3.4 

1982 

-23.2 

28.9 

-4.2 

-2.7 

-3.7 

-4.9 




Federal government 



-G 

T 

-rD 

-7TD 

-zGAP 

B 

RPE 

b 

1971 

-5.8 

8.2 

-1.3 

-.8 

- .5 

-.1 

0 

-. 1 

1972 

-5.8 

7.4 

-1.2 

-1.0 

.0 

-.6 

.2 

-.4 

1973 

-5.7 

7.2 

-.5 

-1.6 

.8 

.3 

.2 

.5 

1974 

-5.8 

7.7 

.2 

-2.2 

.9 

.8 

.6 

1.4 

1975 

-5.7 

6.2 

-.9 

-1.4 

-.6 

-2.3 

.6 

-1.7 

1976 

-5.7 

6.5 

-1.2 

-1.2 

-.3 

-1.8 

.8 

-1.0 

1977 

-5.8 

5.6 

-1.3 

-1.2 

-.9 

-3.5 

.8 

-2.7 

1978 

-6.0 

5.2 

-1.4 

-1.4 

-1.1 

-4.6 

.8 

-3.8 

1979 

-5.2 

5.0 

-1.2 

-1.9 

-.3 

-3.5 

.7 

-2.8 

1980 

-5.0 

5.5 

-1.2 

-2.1 

-.6 

-3.5 

.7 

-2.8 

1981 

-5.3 

7.6 

-1.4 

-2.6 

-.6 

-2.4 

.9 

-1.5 

1982 

-5.5 

7.4 

-2.4 

-2.2 

-2.7 

-5.4 



NOTES 

B = -G + 

T - rD 

- 7tD - 

zGAP; RPE 

= b - B; 

Symbols 

are defined as 


follows (all as percentages of potential GNP): G - nominal spending 
on goods and services; T - non-cyclical component of taxes minus 
transfers excluding interest on public debt; (r + 7i)D - interest 
payments on public debt; zGAP - cyclical component of taxes minus 
transfers; B - nominal budget balance; RPE - relative price effect; 
b - budget balance in 1971 dollars. 


balance by 0.8 and 1.0 percentage points of GNP for the federal govern¬ 
ment and all governments respectively. Hence, using these figures plus 
the first three columns of Table 4, estimated structural balances as a per- 


298 
















centage of potential GNP were as follows: 


Federal 


All governments 


1981 

1982 


1.7% 

0.3% 


3.9% 

2.5% 


For the federal government the estimated structural budget balance in 1982 
is close to zero, while for all levels of government there is an estimated 
structural surplus of about $9 billion. 

It might be useful to repeat what exactly these figures imply and to 
consider their sensitivity to alternative assumptions. If the structural 
balance is zero, the ratio of government debt to GNP is constant over time 
if: (i) the economy is operating at potential output; (ii) there are no 

changes in government spending relative to GNP or in the structure of tax 
and transfer regulations; and (iii) there are no changes in the exogenous 
factors that affect the cyclically adjusted budget balance. In the present 
Canadian context, the most important of these latter factors are probably 
the real interest rate that must be paid on government debt and the world 
price of oil. 

Turning now to a sensitivity analysis, there are several factors that 
might be mentioned: 

The estimates of potential output are based on the assumption of a 
long-run unemployment rate of 6.6 per cent. Suppose instead that 
this rate is 7.5 per cent and that the economy was operating at full 
potential in 1981. In that case, the 1982 output gap would be 6.4 
per cent, a number that is a very conservative estimate when it is 
recalled that real GNP fell by 5 per cent and that unemployment rose 
at an average rate of 3 per cent per year between 1971 and 1981. 
Under these conditions, the 1982 structural balances become a deficit 
of .3 per cent of potential GNP (about $1 billion) for the federal 
government and a surplus of some $6 billion for the government 
sector as a whole. 

If one assumes a long-run inflation rate of 5 per cent rather than 
zero, the additional high-powered money creation adds about $1 billion 
to the structural surpluses. 


299 




Most of the 1981-2 reduction in structural surpluses resulted from an 
increase in the real interest rate that had to be paid on government 
debt. If real interest rates fall to historical levels, there will be a 
further increase in structural surpluses. 

On the other hand, under present regulations a fall in world oil 
prices would reduce government revenues, while the exclusion of the 
net income of government pension plans from the calculations, as 
suggested by Bossons and Dungan (1983), would subtract about one 
percentage point from the all-government surplus. The latter factor, 
however, would have a negligible effect on the federal balance. In 
any case, despite these qualifications, it seems clear that the govern¬ 
ment sector as a whole has a sizable structural surplus, while the 
federal government has close to a balanced budget. 

Has fiscal policy been counter-cyclical? 

A partial answer to this question is provided in Figure 3, which plots two 
alternative measures of the federal deficits as well as estimates of the 
output and unemployment gaps. Policy may be defined as counter-cyclical 
if the deficit increases as the output and unemployment gaps rise, and 
vice versa. The B A series in Figure 1 is the conventional inflation- 
adjusted deficit as defined by equation (2) and as given by the sum of the 
first three columns of Table 4. The other series B is given by: 

B F = -G + .6T - . 54rD, 

where the weights of .6 and .54 are estimates of the fiscal impact of 
changes in taxes and real interest payments relative to a weight of one for 
government spending. The smaller weight on rD relative to T reflects the 
fact that about 10 per cent of federal debt is held by foreigners. It can 
be seen that the two deficit series have moved in a similar fashion over 
the course of the past twelve years. 

It is clear from the graph that federal fiscal policy has been strongly 
counter-cyclical. Deficits fell during the boom years 1972-4, and then 
rose as unemployment rose until 1978. Between 1978 and 1981 unemploy¬ 
ment fell and so did the federal deficit, although the shift towards surplus 


300 



in 1981 was pro-cyclical if one judges by the output gap series. Finally, 
the deficits measures rose in 1982, although by very little indeed in com¬ 
parison with the size of the recession. For the period up to 1981, regres¬ 
sion equations indicate very significant counter-cyclical policy: one-point 
increases in the unemployment and output gaps were associated with incr- 

p L 

eases in the inflation-adjusted deficit (-B ) equal to respectively 1.17 and 
0.47 per cent of potential GNP. Similar, but somewhat smaller, responses 

P 

apply to the weighted budget balance B . As indicated in the appendix, 
provincial fiscal policy was slightly pro-cyclical over this period, offsetting 
10 to 20 per cent of federal actions. Fiscal policy between 1954-70 was, if 
anything, slightly counter-cyclical. 

To judge the significance of these results, one may compare the roles 
of discretionary policy and automatic stabilizers. The results just noted 
imply that when GNP fell by one dollar, the non-cyclical federal deficit 
rose by 47 cents. This effect is greater than that of the automatic stabil¬ 
izers of all levels of government combined. The joint impact of discretion¬ 
ary policy and automatic stabilizers was an increase in the federal deficit 
of over 70 cents for every dollar reduction in GNP relative to potential 
(and vice versa). Allowing for pro-cyclical provincial policy, the corres¬ 
ponding figure for provinces and municipalities was about 5 cents. 

Three further points on this subject may be noted. First, the pat¬ 
tern of federal policy depicted in Figure 3 was not always the result of 
discretionary changes. For example, the tightening of policy in 1973-4 
and the loosening in 1982 were largely the result of reductions and then 
increases in real interest rates - although one can always point out that 
the resulting change in the deficit was discretionary in the sense that 
nothing was done to offset the impact of the real interest rate changes. 
On the other hand, the stated reason for the shift to deficits in 1975 was 
a desire to stabilize the economy. The reduction in the deficits over 
1979-81 may have been carried out partly as a result of a belief that the 
economy was strengthening, although anti-inflation efforts presumably 
played a role as well. In any case, the primary object is to examine the 
after-the-fact effects of federal policy rather than to ask whether the 
implicit fiscal role was in fact explicit. 

A second point is that this counter-cyclical pattern does not prove 
conclusively that policy was stabilizing, because fiscal policy operates with 
a lag. This issue is taken up in the next section, which estimates the 
effect of a hypothetical non-cyclical 'fixed fiscal rule.' Third, it is clear 


301 



that the implicit fiscal rule in operation between 1971 and 1981 was aban¬ 
doned in 1982. What would have happened if fiscal policy in 1982 had 
followed the rule of the past decade, and is it too late to return to that 
rule in 1983? This question will be considered in the final section. 

Consequences of a 'fixed fiscal rule 1 

What would have happened if all of the automatic stabilizers had remained 
in place but there had been no discretionary federal fiscal policy? To 
answer this question, we consider the case where the federal budget 

A 

balance B is held at its average value of -0.7 per cent of GNP over the 
entire period 1971-81. This implies than taxes would have been raised (or 
spending reduced) in the years of deficits, and the opposite would have 
been true in the years of surpluses. Because most of the variations in 
federal policy have taken the form of tax or transfer changes rather than 
spending changes, it will be assumed that all of the adjustments to the 
deficit occur through tax changes. 

Two estimates of the impact of such a fixed fiscal rule are presented. 
The first is based on simple formulas that have the advantage of being 
readily understandable, while the second estimate, based on the Infor- 
metrica econometric model, has the advantage of taking account of more 
complex interactions. The first set of estimates was derived from the 
following formulas: 


(5) 

( 6 ) 
C7) 


DT = B - B, 

DYGAP = -. 6DT - . 2DT_ 1 - .1DT_ 2 , and 

DUGAP = -.35 DYGAP - .15 DYGAP_ X - .10 DYGAP 


For each year, the change in taxes as a percentage of GNP (DT) was set 
equal to the mean value of the federal budget balance (-.7) minus the 
actual value. The impact of the tax change on output (DYGAP) is spread 
over three years, with a first-year multiplier of 0.6 and a cumulative 
multiplier of 0.9. The effects of output changes on unemployment 
(DUGAP) are also spread over three years, with first year and cumulative 
coefficients of .35 and .60. These coefficients are derived from several 
sources, including a comparison of the fiscal multipliers of the major 
Canadian econometric models given in Helliwell (1982). The coefficients are 
intended to represent the average of these models in terms of the effects 


302 



Figure 4 

The fixed fiscal rule and the unemployment gap 


Unemployment 

gap 



of fiscal policy for a given money supply under flexible exhange rates. It 
can be seen that the assumed multiplier is rather low: less than 1.0 even 
after three years. This formulation does not allow for the likelihood that 
the size of multipliers will vary with the extent of economic slack. 

The key results based on both methods are set out in Table 5 and 
Figure 4. The unambiguous conclusion is that the hypothetical fixed fiscal 
rule would have destabilized the economy to a very significant degree. 
Over the 1971-81 period, the unemployment rate would have ranged from a 
low of about 4.6 per cent in 1974 to a high of close to 9 per cent in 1978, 
rather than the actual range of 5.3 per cent to 8.4 per cent. The varia¬ 
bility of the unemployment gap, as measured by its standard deviation, 
would have increased by a third to a half. Similar but somewhat smaller 
effects were found for the output gap. Also, according to the Infor- 
metrica simulations, provincial and municipal policy was very slightly 
destabilizing: if non-federal governments had followed the fixed fiscal rule 
rather than their actual policies, then the standard deviations of UGAP 
and YGAP would have been reduced respectively by 7 per cent and 3 per 
cent. It can be seen too that the two methods yield rather similar results, 
the main difference being the somewhat slower rise in unemployment after 
1977 according to the Informetrica simulations. 

A further point to note is that while the fixed fiscal rule would have 
involved no change in overall fiscal stimulus over the course of the eleven 


303 







TABLE 5 

Effects of fixed fiscal rule and reduced automatic stabilizers on economic 
stability, 1971-81 




Actual 

Standard deviations, 

Fixed fiscal 
rule 

1971-81 

Reduced 

automatic 

stabilizers 

Combined 

case 

A. 

UGAP 






Method A 

.90 

1.36 

1.20 

1.81 


Method B 

.90 

1.21 

1.20 

1.61 

B. 

YGAP 






Method A 

2.03 

2.82 

2.64 

3.76 


Method B 

2.03 

2.59 

2.64 

3.45 


NOTE: Under Method A, column 2 is estimated according to the formulas 
set out in the text, while under Method B column 2 is based on the Infor- 
metrica model. Under both methods, column 3 is one-third larger than 
column 1, while column 4 is one-third larger than column 2. 


years as a whole, it might nevertheless have resulted in higher average 
inflation. This would have been the case if the lower unemployment rates 
in 1973-4 had added more to inflation than the higher unemployment rates 
of 1977-81 subtracted. In this sense federal fiscal policy has been anti- 
inflationary: for a given average level of stimulus, the counter-cyclical 
pattern actually followed was less inflationary than the alternative pattern 
implied by a fixed rule. The second and separate issue is whether the 
average level of stimulus was too much or too little, and that depends 
largely on the priority attached to inflation relative to unemployment as 
well as on questions of monetary-fiscal mix. 

Finally, we may consider the joint impact of the fixed fiscal rule in 
conjunction with thirties-style automatic stabilizers, i.e., a marginal tax- 
minus-transfer rate of .14 rather than .44. As is discussed in the ap¬ 
pendix, such a change may raise the multiplier by about one-third. Hence 
the unemployment gaps would also be inflated by one-third, as is shown in 
Figure 5. Then, if the fixed fiscal rule is imposed on top of the reduction 
in automatic stabilizers, the result is the Case B line of Figure 5 (based 
on the Informetrica results). The unemployment rate now peaks at over 11 
per cent in 1978 and reaches a hypothetical low of 2.9 per cent in 1974 - 


304 








Figure 5 

Effects of alternative fiscal rules on unemployment gap, 1971-81 


Unemployment 

gap 



hypothetical because inflationary pressures would probably prevent the 
economy from reaching such a low level of unemployment. As measured by 
the standard deviation of UGAP, the reduction in automatic stabilizers 
increases instability by a third to a half, the fixed fiscal rule by itself 
raises instability by a third, and the combined effect is to increase insta¬ 
bility by 80 to 100 per cent. 

The 1971-81 policy rule applied to 1982 

The conduct of federal fiscal policy may be likened to an athlete who 
performs brilliantly during practice sessions, but then stays home on the 
day of the race. During the relatively mild ups and downs of the 1970s, 
federal policy exerted a helpful stabilizing influence, but when it came to 
the big recession the stabilizing patterns of the past were abandoned - the 
small increase in the adjusted federal deficit in 1982 was mainly the result 
of a non-discretionary increase in real interest rates. It may be interest¬ 
ing to ask what would have happened in 1982 if the federal government 
had continued to pursue the rule implicit in the fiscal policy of earlier 
years. It is not suggested that the federal government necessarily should 
have done this; and also, because the depth of the recession of 1982 was 
largely unanticipated in 1981, the government would probably have missed 
the target even if it had tried to follow the 1971-81 rule. Nevertheless, 


305 












TABLE 6 

Implications of 1971-81 fiscal rule for key variables in 1982-4 


Impact on: 

1982 

1983 

1984 

Unemployment rate 

-.44 

-.74 

-.98 

Real GNP (per cent) 

+1.3 

+ 1.6 

+ 1.7 

A 

Adjusted budget balance (B" ) 




as per cent of GNP 

-2.1 

-2.0 

-1.9 

Actual budget balance 




as per cent of GNP 

-1.6 

-1.4 

-1.2 


NOTE: Based on fiscal rule B‘ = .37 - .47 YGAP, plus equations (6) and (7) 
of the text, plus assumption that stimulus is entirely in the form of tax 
cuts. See the appendix for further details. 

the results of this hypothetical exercise may be useful as a base case for 
discussion of present policy options. 

The results of this exercise are set out in Table 6. Under the old 
fiscal rule, the 1982 unemployment rate would have been about half a point 
lower and real GNP 1.3 per cent higher. These effects would have in¬ 
creased to a one-point reduction in the unemployment rate and 1.7 extra 
points on GNP by 1984, assuming no private sector recovery (otherwise 
part of the fiscal stimulus would be automatically removed). But these 
rather modest effects would have required an increase in the 1982 deficit 
of just under $6 billion after taking into account induced changes in taxes- 
minus-transfers. Because of the low multipliers and the long lags between 
changes in policy and changes in unemployment, it takes a large increase 
in the deficit to achieve even a very modest reduction in the unemployment 
rate. 

While the order of magnitude of these numbers is typical of Canadian 
econometric models, nevertheless several major points of qualification 
should be noted. First, the first-year output and employment effects 
would be at least double those shown in the table if the fiscal stimulus 
took the form of higher government spending rather than lower taxes. 
Second, while estimates such as those just given may be quite reliable 
during 'normal' years, 1982 was far from normal. I do not think (although 
I may be wmong) that any of the econometric models would have predicted 
anything like the unprecedented drop of 5 per cent in real GNP that was 
recorded in 1982. It seems likely that an explanation of 1982 will have to 


306 









include a major exogenous shock to consumer confidence or security, 
possibly as a result of the sharp initial increases in unemployment and the 
unprecedented level of interest rates. It may be, then, that had fiscal 
policy moved early to provide some support to consumer spending or 
personal income, the shock to confidence would have been less and hence 
the multiplier effects greater than indicated in the table. A third and 
related point is that a substantial easing of monetary policy rather than 
fiscal policy may have been a more efficient and effective way of limiting 
the depth of the 1982 recession. 

CONCLUSIONS AND POLICY CONSIDERATIONS 

A common thread running through this paper has been the proposition that 
fiscal policy matters. Perverse discretionary policy and weak automatic 
stabilizers made Canada's Great Depression of the thirties a great deal 
worse than it would have been if today's policy environment had been in 
place at that time. Conversely, a counter-cyclical federal fiscal policy in 
conjunction with the automatic stabilizers has made a major contribution to 
economic stability since 1970. 

A summary of the quantitative estimates of these effects is set out in 
Table 7. The Depression was an estimated two and a half times as bad as 
it would have been in today's fiscal environment, suggesting that that en¬ 
vironment provides considerable protection against a repeat of the 1930s. 
The results also suggest that the economy of 1971-81 would have been 
considerably less stable under neutral discretionary policy and 1930s-level 
automatic stabilizers. While there is widespread acceptance of the bene¬ 
ficial effects of automatic stabilizers, a key point to emphasize is that over 
the period 1971-81 discretionary federal fiscal policy contributed at least as 
much to stability as did the automatic stabilizers of all levels of govern¬ 

ment combined . 

Two other conclusions may be singled out. First, there is no struc¬ 
tural deficit at the federal level, and there is a sizable structural surplus 
for all levels of government combined. This conclusion, in sharp contrast 
to the American situation, is very similar to that of Bossons and Dungan 
(1983) and Rousseau (1983), despite considerable differences in method¬ 
ology between the two papers. A third point to emphasize is that accor¬ 
ding to conventional estimates, even a very large tax cut in 1982 (say $10 
billion) would have had only a modest braking effect on the recession. 


307 





TABLE 7 

Summary of effects of fiscal policy in the 1930s and 1970s 



Average output gap, 
1930-9 

Standard deviation of 
output or unemployment 
gap, 1971-81 


(Actual=100) 

(Actual=100) 

Actual discretionary policy 
and automatic stabilizers 

100 

100 

A. Neutral discretionary policy 70 

130-150 

B. Reverse stabilizers 

60 

130 

C. Combined case (A+B) 

40 

170-200 


a That is, actual discretionary policy + 1970s automatic stabilizers in 
the thirties and 1930s automatic stabilizers in the seventies. 

SOURCE: Tables 3 and 5. 


Despite considerable grounds for scepticism regarding such conventional 
estimates in the unconventional environment of today, this does suggest 
that straightforward tax cuts in 1983 may have to be very large to have 
much short-term effect on unemployment. 


Longer-term issues 


If one ignores the current recession for a moment and focuses on longer- 
term issues, I think that the findings of this paper provide considerable 
support for continuing with counter-cyclical fiscal policy. As noted above, 
this is quite distinct from the inflation issue, as it relates to the varia¬ 
bility of fiscal stimulus rather than its average level. The desired average 
level of stimulus depends partly on inflation-unemployment priorities and 
partly on the desired monetary-fiscal mix. 

Given the case for counter-cyclical fiscal policy, a key issue concerns 
rules versus discretion. Some readers may accept the evidence that dis¬ 
cretionary federal policy was stabilizing in the seventies, but question 
whether we should count on such good foresight and/or luck in the 
future. If this is the case, then it may be desirable to obtain the same 
effects of the implicit discretionary federal rule of 1971-81 by non-dis- 
cretionary means. This would imply an increase in the role of automatic 
stabilizers, and it would have to be a very large increase indeed if the 
object was to replicate the implicit rule of the seventies. One possibility 


308 









that would go a long way in this direction would be to make unemployment 
insurance premiums depend negatively on the unemployment rate, although 
recent changes in regulations in this area have moved the other way. 
Another possibility would be a variable period of eligibility for collecting 
unemployment insurance, where the period would be a regionally sensitive 
function of the unemployment rate. Yet another possibility is an invest¬ 
ment funds system like the one operated in Sweden, which has recently 
received a favourable evaluation by Taylor (1982). Another longer-run 
issue that may be noted in passing is the old question of whether the 
provinces should play a more active counter-cyclical role or at least refrain 
from the pro-cyclical policy that has often characterized the past. 

Present policy 

The remainder of the paper will outline a possible policy package for the 
present and then comment on likely objections to the proposals. What I 
would suggest is essentially a temporary increase in automatic stabilizers 
or a fiscal stimulus whose size is made contingent on the strength of the 
economy over the coming months. For example, the government could cut 
sales tax by an amount initially set at, say, $5 billion, but the amount of 
the tax cut could be adjusted automatically at quarterly intervals according 
to a formula based on economic indicators such as the unemployment rate 
or the level of industrial production. If there were a rapid recovery, part 
or all of the tax cut would automatically be removed, and vice versa. It 
would be possible to tie the stimulus to an indicator of inflation or wage 
settlements in order to allay fears that the fiscal stimulus would be infla¬ 
tionary. Of course, the principle of a fiscal stimulus contingent on the 
speed of recovery and/or inflation could be applied to measures other than 
sales tax cuts. Also, the principle could either be applied on a temporary 
basis, perhaps as a tax cut for one year but subject to quarterly revi¬ 
sions, or become a permanent addition to the stock of automatic stabilizers. 

As already noted, even a substantial tax reduction may not give much 
short-term improvement in employment. Thus the tax cut could be sup¬ 
plemented by explicitly temporary (or unemployment-contingent) increases 
in government spending on public works and other projects. Also, tax 
relief could be directed to those likely to have high spending propensities, 
e.g., an unemployment-contingent extension of the period of eligibility for 
unemployment insurance benefits. 


309 




Finally, let us briefly consider possible objections to proposals of this 


kind: 


The higher deficit may crowd out private spending through higher 
interest rates or an exchange rate appreciation. This depends in 
part on monetary policy, which, I would argue, should act to prevent 
such developments in the present context. In any case, despite the 
record government deficits in 1982, the current account of the bal¬ 
ance of payments moved sharply into surplus , implying an increase in 
national saving relative to financing requirements. Also, with the 
fiscal stimulus contingent on the speed of recovery, the danger of 
future crowding-out would be alleviated. 

The idea that government debt will eventually be monetized applies to 
the case of a structural deficit. The fiscal stimulus described above 
would not raise the structural deficit. 

Fiscal stimulus is sometimes associated with increased government 
intervention or size of government, but this need not be the case. 

There are those who argue against an increase in the deficit because 
a strong recovery is supposedly under way, and there are those who 
argue against a higher deficit because they do not expect the econ¬ 
omy to recover at all (and hence there is a structural deficit). 
There are even those who argue both points at the same time (e.g., 
Hugh Anderson of the Montreal Gazette ). The first point is taken 
care of by the indexing provision, while the second point may be 
countered on several different grounds. 

Finally, higher deficits may have a negative effect on business con¬ 
fidence . It seems, however, that several Canadian business associa¬ 
tions are now themselves recommending a 'moderate' fiscal stimulus 
and that an explicit indexing of tax cuts to the state of the economy 
and/or inflation would go a long way in dampening concerns. Finally, 
at this stage of the business cycle consumer confidence is probably 
more important than business confidence, and it seems difficult to 
believe that tax cuts directed at consumers would cause a reduction 
in their 'confidence.' 


310 










APPENDIX 


Several items in the text have been referred to the appendix, and the 
order of presentation will correspond to the order in which the points were 
raised in the text. 

Budget balances 1929-42 (Table 1) 


Unless otherwise indicated, all data were taken from Statistics Canada, 
National Income and Expenditure Accounts , Vol. I, 1926-74. The compon¬ 
ents of the budget are defined as follows: 

g - Government spending on goods and services in 1947 dollars (current 
expenditures and capital expenditures each deflated by corresponding 
implicit price deflator) as a percentage of potential GNP in 1947 dollars. 

T - ZGAP - Total taxes minus transfers excluding interest on public debt 
(nominal values deflated by consumer expenditure implicit price deflator, 
1947=100) as a percentage of potential GNP in 1947 dollars. The cyclical 
component consists of losses of government business enterprises plus the 
product of the output gap and cyclically sensitive taxes, defined as all 
direct and indirect taxes other than property tax. The non-cyclical com¬ 
ponent is the residual. 

(r+7i)d - Interest payments on national debt (deflated by consumer expen¬ 
diture price deflator, 1947=100) as a percentage of potential GNP in 1947 
dollars. The inflation premium component is defined as the inflation rate 
(as defined by the GNP deflator) multiplied by a measure of real govern¬ 
ment debt (nominal debt deflated by consumer expenditure price deflator). 
Debt figures, not all fully comparable, were taken from various issues of 
the Canada Yearbook. 

Alternative policies in the thirties (Table 3) 

To obtain the effect of alternative fiscal policies on the output gap, we 
need an estimate of the fiscal multiplier. In general, the multiplier (M) 
may be written: 


311 





M = [1 - c(l-m) (1-t) + x] 


where c, m, and t are the marginal propensities to consume, import, and 
tax, and x is all other factors influencing the multiplier (crowding-out 
effects, accelerator effects, etc.)* If we assume present-day values of M 
= 1.5, c = .85, m = .25, and t = .44, it follows that the present-day value 
of x is .02. If we now assume 1930s values of c = .9, m = .125, and t = 
.14, it follows that if x remains equal to .02 the multiplier becomes 2.9, 
which has been rounded up to 3. The effects on the actual budget bal¬ 
ance shown in Table 3 are given by .58 times the increase in the adjusted 
balance, where the figure of .58 [= 1 - (3x.l4)] allows for an induced rise 
in tax revenue. The actual budget balances as percentages of actual GNP 
are obtained by dividing the balances as percentages of potential GNP by 
(1 - YGAP). Finally the current account estimates are based on the 
assumption that m = .125. 

Budget balances, 1954-70 


These are set out in Table Al, which is based on the same definitions and 
methodology as Table 4. 

Output gap estimates 


The Department of Finance estimates of the output gap for 1978-82 were as 
follows (taken from 'Cyclical and Inflation Adjustment of Government 
Deficits', 11 January 1983, mimeograph): 


1978 

1.5' 

1979 

1.8 

1980 

4.4 

1981 

4.2 

1982 

10.6 


A key ingredient in these estimates is an assumption about the trend rate 
of growth of productivity, a point on which there is a very large amount 
of uncertainty. I was puzzled by these figures primarily because 1979 was 
a year of near-record high employment growth (4.0 per cent) as well as a 
substantial decline in the unemployment rate (from 8.4 per cent to 7.5 per 


312 





TABLE A1 

Components of government budget balances, as percentage of potential 
GNP, 1954-70 


All governments 


18 

t 

-rd 

-nd 

-zGAP 

B 

1954 

-17.8 

19.4 

-1.5 

-1.0 

-.4 

-1.1 

1955 

-17.9 

19.4 

-2.0 

-.4 

.7 

.0 

1956 

-18.4 

19.8 

-.4 

-2.0 

1.7 

1.0 

1957 

-17.9 

19.5 

-1.3 

-1.1 

.6 

.0 

1958 

-17.9 

18.0 

-1.6 

-.8 

-.6 

-3.1 

1959 

-17.5 

19.5 

-1.7 

-1.0 

-.8 

-1.6 

1960 

-17.2 

19.4 

-2.1 

-.7 

-1.3 

-1.8 

1961 

-19.0 

21.7 

-2.7 

-.2 

-1.9 

-1.9 

1962 

-19.3 

21.6 

-2.3 

-.7 

-.9 

-1.6 

1963 

-19.1 

21.5 

-2.1 

-1.0 

-.6 

-1.3 

1964 

-19.0 

22.2 

-1.9 

-1.2 

.0 

.2 

1965 

-19.7 

22.8 

-1.5 

-1.6 

.4 

.4 

1966 

-20.9 

23.7 

-1.1 

-2.0 

1.0 

.7 

1967 

-21.4 

24.3 

-1.5 

-1.7 

.3 

.2 

1968 

-21.8 

25.6 

-1.9 

-1.4 

.3 

.7 

1969 

-21.9 

27.5 

-1.7 

-1.8 

.5 

2.4 

1970 

-22.7 

27.8 

-2.5 

-1.3 

-.5 

.9 




Federal 

government 





18 

t 

-rd -7id 

-zGAP 

B 

1954 

-9.4 

11.4 

1.0 -.8 

-.4 

0 

1955 

-9.1 

10.8 

1.5 -.3 

.7 

.7 

1956 

-8.9 

10.9 

-.1 -1.6 

1.7 

2.0 

1957 

-8.3 

10.4 

-.8 -.8 

.6 

.9 

1958 

-8.1 

8.3 

1.1 -.6 

-.6 

-2.3 

1959 

-7.4 

9.1 

1.2 -.7 

-.5 

-.8 

1960 

-6.7 

9.1 

1.4 -.5 

-1.0 

-.5 

1961 

-7.0 

9.4 

1.8 -.1 

-1.7 

-1.0 

1962 

-6.7 

8.2 

1.5 -.5 

-.7 

-1.1 

1963 

-6.1 

8.1 

1.3 -.7 

-.6 

-.6 

1964 

-5.9 

8.5 

1.1 -.8 

0 

.6 

1965 

-5.8 

8.4 

-.9 -1.1 

.4 

.9 

1966 

-6.1 

7.6 

-.6 -1.3 

.8 

.3 

1967 

-6.1 

7.6 

-.8 -1.1 

.3 

-.2 

1968 

-6.1 

7.8 

1.1 -.9 

. 1 

0 

1969 

-6.0 

8.9 

-.9 -1.1 

.4 

1.3 

1970 

-5.8 

8.5 

1.3 -.9 

-.3 

.3 

NOTES 

: - B = -G + T ■ 

-rD - nD - 

ZGAP. Symbols 

are defined as follows 

(all 

as percentages of potential 

GNP): G = nominal spending 

on goods 

and 

services; T = non-cyclical 

component of 

taxes minus 

transfers 

excluding interest on 

public debt 

; (r+7l)D = interest payments 

on public 

debt; 

ZGAP = cyclical 

component 

of taxes minus 

transfers; B 

= nominal 


budget balance. 


313 













cent), while employment rose at approximately average rates in 1980-1 (2.8 
per cent and 2.6 per cent respectively) and the unemployment rate re¬ 
mained quite stable. Thus, in terms of the employment record or an 
Okun's Law relationship, one should certainly find a lower output gap in 
1979-81 than in 1978. The contrary results shown above result from 
Canada's exceptionally low (negative) productivity growth over this period 
in combination with the assumption that 'trend' productivity growth was 
not so low. However, it seems difficult to make the case that productivity 
growth in 1979-81 was low for cyclical reasons, since employment was 
rising rapidly over this period. 

Consequently, the output gap series used in the text has been based 
on a method that does not force one to make assumptions concerning trend 
productivity growth. Specifically, given the following Okun's Law equa¬ 
tion: 


UGAP = .35YGAP + .15YGAP_ 1 + .10YGAP_ 2 , 

together with UGAP numbers from Fortin and Phaneuf (1981) and Depart¬ 
ment of Finance YGAP numbers for 1969 and 1970 (-1.0 per cent and 1.6 
per cent respectively), it is possible to construct a YGAP series for 1971- 
82. The Department of Finance estimates of cyclically adjusted tax reve¬ 
nue were adjusted to conform with the new YGAP series, as given by: 


1971 

1.3% 

1975 

2.2 

1979 

.6 

1972 

-.5 

1976 

.8 

1980 

1.8 

1973 

-2.6 

1977 

3.1 

1981 

2.0 

1974 

-2.2 

1978 

3.8 




Note that the effect of this change in YGAP series is to reduce the esti¬ 
mates of structural surpluses in 1981-2. On the other hand, fiscal policy 
becomes pro-cyclical in 1980-1 if one uses the Department of Finance 
output gap series. 

Effect of reduced automatic stabilizers (Table 5) 


According to the multiplier formula given above (Item 2), if all of the 
present-day assumed coefficients are unchanged except t (which is reduced 
from .44 to .14), then the multiplier rises from 1.5 to 2.1, or by somewhat 


314 





more than the one-third increase assumed in the text. 


Regression results 

a 

These results are based on B as defined in equation (2), with FBA, and 

A A 

NFBA denoting respectively the federal B and non-federal B . 'Non-fed- 
eral' includes provincial and local governments as well as CPP/QPP. The 
results for 1971-81, with t-statistics in brackets, were as follows: 

FBA = .37 - . 47YGAP DW = 1.24 R 2 = .50 

(1.2) (3.3) 

FBA = .50 - 1.17UGAP DW = 1.09 R 2 = .63 

(1.8) (4.2) 

NFBA = 1.71 + .07YGAP DW = 1.21 R 2 = -.06 

(7.8) (.6) 

NFBA = 1.67 + . 22UGAP DW = 1.34 R 2 = .00 

(7.6) (1.0) 

The corresponding equations for 1954-70 were: 


FBA = .90 - .10YGAP 
(5.3) (1.5) 


DW = 1.86 R 2 = .07 


FBA = .95 - .30UGAP DW = 1.92 R 2 = .18 

(5.8) (2.1) 

NFBA = .03 + .03YGAP DW = .23 R 2 = -.06 

(.14) (.3) 


NFBA = .02 + . 26UGAP 
(.1) (1.4) 


DW = .28 R 2 = .06 


315 



TABLE A2 

Unemployment under a fixed fiscal rule 



Unemployment rate 



UGAP 


Actual 

Simulated 

A B 

Actual 

Simulated 

A B 

1971 

6.2 

5.9 

6.0 

.6 

.3 

.4 

1972 

6.2 

5.9 

6.0 

.1 

- .2 

- .1 

1973 

5.5 

4.9 

5.1 

- .9 

-1.5 

-1.3 

1974 

5.3 

4.3 

4.7 

-1.2 

-2.2 

-1.8 

1975 

6.9 

6.5 

6.5 

.3 

- . 1 

- .1 

1976 

7.1 

6.9 

6.8 

.5 

.3 

.2 

1977 

8.1 

8.3 

8.1 

1.4 

1.6 

1.4 

1978 

8.4 

9.1 

8.7 

1.8 

2.5 

2.1 

1979 

7.5 

8.1 

7.9 

.9 

1.5 

1.3 

1980 

7.5 

8.1 

8.0 

.9 

1.5 

1.4 

1981 

7.6 

7.8 

7.8 

1.0 

1.2 

1.2 

Standard 







deviation 

1.03 

1.53 

1.33 

.90 

1.36 

1.21 

NOTE: Simulation A is 

based on the 

formulas 

set out in 

the text, 

while 


simulation B is based on the Informetrica model. 
Fixed fiscal rule 


The actual and simulated values of U and UGAP are set out in Table A2. 

REFERENCES 

Bossons, J. and D.P. Dungan (1983) 'The government deficit: too high 
or too low?' Canadian Tax Journal 31, 1-29 
Brown, E.C. (1956) 'Fiscal policy in the "thirties": a reappraisal.' 

American Economic Review 46, 857-79 
Fortin, P. and L. Phaneuf (1981) 'Why is the unemployment rate so high 
in Canada?' Universite Laval, cahier 8115 
Helliwell, J.F. (1982) 'Recent evidence from macroeconomic models of the 
Canadian economy.' Paper presented at Conference on Economic 
Policies in the 1980s, Winnipeg, October 28-9 
Rousseau, H. (1983) 'The Dome syndrome: the debt overhanging Canadian 
government and business.' Canadian Public Policy 9, 37-52 


316 











Taylor, J.B. (1982) 'The Swedish investment funds system as a stabilization 
rule.' Brookings Papers on Economic Activity 1, 57-106 


Robert W. Baguley* 

The underlying and unifying theme of Professor McCallum's paper is that 
fiscal policy 'matters.' The paper is strongly Keynesian in tone, and its 
conclusions are most applicable to a world of low aggregate demand, no 
supply restrictions, and high unemployment. But these conclusions are 
not necessarily applicable to a world in which inflationary expectations 
have a role in determining real interest rates - a world in which behaviour 
has been altered by recent experience with high and rising inflation rates. 

I think Professor McCallum's failure to take the effect of this experience 
into account is a major weakness in his paper. 

I will confine my comments to the major analytical and policy issues 
that, in my view, arise in Professor McCallum's paper. 

The paper's first major conclusion is that from an analytical point of 
view, fiscal policy exacerbated the Depression of the 1930s. Discretionary 
policy was perverse, on balance, and the automatic stabilizers built into 
the economy were weak - relative at least to those present in today's 
economic environment. 

I would not quarrel with this finding. It is, after all, the tenet upon 
which Keynesian economic theory is based. Nevertheless, current economic 
thought and policy are too strongly wedded to the idea that the policies 
that should have been pursued half a century ago are still valid in today's 
economic environment. While the fiscal policies of the 1970s and early 
1980s might have alleviated the Depression of the 1930s, it is not clear to 
me that such policies have in fact alleviated the recent recession. Indeed, 
these policies may have deepened and prolonged the recession. 

A second key conclusion of the paper is that there is no structural 
deficit at the federal level and that for all governments combined there is 
a substantial full-employment surplus. This conclusion is loaded with 
implications, not all of them clearly stated. One very relevant and impor¬ 
tant implication is that there is substantial room right now for fiscal stimu¬ 
lus through higher discretionary expenditures and/or through tax cuts, at 

* Vice-President - Economics, the Royal Bank of Canada. 


317 



least at the provincial level and probably at the federal level as well. 

Indeed, the paper goes further than this. When one disentangles the 
details of the analysis, it seems that Professor McCallum is arguing that 
his estimates of the full-employment balance or surplus are conservative 
because he has biased some of his key assumptions so as to make his 
calculations understate government surpluses. For example, he argues 
that the reduction that occurred in structural surpluses in 1981 and 1982 
was partly due to the sharp riae in real interest rates payable on the 
public debt. If real interest rates fall back to historical levels, the paper 
argues, the current deficit will largely disappear. 

This argument, it seems to me, views real interest rates as some sort 
of exogenous variable that affects government fiscal requirements, but that 
is not in turn affected by them. In my view, however, the causality also 
runs in the opposite direction. The inflationary expectations generated by 
huge actual government deficits directly and indirectly push real interest 
rates upward, further raising actual deficits. The high real interest rates 
resulting from such expectations raise household savings rates and reduce 
investment expenditures, and thus hold back economic recovery. In other 
words, real interest rates may not fall at all - especially if government 
deficits rise further. The U.S. economy is in this vicious circle right 
now. 

On the other hand, the paper argues, if oil prices fall, this will 
reduce government revenues and raise the prospect of structural deficits 
in future. While Professor McCallum describes this structural dependency 
upon rising oil prices 'negligible,' he cannot know for sure that it is 
negligible. He does not know how sharply and how far oil prices may fall 
in the near-term future. 

It is not clear, therefore, that he has underestimated the structural 
surplus. He may well have overestimated the surplus. 

A third key conclusion of Professor McCallum's paper is that federal 
fiscal policy has been counter-cyclical on balance in the postwar period 
and especially since 1970. Professor McCallum's conclusions here support 
the discretionary fiscal policy of the past decade, and they appear to make 
a case for increasing the stabilizing role of fiscal policy in the future - 
but Professor McCallum would do so mainly by increasing the strength and 
responsiveness of the built-in stabilizers. The support for continuing with 
counter-cyclical fiscal policy while moving towards more automatic stabiliz¬ 
ation and relatively less discretionary stabilization is perhaps a recognition 


318 


that the lead times of discretionary policy responses to falling output and 
employment are longer than the lead times of automatic or built-in re¬ 
sponses. He also has a feeling that discretionary policy may be applied 
less 'accurately' in future. 

Professor McCallum's policy recommendations echo this shift in 
emphasis. For example, he suggests that the automatic stabilizers be made 
more potent in future, but that they also be tied to various indicators of 
economic recovery, such as the speed of recovery. This would counter 
the objections of those who believe that economic recovery is already 
underway. 

While this is an interesting variation on the theme that fiscal policy is 
the critical stabilizer (the paper virtually ignores monetary policy), 
Professor McCallum does not provide any support for a market response to 
recession. He is a strong advocate of fiscal stabilization, and a key 
feature of the paper is a greater emphasis on automatic than on discretion¬ 
ary fiscal measures. But a third option, where market forces prevail and 
the economy takes its lumps swiftly, is not analyzed at all. Perhaps that 
sort of economic environment would, in fact, be the best option for the 
future. Who knows for sure? What one can say is that the stabilization 
policies that emerged from the experience of the 1930s depended heavily 
upon money illusion. The role of inflationary expectations and the greater 
sophistication of market behaviour today have probably reduced the bene¬ 
ficial impact of these stabilization measures. 

In summary, Professor McCallum's analysis is, perhaps, at least as 
interesting for the questions it raises as it is for the questions it attempts 
to answer. It is a useful jumping-off point for a more in-depth analysis of 
our continuing preoccupation with economic stabilization measures - a pre¬ 
occupation that may be undermining the market economy and limiting our 
medium-term growth prospects. 


Discussion 


MODERATOR: I think while one or two of you are formulating questions 
we should give John McCallum at least one minute to respond to the man 
from Missouri. Do you want to do that, John? 


319 




JOHN MCCALLUM: Yes, please. First, I did not say, or I did not mean 
to say, that oil price increases would have a negligible effect on the 
deficit. Second, Robert’s remarks seem to ignore the point that the 
counter-cyclical policy does work both ways. It’s not just a matter of 
pumping up the economy during the bad times: it's also a matter of 
pumping it down during good times, and I am advocating something that 
would be balanced over the cycle as it appeared to be in the seventies. I 
think one has to look at the matter both ways and not focus in one 
direction. 

Third, my proposals for the current situation do not involve any 
change in the structure of deficit, assuming such a thing exists. They 
are explicitly temporary, and the indexing formula thrown in increases that 
temporariness. 

Fourth, I've nothing against the market system - in fact I'm for 
it - but I don't see why the counter-cyclical policy and the market system 
cannot work together as they did in the seventies. Clearly, the market 
system has not always performed very well, such as during the 1930s. My 
policies do not necessarily involve any increase in the role of government, 
because a counter-cyclical policy could be performed entirely through 
changes in tax transfers, with no changes in government spending on 
goods and services. 

Finally, I don't see how a counter-cyclical policy, whether it is dis¬ 
cretionary or not, rests in any way on money illusion. 

QUESTION: My first question is about the various hypothetical experi¬ 
ments that changed the world of the thirties, or changed the world of the 
seventies, to then ask what happens in that world. What kind of a model 
do we have lying behind those calculations? I didn't see an explicit model 
in the paper and I didn't hear any model mentioned in the discussion, 
other than the Informetrica model. My suspicion is that we have a model 
that is a variant of what I call the Keynesian-cross view of the world - 
that is, a model that generates output by fluctuations in an aggregate 
expenditure function, the slope of which depends on things like the pro¬ 
pensity to consume, the propensity to pay taxes and pay benefits, and so 
on. It's a model in which the monetary sector is either very primitive or 
totally suppressed in the sense that the stock of money is a completely 
accommodative variable. I wonder if we could be told whether that is the 
case. Or am I misinterpreting the monetary aspects of these experiments? 


320 


This raises a related question. We do know that we did not have 
sensible monetary policy in the 1930s and that it wasn't all that hot in the 
1970s either. Could we not with sensible monetary policy do even better 
these things we can allegedly do with fiscal policy? And isn't it indeed 
the case that we are rather overdrawing what we can gain from fiscal 
policy by analyzing it in an environment that is simply too naive in its 
treatment of the monetary sector? 

My final question is very brief. I was struck by the first two 
figures in the paper. The first figure showed the output gap in the 
United States and Canada during the Depression. The second figure 
showed fiscal policy in the two countries. The output gaps were identical; 
fiscal policies were very different. As a non-fiscalist sort of monetarist, I 
took a lesson from that, namely that fiscal policy perhaps is not all that 
important. How do you interpret the equal performance of the two econo¬ 
mies in terms of real output gap and the very different fiscal policy per¬ 
formance? 

JOHN MCCALLUM: The rule of thumb method that I used was based on 
the average of multipliers of Canadian econometric models as summarized by 
John Helliwell, all of which were based on the assumption of a flexible 

exchange rate and a fixed money supply. So they do incorporate the 

monetary sector, although to varying degrees of completeness. My results 
reflect that average. 

Aggregate supply also responds in these models, although not per¬ 
haps as much as you might like. The multipliers are small, on the as¬ 

sumption that the stabilization is performed through tax or transfer 
changes, not through government spending changes. The first year 
multiplier is on the order of .5 and over three years still less than one. 
So they are small multipliers, as you would expect under a flexible ex¬ 
change rate regime. But the shifts have been large, and so the effects, 
while not dramatic, have been significant. That the instability is reduced 
by approximately one-third is not dramatic, but it is significant. 

Regarding the role of monetary policy, as I've just said, the working 
assumption is that the money supply is held fixed, and so these experi¬ 
ments are for fixed money supply. 

I would agree that some of the stabilization should have been done by 
the monetary people, but the problem is that they haven't been very good 
at it. Look at 1982, when real GNP fell by close to 5 per cent. I would 


321 


argue that a lot of that fall, especially in view of the difference between 
what happened in Canada and in the United States, was the result of a 
very substantial tightening of monetary policy in 1981, which led to inter¬ 
est rates of over 20 per cent. It would have been asking far too much of 
fiscal policy to try to moderate that drop very much, given the stance of 
monetary policy. I would certainly agree that we should try to have a 
good monetary policy in addition to good fiscal policy. 

The final point was about the 1930s. The divergence between 
Canadian and U.S. deficits did not occur until after about '32 or '33, and 
so in fact the questioner's point supports my case rather nicely. If, in 
the thirties, you had held the automatic stabilizers constant and done no 
more than make federal fiscal policy in Canada the same as federal fiscal 
policy in the United States, then the Canadian economy would have re¬ 
covered faster and followed the U.S. output gap even more closely than it 
did. My second experiment introduced automatic stabilizers to the thirties 
equal to those of the eighties. Clearly the two countries' gaps in that 
case are not comparable, unless those automatic stabilizers are introduced 
in the United States as well in the 1930s. This would have meant a 
further favourable effect on Canada, because with American output higher 
as a result of that country's automatic stabilizers our output would have 
remained higher whether we had had automatic stabilizers or not. 


322 


Fiscal discipline and rules for controlling 
the deficit: some unpleasant 
Keynesian arithmetic 

Neil Bruce and Douglas D. Purvis* 


Commitment is no less important in affairs of state than in other human 
affairs. For our purposes, a commitment is a pledge to pursue a particular 
course even when circumstances change and alter the costs involved. The 
abiding enemy of commitment is opportunism, often proclaimed as prag¬ 
matism. Under any guise, opportunism is the taking of actions strictly on 
the basis of immediate costs and benefits, regardless of existing arrange¬ 
ments and sometimes regardless of future consequences. 

We want to argue the need for commitment in the conduct of fiscal 
policy - commitment in the form of fiscal policy rules . These are promises 
to pursue certain courses of action, particularly with respect to the deficit, 
even when it is politically opportune to renege. We focus our remarks on 
the design and desirability of rules rather than on their political enactment. 
Rules must somehow be entrenched through the political process, but we 
will leave this question for further discussion. 

After making the case for fiscal policy rules, we discuss some aspects 
of the design of such rules. In the final section we relate fiscal rules to 
monetary policy and, especially, disinflation. 

THE CASE FOR FISCAL RULES 

To develop the case for rules, we first consider the capacity of fiscal 
policy; that is, what can fiscal policy do, and what should it do? We go 
on to argue the need for fiscal prudence, and then discuss some political 
or ideological aspects of fiscal rules. 


* Professors, Department of Economics, Queen's University. 


323 



The impact of fiscal policy 


We believe the evidence is persuasive that fiscal policy matters for the 
short-run determination of income and prices. This is true >even if the 
policy is anticipated and even with due consideration for the openness of 
the Canadian economy. Nevertheless, the effects of fiscal policy are likely 
to be far less predictable and mechanical than the conventional macro- 
economic methodology suggests. This methodology presumes a controllable 
economy operating independently from, but affected by, the fiscal policies 
followed. In policy analysis use is made of fiscal policy multipliers, which 
are correlations between fiscal policy changes and private expenditure 
derived from historical time series data. These multipliers are used to 
determine an appropriate discretionary policy given a particular GNP 
target. 

Our skepticism about the usefulness of such an approach is based on 
the fact that the 'all other things held constant' experiments supposedly 
extracted from the data are ill-defined. This is the well-known 'Lucas 
Policy Critique' (see Lucas 1975). The experiments would be well-defined 
if the economy were populated by myopic agents who spent a predictable 
fraction of their disposable income. But we believe that the effect of 
fiscal policy on private spending cannot be derived, even to a first ap¬ 
proximation, without considering the effects of the policy on private ex¬ 
pectations about future taxes, future expenditures, and future monetary 
policy. 

To illustrate, consider the polar case of a 'rational economy' populated 
by rational maximizers with perfect foresight and access to a well-func¬ 
tioning capital market. In such an economy, private expenditure would be 
based not on current disposable income but on long-run disposable wealth. 
As a result, income tax and transfer policies - and the deficits they 
caused - would be irrelevant for stabilization purposes. Present and 
future taxes would be equivalent, and deficits would simply be promises of 
future taxes. 

There are a number of important implications for fiscal policy in this 
rational economy. First of all, the distinction between government expend¬ 
itures on capital goods and government expenditures on current goods, 
often a key distinction in debates about policy, would be irrelevant in 
determining whether any given deficit were appropriate. If the govern¬ 
ment invested in social capital, such as roads, it would not matter whether 


324 



private sector agents were called upon to pay taxes currently, and bor¬ 
rowed subsequently to smooth their lifetime consumption stream, or 
whether the government borrowed directly. The second implication 
concerns built-in stabilization. Because the cyclical component of the 
government deficit consists largely of induced changes in taxes and trans¬ 
fers, balancing the budget annually by means of tax changes would not be 
destabilizing. However, fiscal policy could still matter in this economy 
because changes in government spending on goods and services, particu¬ 
larly if they are known to be temporary changes, will alter aggregate 
demand. The requirement that fiscal policy take the form of expenditure 
changes would pose problems for stabilization policy, since for good 
reasons fiscal policy is generally implemented through tax changes. 

Most economists do not believe this 'Ricardian' economy to be an 
accurate description of the world. For reasons mentioned by Professor 
Modigliani in this volume and for other reasons, tax policies and deficits 
do influence the economy. Regardless of whether the private decision¬ 
makers have foresight, they may be unable to carry out their expenditure 
plans if capital markets are imperfect. That is, firms and households may 
find expenditure plans constrained by cash-flow considerations. In such a 
situation a government deficit may be seen as financial intermediation 
where the government borrows, at better terms, on behalf of the private 
sector. Balancing the budget annually would mean that the government 
would no longer act as a financial intermediary and provide a flow of funds 
to individuals who would otherwise be unable to fulfil their expenditure 
plans. As a result, private sector expenditure would rise in booms and 
fall in slumps, and thus exacerbate the cyclical swings in income. 1 

While it is true that the world is unlikely to be populated by rational 
maximizing decision-makers operating in perfect capital markets, it is also 
unlikely that it is populated by myopic decision-makers whose behaviour 
can be described by simple Keynesian consumption functions. The actual 
impact of fiscal policy depends on whose incomes are being affected and on 
how the policy alters the liquidity constraints that are binding in the 
economy. It also depends on the effect of the policy on the expectations 
of individuals regarding their lifetime disposable wealth or permanent 
income. For this reason, the mechanical link between tax policy and 
private expenditures that is presumed in the econometric models is unlikely 
to prevail in the real world. That is to say, the link between tax policy 
and private expenditures is likely to be more problematic than the histori- 


325 


cal correlations indicate, since these correlations are average correlations 
and do not necessarily predict how the private sector will respond under 
different circumstances. Furthermore, in the enactment of any current 
discretionary fiscal policy, a key consideration should be its effects on 
future behaviour. 

The need for fiscal prudence 


We use the term fiscal prudence to mean that, in some sense and in some 
time frame, the government must balance its budget. Of course, growth 
of real output allows for a non-zero 'steady-state' deficit. Private debt 
grows with real output, so why not the public debt as well? The budget 
need never be balanced literally; rather, the government can run a deficit 
that is consistent with letting its real debt grow at the trend rate of 
growth in real GNP. Henceforth, when we refer to 'balancing the budget' 
we mean this to include this growth term, a term that could amount to up 
to $3 billion annually in Canada at the present time. 

In the Ricardian economy discussed above, where people discount 
expected future income to estimate their lifetime wealth, it was assumed 
that the fiscal authorities do follow a policy of fiscal prudence. That is, 
the present value of future taxes is equal to the present value of future 
government expenditures by assumption. In practice, however, there is a 
need to ensure that such a fiscal balance is struck. Why? Unlike private 
borrowers, who must convince lenders of the responsibility of their 
intended borrowing with financial ratios, an investment prospectus, and 
the like, the federal government (at least in a sound economy) need never 
convince its creditors that, over some appropriate time period, its budget 
is in balance. The creditors of the public sector will not bear the cost of 
profligate public borrowing, except as citizens like everybody else, so as 
lenders they will not care whether government borrowing is responsible. 
They will lend to the government because of its power to tax and, if need 
be, to create money or to borrow still further in order to pay them 
interest. 

It is for this very reason that deficits are useful as a counter-cyclical 
device: they support private expenditure that would otherwise fall in 

recessions, when lending to private agents falls due to greater uncertainty 
and reduced cash flows. For the same reason, however, creditors would 
constrain government borrowing only when the economy was on the verge 


326 



of ruin - and that, of course, would be too late. Therefore, there is a 
need for non-market - that is political - constraints on public borrowing. 
The question is: what form should these constraints take? 

Before addressing that question, we should like to draw one implica¬ 
tion regarding provincial versus federal deficits. It is well known that the 
discretionary policy of provinces is often pro-cyclical. We do not take this 
to mean that provinces are miserly anti-Keynesians and unaware of the 
merits of counter-cyclical policy. Rather, the provinces find themselves 
very much in the same state as private borrowers; that is, their borrowing 
is constrained by the market. The reason is that provinces, unlike the 
federal government, cannot resort to the printing press to pay interest on 
their debt. Nor can provinces easily raise taxes in order to service a 
large debt. If they did, they would drive out industry and population to 
provinces with lower tax rates. Thus, in prescribing a rule for federal 
fiscal policy, one must take provincial deficit policies as more or less 
given, just as one takes private borrowing as given. 

Discretion and fiscal prudence 

Advocates of discretionary fiscal policies commonly couch their arguments 
in terms of 'balancing the budget over the cycle,' with deficits in reces¬ 
sions offset by surpluses in expansions. These deficits and surpluses 
would reflect both the working of the automatic stabilizers and the use of 
discretionary policies. We doubt that fiscal prudence and fiscal discretion 
can be combined in this way. 

First, the cycle is not a sine wave with predictable amplitude and 
period, but a complex and uncertain series of fluctuations involving cycles 
within cycles. Agreeing upon what constitutes a cycle for the purposes of 
balancing the budget is almost impossible, especially ex post . The fiscal 
authorities could almost always find good reasons for not balancing the 
budget at any particular time. 

Second, it is difficult enough for policymakers to carry out discretion¬ 
ary fiscal policy even without the constraint that the budget must be 
balanced over some cyclical period. The constraint would pose a horren¬ 
dous dynamic optimization problem for the policymaker. In deciding upon 
any action today, the policymaker would have to worry about the implica¬ 
tions for future deficits and, possibly, the implications of not covering 
past deficits. 


327 






Third, there is an ever-present temptation to forgo discretionary 
surpluses during good times and to let bygones (past deficits) be 
bygones. This not only reflects political reality; at a more academic level, 
it also reflects the time-inconsistency of optimizing policies (see Kydland 
and Prescott 1977). 

In addition, there are other, more pragmatic arguments against the 
use of discretionary fiscal policy. For one thing, discretionary policy 
results in recognition and decision lags that could be destabilizing. More¬ 
over, discretionary tax changes invite divisive distributional wrangling. 

For these reasons, we believe that it would be difficult, if not 
impossible, to combine discretionary fiscal policy and fiscal prudence. The 
notion of discretion with budget balance over the cycle provides a useful 
conceptual framework in which to discuss the issues, but is not of very 
much help in designing an operational fiscal policy. 

Political aspects of fiscal rules 


Fiscal rules are commonly advocated and interpreted in terms of 'con¬ 
straining the political system'. A conventional view is that discretionary 
policy will be biased towards expansion, with the cumulative effects of 
increasing the size of the government sector and, through induced 
pressures on the central bank to monetize the debt, accelerating inflation. 
Indeed, some advocates of rules concede that discretionary policy could 
improve macroeconomic performance in the short run but argue that the 
long-run costs dictate against its use. 

We argue that in fact there is reason to believe that fiscal discretion 
has harmful effects, and that adopting and following fiscal rules can lead 
to an improvement in macroeconomic performance. For example, consider 
the pressures that arise from time to time for the government to balance 
the budget. Typically, these pressures are strongest when budget 
imbalances are largest. Those imbalances often reflect the workings of the 
automatic stabilizers, and any discretionary policies introduced to reduce 
the budgetary imbalance will destabilize national income. Policies of 
precisely this sort were adopted in the 1930s, exacerbating and prolonging 
the Great Depression. Many people advocated the same policies in the fall 
of 1982 and the winter of 1983, when the federal deficit reached massive 
proportions for cyclical reasons; indeed, policies of this type were adopted 
at that time by several of the provinces. (Recall our earlier discussion of 


328 




why it is understandable that the provinces should behave this way.) 

The existence of a credible fiscal policy rule would not preclude 
entirely this possibility of perverse discretionary fiscal policy - rules can 
be amended or broken - but it might reduce the chance. Most of the 
pressure to reduce the deficit reflects fear that things are 'out of control' 
and that the government cannot. be trusted to allow a surplus to develop 
once the recovery occurs. Commitment to a fiscal rule that would maintain 
fiscal prudence over some clearly specified time period could quell at least 
some of those fears and reduce the pressure to pursue destabilizing dis¬ 
cretionary policy. Thus, rules may help avoid undesirable fiscal contrac¬ 
tion as well as undesirable fiscal expansion. 

THE DESIGN OF FISCAL RULES 

Fiscal rules may be either rigid or contingent. Rigid rules are easier to 
understand and enforce, but they are harder to defend against the seduc¬ 
tive appeal of discretionary policy. In this section we outline an approach 
to formulating rules and then critically evaluate some currently fashionable 
fiscal rules. 

A framework for formulating fiscal rules 


The basic notion of a fiscal rule is to have in place an established tax and 
expenditure structure that 'balances the budget' over the medium term 
(three to ten years) yet generates counter-cyclical fiscal behaviour in the 
short run. This structure may be augmented with certain policy options, 
contingent on circumstances, that add further stimulus or contraction as 
needed but are 'automatically' removed as the economy returns to a 'nor¬ 
mal' condition; e.g., programs with cyclical 'sunset' clauses. 

Clearly the problem is to design a rule that maintains fiscal prudence 
over the medium term but is counter-cyclical in response to short-run 
fluctuations. One framework for formulating such rules is that of Phillips 
(1954), who first identified the concepts of proportional, integral, and 
derivative automatic stabilizers. 

A proportional stabilizer results in a fiscal stimulus (contraction) that 
is proportional to the output gap - defined here as the difference between 
actual output and the output that is achieved on average. Such a stabil¬ 
izer exists in the present fiscal structure because aggregate taxes and 


329 




transfers are positively related to the level of aggregate output. (It is 
this that most people have in mind when they talk of - automatic stabil¬ 
izers .) The main shortcoming of proportional stabilizers is that they can 
only mitigate, not eliminate, fluctuations in output because an output gap 
is needed before a proportional stabilizer comes into play. 

An integral stabilizer is a positive function (for simplicity, a pro¬ 
portion) of the accumulated output gap over a half-cycle. With this stabil¬ 
izer, the stimulus (contraction) becomes greater the longer the economy 
remains away from its normal or average state. The case for an integral 
stabilizer for fiscal policy rests on the idea the economy will not automatic¬ 
ally return to a normal state but gets 'stuck' in the recessionary (or less 
likely, the expansionary) phase of the cycle. The problem with the integ¬ 
ral stabilizer is that its maximum impact occurs at the end of a half-cycle, 
at which point it may be extraordinarily pro-cyclical. An example of an 
existing 'automatic' integral stabilizer is the transfer resulting from the 
real debt service. During a recession (expansion) government debt rises 
(falls) as long as the output gap persists, assuming that the budget is in 
balance when output is at its normal level. Thus, the real transfer 
resulting from the debt service is largest (smallest) at the end of the 
recession (expansion). 

A 'derivative' stabilizer would be positively related to the difference 
between the trend growth rate and the actual growth rate. This stimulus 
would come into play if the economy grew more slowly than trend, regard¬ 
less of whether the output gap were positive or negative; in essence, it is 
a negative 'accelerator' effect. Phillips argued that there is a strong need 
for such derivative stabilizers. 

The present system relies heavily on automatic proportional stabilizers 
augmented by discretionary policy changes (although, as noted, the debt 
service transfers act as an integral stabilizer because of the dynamics of 
debt accumulation over the cycle). If one wishes to adopt rules in order 
to limit discretionary power for the purpose of fiscal prudence, it may be 
desirable that the fiscal rules augment the existing 'automatic' proportional 
stabilizers with stabilizers of the derivative and integral types. 

Rules that directly constrain the deficit 


A simple - some would say simplistic - rule is the annually balanced bud¬ 
get. It is very easy to criticize this rule, even to ridicule it. Why should 


330 




the budget be balanced over a year rather than over a longer or shorter 
period? 

In response to the guestion 'Why balance over a year?', a balanced- 
budget advocate would answer 'Why not?' There is nothing really sacred 
about a year except that it is a fixed and well-defined period of time. 
Balancing the budget over a year would ensure fiscal prudence. However, 
the fact that the year is a shorter period than the average business cycle 
implies that balancing the budget over a year will be destabilizing, for 
reasons discussed above. 

As mentioned, the problem in designing rules is that of trying to 
reconcile fiscal prudence in the form of a balanced budget over some 
time-frame with the requirement that fiscal policy be at least not pro¬ 
cyclical and preferably counter-cyclical. The answer to this problem seems 
to lie in requiring 'balanced' budgets when the economy is operating 
'normally' but allowing annual deficits or surpluses when it is not. 2 This 
approach differs from discretionary policy in that the rule both prescribes 
and proscribes the range of actions that can be undertaken by the author¬ 
ities. 

Perhaps the simplest rule that is not blatantly pro-cyclical is the 
requirement that the budget be balanced (excluding the growth factor) if 
output is on the trend growth path. Implementing this rule requires the 
calculation of the deficit adjusted for deviations of actual output from its 
trend value, which, in one form or another, is a common practice. A 
major difficulty with a rule that balances the budget subject to a cyclical 
adjustment lies in reaching an agreement on the appropriate cyclical adjust¬ 
ment to make. Specifically, unreasonable estimates of what the economy is 
capable of accomplishing over the medium term will leave the budget in 
deficit for an extended period of time. For example, the cyclical adjust¬ 
ment made by the Department of Finance (1983) has been in deficit for 
seven out of the last eight years and is not expected to be in surplus in 
the foreseeable future. This cyclical adjustment is based on a 'normal' 
unemployment rate of approximately 6.5 per cent. Few, if any, economists 
anticipate a period with unemployment rates below 6.5 per cent long 
enough to offset the 'cyclical' deficits of the past eight years and those 
anticipated over the next few years. In other words, any rule that 
balances the budget at 6.5 per cent does not balance the budget in accor¬ 
dance with any reasonable expectation of the economy's performance over 
the medium term. For this reason, we believe the cyclical adjustment 


331 


should be based on conservative and realistic forecasts of the medium-term 
output performance rather than on potential output estimates based on 
labour market and capacity utilization studies. 

Adjusting the budget balance for deviations in output over the cycle 
is commonplace, but the balance is also affected by other economic 
variables that may vary over time, including the stock of government 
debt, the real interest rate, and the inflation rate. We consider each in 
turn. 

The proportional stabilizer of the tax-transfer system implies changes 
in the stock of government debt, and hence the transfers to the private 
sector resulting from the debt service. These 'debt dynamics' are often 
ignored in discussing the merits of automatic stabilization. As mentioned, 
they act as an integral 'stabilizer', with the level of stimulus depending on 
the accumulated output gap. As such they may be destabilizing. In the 
appendix, we have added these debt dynamics to a simple cyclical model 
that assumes real interest rates remain constant. We find that, at least in 
the case considered, the debt dynamics slightly reduce but do not reverse 
the automatic stabilization provided by the tax-transfer system. In any 
case, the fiscal rule could require that taxes be adjusted to cover changes 
in the real debt service transfers due to cyclical fluctuations in the size of 
the government debt. Whether this is desirable depends on whether there 
is a need for integral stabilizers in the event that the economy becomes 
'stuck' above or below normal output levels. 3 

Fluctuations in the real interest rate also influence the budget 
balance. If such fluctuations are not offset by tax changes, they may be 
either stabilizing or destabilizing depending on whether real interest rates 
vary inversely or directly with output. Fiscal rules could be designed to 
accommodate budget-balance changes resulting from the former and to tax 
finance (and thereby sterilize) the latter. 

Recently, a strong case has been made for adjusting the government 
budget balance for the effects of inflation. The argument is fairly 
straightforward. Inflation reduces the real value of the outstanding 
government debt. Therefore an adjustment equal to the inflation rate 
times the net stock of government fixed dollar liabilities should be added 
to the budget surplus and subtracted from private saving. Adjustments of 
this type have been carried out by other authors in this volume, ourselves 
(1983), and the Department of Finance (1983). 

There seems to us to be a distinct danger in adjusting the govern- 


332 


ment budget balance using ex post inflation rates. This type of inflation 
adjustment implies that it is desirable to create government liabilities that 
can be absorbed only if the existing inflation rate continues. If this 
inflation rate differs from the target inflation rate, this approach would be 
inappropriate. Other adjustments aside, it seems clear that a necessary 
condition for lowering the inflation rate to the target level is to ensure 
that nominal government liabilities grow no faster than the target inflation 
rate. In particular, it is quite unrealistic to think that governments can 
increase their nominal liabilities at a rapid rate over time while the mone¬ 
tary authorities restrict the growth in the money supply by some form of 
the so-called k per cent rule advocated by Friedman. This is, of course, 
the thrust of the argument made by Sargent and Wallace (1981). 

Our views on the inflation adjustment of the budget balance can be 
summarized as follows. The now conventional practice of adjusting the 
balance using ex post inflation rates is fine for descriptive purposes. For 
prescriptive purposes (i.e., the design of a fiscal rule), it is appropriate 
to adjust only for the target inflation rate in determining the permissible 
deficit. 4 Of course, under a gradualist approach to disinflation, the 
medium-term target inflation rate may exceed the long-run steady-state 
inflation rate just as the medium-term unemployment rate may exceed the 
long-term 'natural' rate. 

Targeting on the national debt 


Recently, some observers have turned their attention to the behaviour of 
the national debt, usually scaled by GNP, in their search for something to 

pin a fiscal policy on. The main difficulty with the deficit, even when 

appropriately adjusted, is that it provides a static answer to a static 

question - Is there a structural deficit? - and ignores the accumulation of 
debt that occurs over time. The absence of a structural deficit in a 

recession implies that if the economy were to achieve capacity output 
at that time , there would be no deficit. But since recovery takes time - 
perhaps quite a long time - the cyclical deficits accumulate and the debt to 
GNP ratio rises. Numerous commentators have pointed out that the federal 
net-debt to GNE ratio has been rising steadily from a post-First-World-War 
low of approximately .05 in 1974. The Department of Finance (1983) 
projects this ratio to level off at .34 by 1986. 

There are a number of caveats to keep in mind when focusing on the 


333 









debt to GNP ratio. First, although the ratio automatically adjusts for 
output growth, it does so using actual rather than trend rates. Thus, 
the cyclical movement in the ratio is exaggerated, since the numerator 
(debt) and the denominator (GNP) move in opposite directions over the 
cycle. Second, the cyclical phase of the debt-GNP ratio does not corre¬ 
spond to the phase of the business cycle. The numerator (debt) reaches 
a peak or trough at the mid-point of the output cycle. This circumstance 
may give rise to misplaced concern about the high debt to GNP ratios 
expected towards the end of the current recovery. In fact, this is the 
time when the debt to GNP ratio is largest; it does not imply that the 
cyclical average debt to GNP ratio is rising over time. 

The third caveat concerns the implicit inflation adjustment one makes 
when targeting on debt to GNP ratios. Since nominal debt is divided by 
nominal GNP, the implicit adjustment makes use of the ex post GNP de¬ 
flator. We have already pointed out the dangers of making such an infla¬ 
tion adjustment to measured deficits. 

In conclusion, the danger of a pro-cyclical rule that validates existing 
inflation rates is as large, if not larger, when the focus is on debt to GNP 
ratios as it is when the focus is on balanced budgets. Nevertheless, the 
debt dynamics arising from the accumulated deficits over the cycle may be 
destabilizing, and, to the extent that an appropriate rule can be designed, 
targeting on the debt-GNP ratio could play a useful role in ensuring fiscal 
prudence. 

FISCAL RULES AND MONETARY DISINFLATION: LESSONS FOR 
CURRENT POLICY 

Consider the typical policy problem perceived by most Western economies in 
the last decade: to reduce the underlying inflation rate. There is wide¬ 
spread agreement among economists that a reduction in the rate of expan¬ 
sion of the money supply is necessary to ultimately achieve this goal. 
However, controversy remains about what strategies best ease the trans¬ 
ition from high to low inflation. Central to this debate is the question 
whether monetary growth should be reduced gradually or sharply. We 
believe that the emphasis on this question has crowded out the equally 
important question of the role for fiscal policy in the disinflation process. 

A necessary condition for a lower inflation rate is that all nominal 
liquid government liabilities grow no faster than the real growth rate plus 


334 



the target inflation rate. It is unrealistic to think that the government 
can continue to inject its overall liabilities into the economy at (say) 15 
per cent per annum over the foreseeable future and yet require that 
monetary authorities increase the money supply by (say) only 5 per cent 
per year. 

The considerations raised earlier in this paper suggest that these 
issues are of concern not only in terms of the long-run consistency of a 
program of reducing the rate of monetary expansion not accompanied by 
fiscal contraction, but also in terms of the design of a short-term policy to 
facilitate the transition. 

Consider a typical disinflation policy initiated by monetary restraint. 
This policy drives up interest rates, curbing expenditure and generating a 
recession. As the recession feeds into prices, it is argued, inflationary 
expectations will adjust and full employment will eventually be reattained at 
a lower inflation rate. During the adjustment, the measured government 
budget balance moves into deficit due to the recession. Further, the 
temporary rise in real interest rates leads to a worsening of the inflation- 
adjusted deficit. 

These deficits have recently been the source of considerable contro¬ 
versy in many countries that have embarked on policies of monetary dis¬ 
inflation, including Canada. Our own initial reaction was to accept the 
traditional adjustments that 'explained the deficit away' and, hence, to 
reject the concerns about the deficit being expressed, primarily but not 
exclusively, by analysts in the capital markets. While we still feel that 
many of these concerns were overstated, further consideration has led us 
to be less sanguine about the deficit. The cyclical adjustment is, as we 
have argued above, too static a concept. The inflation adjustment is 
difficult to do and even harder to explain, and it is often used in a way 
that confuses the roles of the actual and expected inflation rates. 

The cumulative effect of the large projected deficits, even corrected 
for target inflation, would be a dramatic increase in the debt to GNP ratio 
over the medium term. The effect of this increase on interest rates 
remains an unsettled question, but there is no argument to suggest that 
the rising debt to GNP ratio lowered interest rates. Its effects may have 
been small, but it likely contributed something to the troubling persistence 
of high real interest rates that seemed to inhibit recovery. Both through 
this and other channels, the rising debt to GNP ratio seemed to unsettle 
expectations about the credibility of the policies being pursued, and to 


335 



raise the possibility of a return to a high-inflation world. 

For these reasons, we feel that fiscal contraction is an integral part 
of a policy mix designed to reduce the rate of inflation in the economy. 
Not only is it required in the long run to limit the rate at which govern¬ 
ment liabilities are being injected into the system, it is also beneficial to 
the short-run goals of moderating expectations and easing the transition 
period. (The view that a tighter fiscal policy, accompanied by a less 
stringent monetary policy, would have eased the transition and aided a 
non-inflationary recovery has also been put forward by many American 
economists, including James Tobin.) 

In late 1982 we started looking at the deficit from the point of view of 
determining whether it was consistent with fiscal prudence and discipline. 
Our first approach was to ask whether there existed a 'structural deficit' - 
i.e., one that existed after adjusting for inflation and the cycle. 5 
However, the considerations discussed above soon led us to elaborate on 
that approach. Our view was that because of the projected large deficits 
into the medium term, the deficit did in fact impose a constraint on dis¬ 
cretionary fiscal expansion. 

In our 1983 paper, we calculated a budget deficit using target 
inflation rates on the order of 4 per cent, an unemployment rate (con¬ 
sistent with disinflation) of 9 per cent, normal real interest rates on the 
order of 4 per cent, and trend growth on the order of 3 per cent. We 
found that a 1983 calendar-year deficit, consolidated for all levels of 
government, in the mid-twenty billions seemed perfectly consistent with a 
rule that 'balances' over the anticipated future. For this reason, we 
advocated mild and temporary expansion on the order of $3 billion. 

Given the GNP gap of approximately $40 billion, this would be a very 
small stimulus. The prospect of a strong private-sector recovery in the 
U.S. and Canada, with the concomitant fear that fiscal stimulus operating 
with a lag would be pro-cyclical rather than counter-cyclical, strongly 
mitigated but did not eliminate the need for stimulus. The $3 billion 
figure represented our attempt to quantify the constraint that we 
perceived arose from the deficit. 6 

We also argued that a key part of the budget strategy should be that 
any stimulus that is provided should not be at the expense of an increased 
structural deficit. Hence, stimulus measures should be self-eliminating. 

The budget brought down by Finance Minister Lalonde in fact met 
these basic specifications. Only $1.9 billion in stimulus was provided in 


336 


the first year, and the deficit was 'tilted' so that it shrank with time. 
For a detailed discussion of this, see Bruce (1983). 

A central question respecting the performance of fiscal policy in the 
future is whether or not the government honours the commitment made in 
this budget to reduce the deficit as recovery proceeds. An essential 
ingredient of this budget was not direct stimulus but rather the winning- 
back of the confidence of the private sector. Thus, the rules must be 
'obeyed', and there is little room for discretionary expansion over the next 
few years. If the rule is broken, credibility of future policies will be 
seriously in question. 

APPENDIX 

In this appendix, we examine the effects of debt service transfers on the 
degree of automatic stabilization provided by the tax-transfer system. Let 
y t denote the deviation of output from trend at time t, b t denote the stock 
of government debt, and ♦ denote the time derivative; so b^ denotes the 
deficit at time t. Assume 

y t = sin t + y b t , (1) 

where y > 0 is the effect of the deficit on output. In the absence of the 
deficit, output would fluctuate regularly around a normal output level 
according to the sine function. 

Let 

b t = o t - py t + rb t , (2) 

where -p denotes the impact of output fluctuations on the deficit through 
changes in taxes and transfers, -a is an exogenous tax component, and r 
is the real interest rate (assumed constant) on the government debt. 

We can consider three simple rules. (i) Setting = py t - rb^ 
balances the budget at each point in time, (ii) Setting a = -rb^ balances 
the budget over the output cycle and allows deficits due to variations in 
output from trend, but tax-finances all fluctuations in debt service re¬ 
quirements. (iii) Finally, setting = rb, where b is the average stock 
of government debt over the cycle, balances the budget on average and 
allows deficits contingent on fluctuations in output and in debt service 


337 


requirements. 

Substituting (2) into (1), we can solve for 

y t = A[sin t + y(a t + rb^)], where (1') 

k = (1 + yp) ^ < 1. With a balanced budget, case (i), + rb^ = py t , so 

y t = sin t; that is, no automatic stabilization is provided. With a simple 
automatic stabilizer, case (ii), where a + rb^ = 0, we have y^ = k sin t. 
Since k < 1, the cyclical fluctuations in y are reduced. Given a conserva¬ 
tive 'multiplier' y of 1.5 and p = 0.30, we have k = 0.69; so fluctuations 
are reduced on the order of 31 per cent. 

Now let us consider the more complicated final case, where the fiscal 
rule also accommodates temporary fluctuations in the debt service payments 
due to fluctuations in the government debt over the cycle. First, substi¬ 
tuting (1') into (2) and setting = -rb yields the differential equation in 
government debt: 

b = A[-p sin t + r(b t - 6)], (2') 

which can be solved for: 7 

b t = b + pA./<5[Ar sin t + cos t], (3) 

where 6 = [1 + (Ar)^]. Substituting (3) into (2') gives 
y t = (A/S)[(l + kr 2 ) sin t + (1 - A) r cos t]. 

We can now calculate the accumulated absolute output deviations from 
trend over the cycle - i.e., 

t°+27X 9 i 

a = cr (y t r dtr 2 , 

t 

0 

where A^, A^, and denote the three cases described above: 


where 6 = [((1 + Ar 2 ) 2 + (1 - A) 2 r 2 )/6 2 ]. 

It can be shown algebraically that 0 must exceed unity if A is less 
than 1; consequently, A^ exceeds A2. It can also be shown that A6 must 

be less than unity if A is less than unity. Thus we have 

A 2 - A 3 - A r 

Both of the balance-over-the-cycle rules reduce output fluctuations relative 
to the strict-balance budget rules, but it is destabilizing to allow the 
deficit to absorb fluctuations in debt service expenditures resulting from 
cyclical fluctuations in the government debt. However, using r = 0.05, p 

= 0.3, and y = 1.5, we can evaluate A ^ as 0.6897 and A^ as 0.6906. 

Consequently, the effects of the debt dynamics are small, at least in the 
model considered. 


NOTES 


1 A second reason why the annually balanced budget may be destabilizing 
is because the tax increases enacted in order to balance the budget 
during recessions might be assumed to be permanent. In that case, 
households would believe their lifetime disposable wealth to be re¬ 
duced and would reduce their consumption accordingly. 

2 The true balanced-budget advocate would reject this approach. Milton 
Friedman is fond of drawing analogies between policymakers and alco¬ 
holics; in this context, the analogy would be the total-abstinence 
requirement that alcoholics refrain from drinking altogether. Simi¬ 
larly, in this view policymakers must eschew annual deficits. 

3 The model in the appendix assumes that output fluctuates regularly 
around the trend. Thus, integral stabilizers do not play any positive 
role and, as mentioned, reduce automatic stabilization slightly. 

4 Alternatively, if the budget balance is adjusted for the actual infla¬ 
tion rate, then some non-zero value must be targeted for if the actual 
and target inflation rates differ. 

5 For reasons noted earlier, we focus on the consolidated deficit. 

6 However, events have overtaken this policy prescription somewhat. 
Soon after we wrote the paper, the estimated federal deficit for 1983 
increased from $23 billion to nearly $30 billion. This increase did 
not come, at least totally, from a worsening of the cycle but from 
structural changes, a major cause of which was the collapse of gov¬ 
ernment revenues from oil. Also, the sharp fall in inflation raised 
real interest rates on the public debt (a situation that is temporary 
and not serious). More important, the fall in inflation nullified the 
real tax increase expected on the basis of the 6 and 5 per cent index- 


339 



ation cap of the tax brackets in 1983 and 1984. 

7 This solution 'starts' the economy a_t^ a particular point in the cycle 
(specifically, t^ = arc tan (-(rA) )) such that has no trend. 

REFERENCES 

Bruce, N. (1983) Report of the Policy Forum on the Federal Budget 

(Kingston, Ontario: John Deutsch Memorial for Economic Policy, 
Queen's University) 

Bruce, N., and D. Purvis (1983) 'Fiscal policy and recovery from the 
Great Recession.' Canadian Public Policy/Analyse de Politique , 53-70 

Canada, Department of Finance (1983) The Federal Deficit in Perspective , 
Budget Paper (Ottawa) 

Kydland, F., and E. Prescott (1977) 'Rules rather than discretion: the 
inconsistency of optimal plans.' Journal of Political Economy 85, 
473-93 

Lucas, R.E., Jr. (1975) 'Econometric policy evaluation: a critique.' 

Journal of Monetary Economics , Supplement I, 19-46 

Phillips, A.W. (1954) 'Stabilization policy in a closed economy.' Economic 
Journal 64, 290-323 

Sargent, T., and N. Wallace (1982) Some Unpleasant Monetarist Arithmetic 
(Minneapolis: Federal Reserve Bank of Minneapolis) 


William Mackness* 

I am the last speaker but one in today's very long and very interesting 
program. In these circumstances, I shall do my best to be brief. More¬ 
over, my not having had prior access to the Bruce-Purvis paper also 
recommends some economy in my remarks. However, lack of access to the 
paper also allows me some latitude in the scope of my remarks. 

Let me open with some general remarks. There seem to be a number 
of underlying propositions about public finance that have been accepted 
quite uncritically today. For example, there appears to have been a 
presumption throughout these proceedings that, as a practical matter, a 
higher level of deficit increases aggregate demand and a smaller level of 

* Vice-President and Chief Economist, The Bank of Nova Scotia. 


340 











deficit reduces aggregate demand. That is a handy framework, something 
of an analytical crutch. However, an analytical crutch is useful only if it 
is a relatively accurate reflection of the real world. Notwithstanding the 
benefit of economy of thought and analysis, it is simply not helpful to 
create the impression that output and employment necessarily benefit from 
expanding public spending and the deficit - and vice versa. 

Cast your imagination on this argument. It is an empirical fact that 
if the worst-managed economies in the world can accomplish anything, they 
can do the following two things: they can devalue and they can run 
deficits - on scales that are often boundless. One would always be cau¬ 
tious of the simple argument that there is some easy way out of economic 
difficulties - for example, all we have to do is devalue and we shall export 
our unemployment to our trading partners. In Canada, we have already 
tried devaluation, the first of the easy policy options. The tone of much 
of today's discussion indicates a belief that the second of the slick and 
easy policy options - deficit finance - may just work where devaluation 
certainly has not. 

It is so convenient and enticing to think that if you had the proper 
management team in place, you might just be able to get the deficit a little 
higher and somehow walk away from your difficulties. This is an unsound 
and dangerous argument. I think, in fact, that most of the deficit argu¬ 
ments rest on an extremely limited case. There can be serious and persis¬ 
tent imbalances of saving and investment, as there clearly were in the 
1930s. In such exceptional circumstances, where trust, confidence and 
the financial system have been devastated, there is both virtue and benefit 
in major recourse to public spending and deficit finance. However, to use 
the same techniques to finance buns-and-butter policies, policy-induced 
capital outflows, and the like is a plainly deficient policy on both a priori 
and empirical grounds. 

More to the point of today's discussion, I don't think that in a small, 
open economy it is a very bright idea at all to have a highly levered fiscal 
system that produces fearsome and unpredictable deficits. Anyone familiar 
with the official forecasting record of the most recent deficit knows for 
certain that the deficit is not a finely calibrated policy instrument. Even 
a kind assessment would record that the policymakers had virtually no 
ability to control or even to predict the future course of the policy instru¬ 
ment they were wielding or, more accurately, were attempting to wield. 

In fact, if the real problem with the economy is a lack of foreign 


341 



demand, it is not clear that abusing the fiscal system is going to move the 
economy a long way ahead - or move a lot of nickel out of Sudbury - or 
move a lot of lumber out of Port Alberni. Similarly, if one of the key 
reasons that the recession is so severe is that the private sector of the 
economy is vastly under-capitalized and caught up in a liquidity crisis - a 
liquidity crisis that is forcing layoffs and asset sales at a rate never seen 
since the 1930s - it is not clear that running up the fiscal deficit in 
Ottawa or the provinces is going to be particularly helpful. In Canada, 
the public sector deficit as a proportion of GNP is the highest among the 
industrial economies, with the single exception of Italy. 

Let me pose a question. Does Canada's deficit reflect in the main bad 
luck, deficient economic management, or a determined and sophisticated 
economic policy? I think the real fiscal and economic problems are far too 
complex to run on a simple-minded, single-variable policy rule. I liked 
the degree of skepticism that I heard in Neil Bruce's paper. Canada's 
record of fiscal management simply does not inspire much confidence. 

Let me make a quick aside on the issue of inter-generational shifts of 
resources. It is preposterous to argue that public assets should neces¬ 
sarily or desirably pass between generations accompanied by large indebt¬ 
edness. The inter-generational shift argument would have it that public 
assets should be accompanied by millions or billions of dollars of perpetual 
debt so that others can worry about paying for public assets, long-lasting 
or otherwise. Those who point to eventual redemption of indebtedness 
should be reminded that Second World War Victory Bonds will be rolled 
over this year, not redeemed. I think the whole inter-generational shift 
argument is idle and fallacious nonsense. Closer to home, the argument 
would have it that family assets passed between generations should come 
with a matching mortgage or other indebtedness. The inter-generational 
shift argument is both mean-minded and silly. 

So much of today's discussion of fiscal policy rules and targeting 

presupposes a completely unrealistic degree of predictive accuracy. Just 

for the record, only fifteen months ago, in November 1981, the first 

estimate of the federal deficit in fiscal 1982-3 was $6.8 billion. Subse¬ 

quently, this estimate was amended to $17 billion, then to $19 billion, and 
most recently it is rumoured to be approaching $30 billion. There is an 
important lesson here. Fiscal forecasting and fine-tuning are not a very 
precise science. For the record, it is worth remembering that the $6.8 
billion estimate was produced when the economy was already five months 


342 


into the worst recession in fifty years. I have been in the forecasting 
business too long not to realize that all forecasters live in glass houses. 
The point of reviewing the record of federal fiscal forecasting is simply to 
underline the severe limitations of fiscal forecasting and manipulation, 
particularly in a highly levered, open economy such as the Canadian 
economy. 

We have raised here today the question of crowding-out. Plainly and 
simply, we don't know if there is going to be crowding-out. We have 
never had today's scale of financial deficit in peacetime or on such a 
sustained basis. By our calculation at the Bank of Nova Scotia, the 
overall government financial requirements are going to amount to something 
on the order of 8 per cent of GNP this year and again next year. That 
represents a range of $30 to $35 billion on a conventional national accounts 
basis. On a cash basis, taking in the off-the-books financing for EDC 
and other federal and provincial crown corporations and so on, there will 
be an overall government borrowing requirement on the order of $45 billion 
dollars this year and again next year. This unprecedented government 
cash requirement should be measured against a total capital pool that might 
reasonably be estimated at $60 billion. Government borrowing has never 
accounted for half of the capital pool for two consecutive years before. 
Over the next two years, about 75 per cent of the savings pool will be 
devoted to meeting public sector borrowing requirements. I wouldn't care 
to wager a great deal of money that anyone can predict with any useful 
degree of accuracy what the results are going to be. However, I do not 
think these unprecedented levels of public sector borrowing create a good 
environment for an economy that has a very serious capitalization problem 
across virtually every industry in the country. 

The Canadian corporate sector is severely under-capitalized and has 
the highest degree of foreign ownership among the industrial economies. 
It is not helpful, I think, to point to the aggregate demand for corporate 
borrowing over the next couple of years as being quite low. To focus on 
the prospect of low net corporate borrowing demands over the next year 
or so is to miss the point. The problem in the corporate sector is that 
there is a vastly too great dependence on short-term floating rate debt, 
bank loans, and the like. The issue is that if the economy is going to 
grow again, the corporate sector must restructure the balance sheet. 
Even if the net borrowing demand in the corporate sector is fairly close to 
zero, it is not a good idea for the government sector to be into the market 


343 


for tens of billions of dollars at a time when it is necessary to restructure 
corporate balance sheets across the entire economy. There is a pressing 
need for new issues of corporate equity and long-term debt with which to 
retire today's excessively high levels of short-term debt. 

Let me comment here on the basic idea of a fiscal rule. I think the 
the Bruce-Purvis paper, if I interpret it properly, presents a sound 
argument on the question of a fiscal rule. Namely, there is no need for a 
rule that says you should raise taxes in the middle of a recession. I 
think, moreover, that it is useful to have some approximation of the 
amount of the deficit that is occurring because of what is happening in the 
economy, and I would not advocate strenuous efforts to correct that sort 
of deficit. I would argue, though, that it would be prudent in this 

country, where the economy is so open and so exposed, to have a rela¬ 
tively unlevered tax system. Moreover, I don't think it is a good idea, 

given the way energy prices are moving around, that the fiscal system be 

highly dependent on energy revenues. The general rule is that income- 

based tax revenues are inherently unstable. Consumption-based tax 
revenues are quite stable. 

A concluding point that is very, very important is that Canada has 
an extremely open economy with respect to capital markets and inter¬ 

national capital flows. To see the government deficit appearing to move 
out of control is very destructive of confidence and produces extremely 
destabilizing capital outflows. There is compelling empirical evidence that 
a rising deficit is probably the single best advance warning that a busi¬ 
nessman or a taxpayer can get of an impending tax increase, or worse. 
International investors are quick to confirm that loss of fiscal control 
correlates very highly with rising inflation, economic dislocation, rising 
government intervention, and sub-standard economic performance. Inter¬ 
national investors also have a record of acting on this perception. 

Let me make a closing remark about the politics of deficit finance. I 

think that a regular, heavy reliance on deficit financing is the antithesis 
of the democratic process. The whole point of representative government, 
going back better than a thousand years, has been parliamentary control 
of the pursestrings. The deficit subverts the process by providing for a 
level of public spending that is far in excess of what would be tolerated 
by the electorate. At this point I would remind you that federal govern¬ 
ment spending in Canada is running at a level fully 50 per cent above 
revenues. To me the evidence is compelling that if the books had to bal- 


344 










ance over the cycle, we would not have nearly as large a public 











sector 















Rapporteur’s remarks 

David Laidler* 


This conference has dealt with a wide range of issues having to do with 
deficits. We have had papers about the Canadian federal deficit (Wilson 
and Dungan, McCallum), about the Ontario provincial deficit (Auld), about 
the way in which financial markets view those deficits (Grant), about their 
significance when viewed in the light of the extreme openness of the 
Canadian economy (Wirick), and about the underlying principles of econo¬ 
mic, and indeed social, analysis that must inform any properly structured 
discussion of the conduct of fiscal policy (Modigliani, Parkin, and Purvis 
and Bruce). The conference has, I believe, been a success precisely 
because it has been so wide ranging, and because our theoretical discus¬ 
sions have taken place, not in a vacuum, but in the context of a pressing 
and important problem of current policy. It is as well, then, to begin this 
attempt to sum up our proceedings with some comments on the nature of 
that policy problem. 

THE CURRENT RECESSION AND THE DEFICIT 

To put it simply, the Canadian economy is, at this moment, in the very 
trough of the most serious recession it has experienced since the 1930s. 
Policymakers are under considerable pressure to 'do something' to alleviate 
this recession. Naturally enough, they turn to the traditional tools of 
demand management policies, developed by economists over the last four 
decades, to see what is available to them; and perhaps because of some 
residual 'Keynesian' instinct in their thinking, they seem to be looking 
mainly to fiscal policy. Fiscal policy is certainly important enough in its 
own right to merit single-minded attention, but I believe that at this point 


Chairman, Department of Economics, University of Western Ontario 


346 


I should say a word or two - necessarily assertive, given time and space 
constraints - about monetary policy. 

I have no doubt that the major cause of the current recession has 
been excessively tight monetary policy. It is about two years since the 
Bank of Canada's own choice of monetary aggregate. Ml, to all intents and 
purposes stopped growing. Other aggregates too showed a marked tend¬ 
ency to grow at slower rates. In the wake of this development, interest 
rates, both nominal and real, reached unprecedented heights from which 
only the recession has dislodged them. Of all the possible indicators of 
the stance of monetary policy, only the $US exchange rate has failed to 
show the traditional signs of 'tight money' by significantly appreciating; 
but that, as we all know, is because the United States too has had tight 
money. On a trade-weighted basis, the Canadian dollar has indeed appre¬ 
ciated. It is seldom that all possible indicators of the stance of monetary 
policy point in one direction, and economists of all schools should be able 
to agree on the importance of its role in generating the current recession. 

Be that as it may, because one factor causes a recession does not 
mean that another factor cannot alleviate it, and since the monetary au¬ 
thorities seem to believe - probably rightly, in my view - that any signifi¬ 
cant slackening of monetary policy carries with it a great risk of rekind¬ 
ling inflation, we are bound to consider carefully what can be accomplished 
by fiscal means. At first sight, this alternative seems to offer little 
promise. Fiscal deficits, both federal and provincial, appear to be run¬ 
ning at unprecedented levels. There does not seem to be room, within the 
financial constraints facing governments, in which to implement the 
spending or tax-cut programs traditionally brought into play in times of 
low output and high unemployment. It is a strongly held belief in certain 
quarters of the financial community, as Grant in particular has told us 
(without committing himself personally to that view), that past fiscal 
excesses have left governments powerless to deal with the current situation 
by any policies that will further increase their deficits (or even, according 
to some, leave deficits at their current levels). 

The underlying worry of the financial community is that ultimately 
capital markets will prove unwilling to lend to governments, so that the 
one government with access to the printing press, namely the federal 
government, will be forced to resort to it. It is fear of renewed inflation 
in the longer term - a fear well grounded in economic theory, as Michael 
Parkins's paper has shown us - that provides the foundation for current 


347 


concerns about the deficit, and particularly the federal deficit. As Auld 
has pointed out, the provinces cannot resort to the printing press and 
hence must eventually give in to pressures from capital markets to put 
their finances in order. For this reason, provincial deficits do not attract 
much attention. 

Much of the quantitative work presented to this conference has been 
directly addressed to the concerns I have just mentioned. In particular, 
it has been almost universally argued that current deficits are not the 
result of irresponsible fiscal policies implemented in the past, and that the 
bald figures in which they are typically measured (about $30 billion for the 
1983-4 federal deficit, given current policies) grossly exaggerate the 
strains they place on financial markets. It has also been argued, although 
with much less unanimity, that there is, even now, considerable room for 
fiscal stimulus in the Canadian economy. McCallum in particular has 
argued this position, and Wilson and Dungan have taken up a more re¬ 
strained version of it. 

As I discuss these matters, my own views will necessarily intrude, so 
let them be clear at the outset. I share the view that the current situa¬ 
tion respecting the deficit was not produced by irresponsible policies, but 
I am nevertheless doubtful that any room remains for further fiscal expan¬ 
sion. My reasons for taking this view will become apparent in due course. 

THE NOTION OF THE 'STRUCTURAL DEFICIT' 

Franco Modigliani surely set out a basic truth when he suggested that high 
deficits are in large measure the result, rather than the cause, of that 
ugly phenomenon (with an equally ugly name) 'stagflation'. To see why, 
it is helpful to begin with what Wilson and Dungan have called the 'struc¬ 
tural deficit'. This is the deficit that the government of any country with 
a given set of policies in place would incur, if its economy were running 
at a sustainable level of 'full-employment' output at a zero inflation rate. 
Now ask what would happen to the actual deficit of such a country if, with 
fiscal policies unchanged, it began to experience positive inflation and/or a 
level of output below 'capacity'. The answer in either case must be that 
the actual deficit would grow. 

With inflation, the nominal interest rate paid on public debt would 
rise and government cash outlays would grow. However, the related infla¬ 
tion-induced fall in the real value of government debt outstanding, which 


348 


represents debt retirement, would not enter into the government's 
accounts. If all government debt bore interest at a nominal rate fully 
indexed to the current inflation rate, these two effects would cancel each 
other out completely as far as the real state of government finances was 
concerned. However, given accounting procedures that do not adjust 
indebtedness for inflation, the government's measured deficit would 
increase. When debt is not fully and automatically indexed, the matter is 
more complicated; I will return to this point in a moment. 

The notion of adjusting the deficit for inflation is a fairly novel one, 
at least as far as Canadian policy discussions are concerned, but it ought 
to be relatively uncontroversial. On the other hand, the effects of reces¬ 
sion on deficits are well understood. As the economy slows down, tax 
receipts fall and benefit payments of one sort or another increase. That 
is how built-in fiscal stabilizers are supposed to work, and inherent in 
their operation is a deficit that increases relative to its structural level in 
a depressed economy. 

An economy suffering from inflation and recession simultaneously will 
find its deficit influenced by both of the factors I have just discussed, 
and the Canadian economy is clearly such an economy. Moreover, in the 
Canadian context the effects we are talking about are large. In round 
numbers, the inflation adjustment reduces the current deficit by about $8 
billion in 1982; and, if one treats the kind of unemployment rates that 
prevailed in the mid-1970s as being attainable and sustainable in the 1980s, 
the so-called 'cyclical' adjustment to the deficit essentially eliminates any 
remaining fiscal imbalance. Indeed, on some calculations it even puts the 
budget into structural surplus. 

All of this is little more than arithmetic, though it is fairly complex 
arithmetic, as those who look at the papers of Wilson and Dungan, 
McCallum, and Auld will soon discover. Even so, the arithmetic is un¬ 
controversial qua arithmetic. Its economic interpretation is altogether 
another matter, because the significance of the difference between the 
'structural' deficit and its actual value varies with the question being 
addressed. 

PITFALLS IN INTERPRETING THE STRUCTURAL DEFICIT 

The first question we may want to ask is whether current deficits reflect 
the inevitable outcome of persistent fiscal irresponsibility on the part of 


349 



I 

governments. To answer this question, I would argue, it makes good 
sense to ask what the deficit would be if the unemployment rates of the 
mid-1970s were feasible now, and also to ask what the deficit would be if 
there were no inflation. That is to say, the full adjustments made by 
McCallum, Wilson, and Dungan for the federal deficit, and by Auld for the 
provincial deficit, are relevant. When they are made, the answer is un¬ 
equivocal: the deficits we now face are not the result of fiscal irrespons¬ 
ibility. They are indeed the consequence of stagflation. 

At the provincial level, we may take this answer at its full face 
value. Stagflation was not caused by the policies of the Government of 
Ontario or the government of any other province. In relation to the 
provinces, stagflation has been an exogenous phenomenon that may truly 
be said to have played havoc with provincial finances. Auld exonerates 
the Government of Ontario of fiscal irresponsibility, and he is right to do 
so. 

At the federal level, the matter is more complex. Fiscal policy, in 
the sense of expenditure and tax or transfer changes, has not caused the 
deficit. Stagflation, however, has, and I have already suggested that 
much of the blame for stagflation is to be laid at the door of monetary 
policy. Thus, in saying that I agree with Wilson and Dungan, not to 
mention McCallum, that fiscal policy per se has not caused the federal 
deficit, I do not mean to imply that the federal government is a victim of 
circumstances. There has been a long-standing failure to coordinate 
federal fiscal and monetary policies in Canada, and this failure has had 
much to do with the current state of the economy, including the state of 
the federal deficit. The federal government cannot be absolved of its 
responsibility for this policy failure. 

The question of how we got here is one thing; the question of what 
to do about it is another thing altogether. The adjustments of the 
measured deficit that we have discussed so far are certainly relevant to 
the latter question, but for several reasons they must be treated with 
great care. 

To begin with, note that the inflation and cyclical adjustments are 
conceptually very different from one another. We make the former adjust¬ 
ment because, in an inflationary economy, the measured deficit overstates 
the rate at which real government debt is being accumulated by the 
general public, just as on the other side of the national accounts measured 
private sector saving overstates the very same rate of accumulation. In 


350 



other words, the inflation adjustment is a simple accounting change 
designed to distinguish more accurately between current government cash 

outlays and debt retirement. The cyclical adjustment is a very different 

matter. The debt generated by the operations of built-in stabilizers is not 
a statistical illusion. It is real debt with real consequences for financial 
markets. Moreover, the magnitude of any cyclical adjustment must rest on 
a whole set of assumptions about how the economy functions now and is 
likely to function in the future. 

Though the inflation adjustment is conceptually straightforward, in 
practice it presents pitfalls that need careful attention when we discuss 
the appropriate stance for current fiscal policy. If the interest rate on 
the national debt were fully indexed to the current inflation rate, there 
would be no difficulty. Every change in the inflation ..rate would be 
reflected in nominal interest payments, and the inflation adjustment could 

be made by transferring the indexation portion of interest payments from 

current expenditures to some debt retirement account. The difficulty is 
that, in contemporary Canada, the national debt consists of instruments 
bearing interest at rates that reflect inflation expectations at the time of 
issue, and that will not vary until the instruments in question are retired, 
regardless of what may happen to the inflation rate in the interim. 

I have already remarked that the inflation adjustment to the deficit 
for 1982 in Canada is about $8 billion. If, as now seems quite likely, the 
year-on-year inflation rate up to the end of 1983 turns out to be in the 
region of 6 per cent, the relevant adjustment for 1983 will be somewhere 
around $5 billion plus. (I am indebted to Douglas Purvis for this very 
rough estimate.) Interest payments on debt not maturing in 1983 will 
continue at their old rates, but inflation will not erode the real value of 
that debt as quickly; consequently, the economy will receive a real stim¬ 
ulus of close to $3 billion this year from falling inflation alone. This does 
not necessarily mean that further active stimulus should be avoided, but it 
is something which must be taken into account when considering this 
question - as indeed it is, implicitly, in the econometric projections of 
Wilson and Dungan and McCallum. 

As I suggested earlier, the cyclical adjustments to the deficit that 
have been discussed at this conference are altogether more difficult to 
assess than inflation adjustments, not only because the debt created when 
the economy is in recession is genuine debt, but also because quantitative 
estimates of just how far below its feasible long-run operating ceiling the 


351 


economy now is are controversial. I am bound to say that estimates based 
on the assumption that unemployment rates of between 7 and 8 per cent 
are on average feasible over any reasonable medium-term horizon seem 
optimistic to me. I say this for a number of reasons. 

First, it is now five years since we have seen unemployment rates 
that low, and it is likely to be even longer before we see them again, even 
according to the forecasts of Wilson and Dungan. An unemployment rate 
that does not seem attainable for even a short period in more than a 
decade does not look to me like a feasible long-run average unemployment 
rate on which to base estimates of the Canadian government's structural 
deficit. 

A second and related reason is this: though the inflation rate is now 
well down and falling fast, the expectations generated by a decade of 
serious inflation are not going to erode quickly. Through this cycle and 
the next, down to the end of the decade, say, the economy is probably 
going to have to run, on average, with excess capacity, if inflation is to 
be squeezed out of the system. If 7 to 8 per cent unemployment could be 
maintained on average in the Canadian economy in the absence of deeply 
ingrained inflation expectations - and perhaps it could be - this fact would 
still not be of much relevance in the presence of such expectations. 

I would very much like to be wrong about this matter, because I get 
no pleasure from contemplating so long a period of slack economic activity. 
However, I get even less pleasure from contemplating the consequences of 
an expansionary policy that, in seeking to tackle unemployment, under¬ 
estimates the risks of rekindling inflation and puts Canada in the position 
of having to undergo an even deeper and longer recession than the cur¬ 
rent one later in this decade. 

My last concern about estimates of what is a feasible long-run un¬ 
employment rate stems from a factor drawn to our attention by Michael 
Parkin. The average unemployment rate that is compatible in the long run 
with a zero average inflation rate is not a unique number. It depends 
upon how much the inflation rate varies about an average value of zero. 
In the short run, it seems to take a greater increase in unemployment to 
reduce the inflation rate by a given amount than it takes a decrease in 
unemployment to increase the inflation rate by the same amount. There¬ 
fore, the more variable is the inflation rate, the higher is the average 
unemployment rate compatible with an average inflation rate of zero. If I 
understand them correctly, the estimates of a sustainable unemployment 


352 


rate that underlie the cyclical adjustments to the deficit discussed at this 
conference are based on the assumption of a constant and not a variable 
inflation rate. Hence they must be regarded as optimistic by anyone who 
thinks that an average inflation rate of zero, accompanied by some short¬ 
term variability , is the best that we can ever hope to achieve. 

THE DANGERS OF CROWDING-OUT 

The arguments I have sketched suggest that the structural deficit in this 
country is larger than some of the participants in this conference have 
suggested, but it does not follow that it is intolerably large. To begin 
with, it is quite fallacious to think that the 'right' value for the structural 
deficit in a well-run economy is zero. As Bruce and Purvis have pointed 
out, in a growing economy the private sector's wealth will grow. If that 
sector wishes to hold a roughly constant fraction of its wealth in govern¬ 
ment debt, then, on average, the government must run a deficit - and 
perhaps quite a sizeable deficit - if that demand is to be satisfied. In an 
economy growing at 4 per cent per annum, in which the private sector 
wished to hold public debt to the value of 50 per cent of one year's 
income, the structural deficit would have to run, on average, at an amount 
equal to 2 per cent of national income. 

Wilson and Dungan have in fact presented us with some information on 
the way in which the debt-to-income ratio has been moving over time in 
Canada. This is a particularly useful statistic, because, as both they and 
Modigliani have told us, the inflation adjustment to the current deficit is 
automatically built in when it is calculated. What Wilson's and Dungan's 
data show is that this ratio, which was well over 100 per cent as a result 
of the Second World War, steadily fell until the early 1970s to a value less 
than 30 per cent and has since been climbing. It is now back to the 
levels of the early 1960s. According to Wilson's and Dungan's estimates, it 
will stabilize at about 50 per cent by 1990 if current policies do not 
change. 

Thus, one question we face is whether a national debt equal to about 
one half of one year's income will put an undue burden on the Canadian 
economy. The typical layman's view of such an issue is that the more 
government debt there is outstanding, the higher will be the interest rates 
that the private sector must pay. Thus, so the argument goes, govern¬ 
ment deficits 'crowd out' private sector capital formation and undermine 
growth. 


353 





There are at least two fallacies involved in this position. First, as a 
number of speakers have reminded us (particularly Auld in his discussion 
of the Ontario government's deficit), governments themselves participate in 
the process of capital formation: they do build roads, airports, schools, 
hospitals, and so forth. Expenditure transferred from the private sector 
to the public sector is not necessarily expenditure transferred from a 
productive purpose to a wasteful one. Second, and equally important, the 
capital market in which Canada can borrow, on both private and public 
sector accounts, is far wider than the borders of this country. 

Wirick's careful and useful study has addressed the latter point. He 
has shown that on the basis of the historical evidence Canadian borrowing 
is so small a part of the North American capital market that we are essen¬ 
tially price-takers in that market. This does not mean that we can always 
borrow as much as we like as cheaply as we like, but it does mean that 
variations in our borrowing, if they remain within the boundaries of past 
observations, are unlikely to affect the terms we can get. That is to say, 
whatever else the deficit might do to Canada, it is unlikely to crowd out 
private investment. 

However, there is an important qualification here. If the ability to 
borrow abroad means that we can continue to invest, regardless of the 
deficit, it does not mean that we can continue to enjoy the full benefits of 
that investment. Interest payments to foreigners represent a real drain on 
Canada's resources. For a given level of national output, the lower is our 
foreign debt, the higher are our living standards. 

More crucial still, in an open economy, the interest rate/investment 
channel is not the only one whereby public sector expenditure crowds out 
the private sector. Indeed, in the Mundell-Fleming model that Wirick 
cites, government expenditure completely crowds out private sector 
activity - not through the interest rate, but through the exchange rate - 
given that this rate is flexible. The capital inflow generated by borrowing 
abroad appreciates the currency and hits the demand for exports and 
import substitutes. 

This Mundell-Fleming result is not irrelevant. Something very like it 
was going on in Canada in 1975-6. It is nevertheless true that no one 
nowadays fears that rising deficits will unduly strengthen the Canadian 
dollar on foreign exchange markets. Quite the contrary! The great, and 
widespread, fear is that the size of the current federal deficit is under¬ 
mining confidence in the stability of the Canadian dollar. If it is, then 


354 


any attempts on the part of the authorities to combat the recession by 
applying fiscal stimulus might result in a substantial exchange deprecia¬ 
tion, a renewal of inflation, and ultimately an even more severe recession, 
as the monetary authorities attempt to offset these effects through even 
tighter monetary policy. 

To put it simply, the longstanding inconsistency between monetary 
and fiscal policy in Canada, to which I referred earlier, is still perceived 
in many quarters to exist. This perception is generating expectations in 
financial markets that make fiscal expansion in current circumstances 
possibly counter-productive and even downright dangerous. 

THE ROLE OF EXPECTATIONS 

Expectations are notoriously difficult to handle in economic analysis, but 
nothing has emerged more clearly from this conference than an under¬ 
standing of how crucial they are to our current policy dilemma. John 
Grant has gone so far as to refer to 'paranoia' in financial markets. He 
means by this, I believe, that fears of the imminent collapse of the 
Canadian dollar are hard to justify in the basis of present evidence. 
Nevertheless, as Grant has warned us, the past experience of financial 
markets has given them a lot to be paranoid about. Given the current 
situation, and given the forecasts set out in the papers by Wilson and 
Dungan and by McCallum, the deficit seems manageable both now and in 
the future. But two or three years ago it seemed as though the federal 
deficit for 1983 would be less than $10 billion. Instead it is three times as 
big. Why, then, should financial markets believe the forecasts this time 
around? After all, what would become of them if OPEC collapsed and there 
were a full-scale price war in the oil market? 

We must turn to the papers by Parkin and by Bruce and Purvis for 
further guidance here, and I say 'guidance' rather than 'comfort' because 
the implications for the current policy debate of these apparently abstract 
pieces of analysis are far from encouraging. The essential results of 
Parkin's survey of recent work in the 'rational expectations' tradition of 
macroeconomics are easily summarized. If a government sets out on a 
course of policy that will ultimately - and 'ultimately' can be a very remote 
time in the future - force it into inflationary finance, and if the public 
believes that the government will stick to that policy, then the public will 
act in a way that will cause that 'ultimate' inflation to occur here and now. 


355 


It is important not to take such theoretical results literally. Tradi¬ 
tional, and now discredited, Keynesian macro-models used to treat the 
government as farsighted and all-wise and the private sector as myopic, as 
if it behaved according to the same old rules of thumb regardless of what 
the government did. By the same token, many modern 'rational expecta¬ 
tions' models deal with a world in which a farsighted and all-wise private 
sector is confronted by a totally stupid government, which pursues the 
same policies regardless of private sector reaction. Both approaches are 
unrealistic. The public learns from experience, but so do governments. 

However, not taking results literally is not the same thing as not 
taking them seriously. The models that Parkin discusses tell us what will 
happen if adverse expectations about future government behaviour are 
engendered by current actions. Hence they tell us that, in the conduct of 
policy, avoiding the creation of such adverse expectations is crucial. 
Purvis and Bruce, in advocating the governing of fiscal policy by some 
rule requiring balance in the structural deficit (with due allowance for 
economic growth), are advocating the establishment of a social institution 
that would generate favourable expectations about deficits in the private 
sector, regardless of current conditions. They hope that such favourable 
long-term expectations would stave off the catastrophic consequences, 

implicit in some of the models that Parkin has analysed, of current in¬ 
creases in the deficit. 

The trouble is that the kind of rule Purvis and Bruce advocate would 
not attain credibility simply be being legislated. Its credibility could 
emerge only from the constant interaction of the authorities and the priv¬ 
ate sector, interaction not just in their economic activities, but also in 
their political activities. In other words, credibility is born of experience. 
Moreover, once a credible rule is established, its durability should not be 

taken for granted. A rule for the conduct of economic policy will be kept 

credible only by constant political activity designed to stave off whatever 
pressures might be put on the authorities to abandon it. 

This amounts to saying that certainty concerning the responsible 

conduct of economic policy in the future is simply not available in this 
world. Fortunately, it would appear that the economic system can live 
tolerably well with a good deal less than certainty. If it could not, it 
would have collapsed long ago. Nevertheless, what we are learning in 
Canada at this moment, or rather what we are being reminded of, is that 
the stability of the economic system cannot be taken for granted; that 


356 


trust in the responsible behaviour of government, and in the capacity of 
the political process to enforce such responsible behaviour, can be danger¬ 
ously eroded. It is precisely because trust in the authorities has been 
eroded over the last few years, particularly the trust of those operating 
our financial system, that the federal deficit presents so difficult a prob¬ 
lem at this moment. Grant may believe that the markets are 'paranoid', 
but he is also warning us that the actions of paranoid people have real 
consequences, even if they stem from what the rest of us regard as delu¬ 
sions. He is right to give us this warning, and we should heed it. 

CONCLUSIONS 

It is not difficult to state succinctly the major conclusions that seem to me 
to have emerged from the papers presented at this conference. First, the 
deficits we are now seeing in this country, at both the federal and the 
provincial levels, ought not to be blamed on irresponsible fiscal policy. 
They are not the result of our elected representatives' going on an undis¬ 
ciplined spending spree. Rather the deficits in question are the result of 
'stagflation'. While provincial governments are not to be blamed for this 
phenomenon, and while the Canadian economy is undoubtedly in some 
measure the victim of circumstances prevailing throughout the western 
world, it is not an accident that Canada is suffering stagflation on a 
grander scale than most countries. The federal authorities have failed to 
coordinate their various policies over the last few years, particularly their 
fiscal and monetary policies, and that failure has contributed in an impor¬ 
tant way to the current situation. 

Some comfort is to be found if we simply compare the raw data on the 
current deficit with data generated in the past. A deficit that is largely 
due to stagflation is likely to go away as stagflation abates, provided 
nothing happens in the interim. Furthermore, even now, the country's 
debt-to-income ratio is at a rather low level by historical standards. It is 
no higher than it was in the 1950s and early 1960s, and those were years 
of steadily growing prosperity in Canada. Thus it is hard to argue that 
the national debt currently imposes a burden on the private sector that 
the latter cannot bear. 

The trouble is that government debt is growing rapidly at this 
moment, and fears that it might become unmanageable in the future are 
already undermining confidence in certain important sectors of the eco- 


357 


nomy, notably the financial sector. Conventional forecasts suggest that 
such fears, if not quite groundless, are certainly exaggerated, but they 
are there and must be taken into account in the design of policy. In 
large measure the authorities are boxed in, not by objective factors 
per se , but by the way in which those factors are interpreted by key 
agents in the economy. This is unfortunate, and the authorities' task 
would be much easier if these subjective elements has not intruded upon 
the scene; but they have intruded and they are important. 

What, then, can be done in the current situation? My own view is 
that, above all, we must avoid a resurgence of double-digit inflation. 
Given the current state of knowledge, it is pointless to pretend that we 
know of any cure for inflation other than slack economic activity, because 
we don't. The more deeply inflation expectations are embedded in the 
economy, the more slack we need in order to eradicate them. The current 
recession is providing that slack. It is, in large measure, a fait accompli , 
and we may as well take advantage of it. If we end it too quickly, we 
shall have suffered all of its costs to obtain no lasting benefits. We shall 
simply have to repeat the whole process later on an even more painful 
scale. 

Thus, sadly and not very helpfully, I must conclude that the current 
deficit and the way in which it is perceived effectively prohibits the author¬ 
ities from undertaking anything more than a short-term token stimulative 
policy. If the Canadian economy is to be put upon a stable long-term 
footing, we must let the current recession take its course. The risks of 
doing otherwise are too great. 


Discussion 


JOHN BOSSONS: David, I liked your way of summarizing where we end 
up, because I think it is terribly important to recognize the importance of 
policy credibility and the strong appeal of rules. Nevertheless, it is also 
necessary to recognize that politcal pressures will force the government to 
be pragmatic and to respond to the pressures that are implied by what is 
likely to happen over the medium term. 

One of the problems in talking about the deficit is that we are not 
always talking the same language. I personally believe that it is very 


358 





important over the long run to control government expenditures. Indeed, 
one reason for my emphasizing the importance of measuring the structural 
deficit in cyclically adjusted terms is that I feel we must have an effective 
way of focusing attention on what is happening to the long-term stance of 
the government fiscal balance, so that we don't get misled into thinking 
that things are better or worse than they are. 

Measuring the structural deficit is something that helps us understand 
where we are. It is a very different matter to ask where we should go. 

There are major advantages in adhering to a fixed policy rule in the 
long term. Nevertheless I would be very concerned if people thought 
these long-term advantages sufficient to warrant not deviating from such a 
rule in the short term. 

What has happened as a result of monetary policy (and to a lesser 
extent fiscal policy) over the last few years is that people have been 
burned. With the substantial volatility of interest rates and the sharp 
plunge in the economy, businessmen and investors in the financial markets 
have become risk averse to a degree that makes our present situation very 
different from any other post-war recession. It is this qualitative differ¬ 
ence between where we are now and where we have been in any other 
recession since the Second World War that worries me. I am concerned 
about our getting misled by the high level of the deficit into thinking that 
we should not do anything about the secular stagnation that is likely to 
persist over the next half decade. 

Even though one would want in the longer term to build up the 
credibility of policy rules, I think it is very important to recognize that 
this is one of those times when there would be real advantages to devi¬ 
ating temporarily from the policy rule that would be optimal in the long 
term in order to deal with a serious middle-term problem. This middle- 
term problem is the likelihood that we won't come within 3 or 4 percentage 
points of full employment even five years from now. If that is a correct 
characterization of the current medium-term outlook for the economy, and I 
think it is, then achieving better economic performance has to be given 
high priority. 

My own prescription is fiscal stimuli that are primarily directed to 
increasing business capital formation. I am sure there are many others. 

In any case, I do think we have to recognize that at least one dimen¬ 
sion of the problem is that there is an unacceptably high probability of 
staying in a period of relative stagnation as a result of the risk aversion 


359 


that has been built into the system by what has happened to monetary and 
fiscal policy over the last few years. The fact that what has happened is 
itself a result of deviation from the desired long-term policy rule does not 
make blind allegiance to such a rule good policy once such a deviation has 
occurred. 

'Letting the current recession take its course' might well be an ideal 
anti-inflationary strategy if one could afford to ignore the effect of con¬ 
tinued high levels of unemployment on the political system. However, 
doing so for a protracted period risks building up politically irresistible 
pressures for greater stimulus. If such pressures cannot be reflected in 
fiscal stimulus because of concern over the deficit, there is a high risk 
that they will be reflected in monetary stimulus. It is dangerous to 
assume that it is politically feasible to maintain a policy stance that implies 
a protracted period of high unemployment. 


360 


Members of Ontario Economic Council 


Thomas J. Courchene 
Chairman 

Ontario Economic Council 


Gerard R. Beaulieu 
Administrator 

Seafarers' Training Institute 
Morrisburg, Ontario 

Jalynn H. Bennett 
Financial Vice-President 
Manufacturers Life Insurance 
Company 
Toronto, Ontario 

W. Lyle Black 
Greig Medical Group 
Bracebridge, Ontario 


Gail C.A. Cook 
Executive Vice-President 
Bennecon Limited 
Toronto, Ontario 

Lucien P. Delean 
Partner 

Critchley & Delean Architects 
North Bay, Ontario 

E. Gerard Docquier 
National Director 
United Steelworkers of America 
Toronto, Ontario 

John Grant 

Director and Chief Economist 
Wood Gundy Limited 
Toronto, Ontario 

William A. Jones 
Secretary-Treasurer 
Ontario Teachers' Federation 
Toronto, Ontario 


Robert W. Korthals 
President 

Toronto Dominion Bank 
Toronto, Ontario 


William Mackness 

Vice President & Chief Economist 
The Bank of Nova Scotia 
Toronto, Ontario 

Samuel A. Martin 
School of Business Administration 
University of Western Ontario 
London, Ontario 

Elizabeth Parr-Johnston 
Corporate Strategies 
Shell Canada Limited 
Toronto, Ontario 


Clifford G. Pilkey 
President 

Ontario Federation of Labour 
Don Mills, Ontario 

Bruno R. Rubess 
President 

Volkswagen Canada Inc. 
Scarborough, Ontario 

Murray Rumack 
Partner 

Clarkson Gordon 
Toronto, Ontario 

David C. Smith 
Department of Economics 
Queen's University 
Kingston, Ontario 

Clayton M. Switzer 
Dean of Agriculture 
University of Guelph 
Guelph, Ontario 

Donald J. Taylor 
Executive Vice-President 
and Director 
Shell Canada Limited 
Toronto, Ontario 

David M. Winch 
Department of Economics 
McMaster University 
Hamilton, Ontario 


361 



HC 
117 
. S741 
. D313 


Ontario Economic Council 
Deficits. 

dd \p 









Ontario Economic Council Papers 
Special Research Reports 


Energy Policies for the 1980s: an economic analysis 

Developments Abroad and the Domestic Economy 

Policies for Stagflation: Focus on Supply 

Inflation and the Taxation of Personal Investment Income: 
an analysis and evaluation of the Canadian 1982 Reform Proposals 

Inflation and the Taxation of Personal Investment Income: 
an Ontario Economic Council Position Paper on the Canadian 1982 
Reform Proposals 

A Separate Personal Income Tax For Ontario: 

An Ontario Economic Council Position Paper 

A Separate Personal Income Tax For Ontario: An Economic Analysis 

The U.S. Bill of Rights and the CanadiarfCharter of Rights and Freedoms 

Pensions Today and Tomorrow 

Deficits: Flow Big and Flow B^d? 


Copies of these publications are available at a nominal charge 
from the Ontario Government Bookstore, 880 Bay Street, Toronto 
to those shopping in person. Out-of-town customers may write: 
Publications Section, Fifth Floor, 880 Bay Street, Toronto, Ontario, 
M7A1N8, or telephone 965-6015 (toll-free long distance, 
1-800-268-7540; in northwestern Ontario, O-Zenith 67200). 


ISSN 0225-591X 


ISBN 0-7743-8497-2