Daniel L. Thornton

Dan/el L. Thornton is an assistant vice president at the FederalResence Bank of St. Louis. Lynn D. Dietrich provided researchassistance.

Do Government DeficitsMatter?

nflSit, HERE HAS BEEN considerable concern

about the size of the federal deficit and therather modest success of the Gramm-Rudman-Hollings Act to reduce it. The traditional, that is,Keynesian, view of deficit spending in macro-economics was that it could smooth out fluctua-tions in economic activity due to gaps betweensaving and investment that were primarily theresult of exogenous shifts in investment. Fromthis perspective, deficit spending was seen asboth desirable and necessary to offset cyclicalfluctuations in economic activity that were char-acteristic of capitalist, free-market economies.

Recent discussions of government deficits,however, have focused on their alleged adverseeffects. It is now common to blame deficitspending for high real rates of interest and thelarge and persistent trade deficit. Moreover,there is widespread concern that large and per-sistent federal deficits will produce stagnant

economic growth and result in renewed infla-tionary pressures.

An alternative view, called Ricardian equiva-lence, however, sees deficit spending as a har-binger of neither good nor ill. According to thisview, deficit spending cannot offset fluctuationsin economic activity due to exogenous shifts ineither private saving or investment; nor, can itbe blamed for high real interest rates or thelarge trade deficit. Moreover, it has no influenceon the outlook for economic growth or infla-tion. Macroeconomic tests of the effects ofdeficits on the economy using U.S. time-seriesdata have generally favored the Ricardian view.’

The purpose of this article is to review bothviews of deficit spending to determine whichview of the relationships between deficit spen-ding and various associated macroeconomic var-iables is supported by evidence from 16 OECDcountries over the period 1975.86.2

‘Recently, there have been three excellent surveys of theempirical investigations of Ricardian equivalence. SeeBarro (1987), Bernheim (1987) and Aschauer (1988). Also,see Evans (1988), Koray and Hill (1988) and Leidermanand Razin (1988) for more recent work that is not sum-marized in the three surveys. For some very recent workon this issue, see Feldstein and Elmendorf (1990),Modigliani and Sterling (1990) and Kormendi and Meguire(1990). Most of the empirical evidence against the Ricar-dian view is at the microeconomic level using cross-sectional or panel data.

reasons. First, strictly speaking, Ricardian equivalence is aproposition about what happens when deficit spending issubstituted for taxes at an unchanged expenditure level.Second, in a rational expectations framework, only unan-ticipated changes in deficits should affect macroeconomicvariables, and there is no plausible way to isolate theunanticipated component of deficits from these annualdata. Third, the deficit measures used here are notcyclically adjusted. Hence, they are endogenous, at leastin part.

2The reader is cautioned that the evidence presented herecan be considered suggestive for only a couple of

I)J:tIH:1(lJflFSt)JF;i%T1)t.

Total saving has two components: public andprivate saving. Public saving measures thegovernment’s surplus or deficit position: sur-pluses represent positive public saving, whiledeficits represent public dissaving. In the Keynes-ian view, deficits can be used to offset gapsbetween saving and investment, thereby stabiliz-ing output around its potential (full-employment)level. For example, if private saving is too highrelative to investment to achieve potential out-put, a government deficit will reduce the amountof total saving and close the gap between savingand investment. Conversely, if private saving istoo low relative to investment, a governmentsurplus can increase total saving. Hence, cyclicalswings in economic activity arising from move-ments in the saving/investment gap can beshortened by changing public saving appropri-ately to maintain total saving at a level consis-tent with potential output.

•L~~t~:f/.(5anti .i’n’vateSavi.ng

The above analysis is based on the assumptionthat public and private saving are separate anddistinct activities.’ In the extreme, they areviewed as totally independent: changes in publicsaving have no direct effect on private savingand vice versa.4 In this case, total saving willchange one-for-one with changes in public sav-ing. In the less extreme case, total saving issimply positively correlated with public saving.’

Lktti/ICttS ~SFt/%4:n,iEflt’SL .Hatat;

In the conventional view, increases in thedeficit cause the interest rate to rise.°Since the

interest rate is the price of credit, it is deter-mined by the supply and demand for credit. Ac-cording to the traditional view, an increase inthe deficit increases the demand for creditrelative to the supply and, consequently, in-creases the interest rate.7

IlPe:JicIis tDId th.e Tracts iJeiic~ii

The effect of deficits on credit demand canalso affect the relationship between governmentdeficits and the trade deficit. The effect ofdeficit changes on interest rates depends on theslope of the credit supply curve—the steeperthe slope, the larger the effect on interest rates.Among other things, the slope of the creditsupply curve depends on the degree of open-ness of the nation’s money and capital marketsto the rest of the world- In general, the moreopen these markets are, the flatter is the supplyof credit.

To see this more easily, let’s recast the discus-sion in terms of the demand for securities. Inthis framework, the price of bonds and, hence,the interest rate are determined by the supplyand demand for bonds. The bond and creditmarkets are related inversely: those who supplycredit are demanding bonds while those whodemand credit are supplying bonds. If financialmarkets are open and competitive, the demandcurve for bonds seen by individual bond sup-pliers, including the government, is flat at themarket price of securities. In other words, nomatter how many bonds that individuals, firmsor the government may supply, individually,they see no effect on the market price of bondsand, hence, on the interest rate. All market par-ticipants, including the federal government, are

‘The following discussion is based on the standard lS/LMaggregate demand-aggregate supply model of themacroeconomy, where deficit spending is treated asexogenous.

4Also, deficits can have an indirect effect on saving if theincrease in the deficit spending raises real income and,hence, saving. Likewise, if a shift in the saving or invest-ment functions cause a change in income, the deficit willrespond endogenously because both expenditures andtaxes are related to the level of income. Hence, theassumption that the series are independent applies to theirautonomous components.

5Strictly speaking, the above conclusion holds as long aspublic and private savings are not perfect substitutes.Note, however, fiscal policy is used to offset changes inprivate savings, there may be a negative correlation bet-ween public and private savings even, if at a moreabstract level, the series are unrelated. That is, there maybe a policy reaction function where deficits respond,presumably with a lag, to changes in private savings.

°Thediscussion here abstracts from the possible effects ofdeficits on the rate of inflation, so the hypothesized effectin this section is on the real interest rate.

‘Most standard IS/LM aggregate demand-aggregate supplymodels assume that there is no direct effect of the in-crease in the deficit on private savings; however, thisassumption is not necessary to obtain the usual implica-tions. All that is required is that the direct effect on privatesavings (if the effect is to increase it) be less than the in-crease in the deficit. Also, the effect on the interest ratewould be small if the monetary authority monetized debt tokeep real interest rates low. While this is a controversialissue, the evidence for the U.S. suggests that the FederalReserve has not monetized the debt. See Thornton (1984).This discussion abstracts from the effect of the deficit oninterest rates via a deficit-induced increase in the rate ofinflation.

Figure 1Price Level and Output Determination

“small” relative to the total market, that is, eachis a price-taker.”~Thus, if the economy is suffi-ciently open, the money and capital markets arecompetitive and the government is a price-taker,changes in deficit spending will have no effecton the interest rate. Increases or decreases inthe supply of bonds will be fully absorbed atthe current market price.

This does not mean, however, that deficitspending has no effect on the economy. In theabove case, the deficit will be matched byforeign claims on U.S. assets. That is, there willbe a capital inflow which, given the balance-of-payments identity, must be matched by a tradedeficit. Hence, if the economy is “sufficiently”

open, there may be no (zero) correlation be-tween government deficits and interest rates,but there may be a positive correlation betweenthe government and trade deficits.

The strength of this correlation depends onthe degree of openness of financial markets. Iffinancial markets are only partially open or ap-proximated by the competitive model, thegovernment deficit will be positively correlatedwith both the interest rate and the trade deficit.

The effect of deficits on economic activity canbe illustrated in terms of the aggregate de-mandiaggregate supply model. The traditionalview asserts that increases in the deficit will in-crease the demand for goods and services, as il-lustrated in figure 1. The extent to which theincrease in aggregate demand affects output orprices is determined (among other things) bythe slope of the aggregate supply curve. The

Output steeper the slope of the aggregate supply curve,the larger the effect on the price level and thesmaller the effect on output. If the aggregatesupply curve were vertical, the deficit would af-fect only the price level.

Economic theory and empirical evidence sug-gest that the aggregate supply curve is upward-sloping in the short run.°If this is true, thereshould be a positive correlation between deficitsand both prices and output.

In addition to the positive effects of deficitson output, it is often suggested that deficits canaffect output adversely because of their impacton interest rates. In a closed economy, someargue, the deficits raise the real interest rateand, thereby, “crowd out” private investment.bo

eThis argument also applies to the Federal Reserve’s con-trol over real interest rates in an open economy, and mayhave implications for the Fed’s control over interest ratesin a closed” economy as well.

9The “long run” tends to mean different things inmacroeconomics depending on the context in which it isused. In the present ISILM aggregate demand-aggregatesupply framework, it means the period of time for allwages and prices to adjust to exogenous shocks. If allwages and prices adjust instantaneously, the aggregatesupply curve would be vertical for all time periods. Hence,the positively sloped aggregate supply curve in this modelresults from some imperfections” which keep wages andprices from adjusting instantaneously to shocks.

‘°Bernheim(1989) and others call this the neoclassical view.In the neoclassical framework, crowding out necessarily

results from a full-employment assumption. The only waythat the government can spend more is for the private sec-tor to spend less. That is, interest rates must rise to in-crease private savings by the amount of the decrease inpublic savings. The much-hypothesized adverse effects ofgovernment deficits in the neoclassical view come eitherthrough current government consumption replacing privateinvestment or the assumed lower marginal productivity ofpublic capital (see footnote ii). In any event, no explicitdistinction between the conventional and neoclassicalviews is made in the text because they are comparable intheir implications (though certainly not with the full-employment assumptions).

PriceLevel

p2

p

Pt

Other things the same, the degree of crowding For deficits to really be considered infla-out will be larger, the more responsive interestrates are to deficit spending and the moreresponsive investment spending is to changes inthe interest rate. Thus, the larger the effect ofdeficits on interest rates, the smaller should bethe effect on current output. Even with thiscrowding out effect, however, deficits and out-put are still expected to be positively corre-lated—at least in the short run.”

— >“ .trl,.flLc.t’

11it11:ttFtDI/tN 511 T~flT5!TYi5

An alternative view of the effects of deficitspending is called Ricardian equivalence. Unllkethe Keynesian view, which sees public andprivate saving as essentially unrelated, theRicardian view sees them as perfect substitutes.”According to this view, changes in public savingare matched by an equal but opposite change inprivate saving. This fundamental differencemanifests itself in the answer to the question: Isgovernment debt part of society’s net wealth?

In the Keynesian view, the answer is “Yes,”while the Ricardian alternative would answer“No.” According to the Keynesian view, when

‘1Some people fear that deficits could have some longer-term adverse effects due to crowding out. They argue that,if deficits lead to higher interest rates and lower invest-ment, society will be left with a smaller capital stock,which, in turn, means that future output will be smallerthan it would have been in the absence of deficit-financedexpenditures. Consequently, in the short run, deficits andoutput are expected to be positively correlated. In the longrun, however, they may be negatively correlated, or atleast negatively correlated with output growth. There is animplicit assumption in this argument about the nature ofpublic expenditures or the relative productivity of “public”and “private” capital. To see this, note that the supply ofoutput, Y,, is a function of labor, L,, and the capital stock,K,, i.e., Y, f(L,, K,).

Output rises with both labor and the stock of capital.Consequently, if a deficit-induced rise in interest ratescauses investment to fall in this period, the next period’scapital stock will be smaller, as will the next period’s out-put. This assumes that all of the rise in the deficit was us-ed for current consumption. Alternatively, aggregate outputcould be expressed as a function of labor and the publicand private capital stocks, K”, and K”,, respectively, i.e.,V = h(L., K”,, K”,). If the rise in the deficit is used to ac-quire public capital, the immediate effect on output is in-determinant even in the case of complete crowding out,i.e., the decline in private capital from what it would haveotherwise been is equal to the increase in public capital.See Aschauer (1989a, 1989b).

‘2Deficits can sustain inflation if the monetary authority at-tempts to peg the nominal interest rate [see Friedman(1968)J or if deficits become explosive. In the latter case,the ratio of interest-bearing government debt to output in-creases without limit forcing the monetary authority tomonetize the debt. See Sargent and Wallace (1981) andMiller and Sargent (1984) for a discussion of these pointsand McCallum (1984) for a qualification.

“This is merely a convenient, equivalent characterization.Ricardian equivalence stems directly from an intertemporalresource constraint. That is, for a given path of expen-ditures, a cut in present taxes necessarily implies an in-crease in future taxes of equal present value. If an in-dividual’s demand for goods and services depends on his‘after tax” net worth—the present value of assets, thepresent value of liabilities and the present value oftaxes—a fall in current taxes would necessarily be match-ed by an increase in the present value of future taxes.Since a budget deficit merely rearranges the timing of thetax liability, it cannot affect aggregate demand. Thedecrease in public saving (the deficit) must be matched byan increase in private saving.

Strictly speaking, however, Ricardian equivalence holdsonly for a given path of government expenditures. Therecan be real effects associated with changing the level ortiming of real government expenditures.

29

Figure 2The Trade Deficit and the Government Deficit

Trade Delicil

the government issues debt, the holder of thedebt views it as an asset; but taxpayers do notview it as their liability. That is, they do notbelieve that they will have to pay current orfuture taxes to service or retire the debt. Conse-quently, their saving behavior is unaffected bythe debt issuance.

The Ricardian view, on the other hand, as-serts that individuals believe that they, or theirheirs, will have to pay taxes to service and re-tire the increased debt. Because they perceivean increase in the present value of their taxesthat just offsets deficit-financed expenditures,the stock of government debt is not part of so-ciety’s net wealth. Hence, a rise in the deficit (afall in public saving) will be matched by a risein private saving in anticipation of future taxes.

If public and private saving are perfect sub-stitutes, a decline in public saving—the increasein the deficit—is offset by an increase in privatesaving, with no effect on the gap between sav-ing and investment. According to this view,then, deficit spending cannot be used to smoothout cyclical variations in economic activity dueto exogenous swings in the gap between savingand investment. Moreover, the increased de-mand for credit due to deficit spending willbe matched by an increase in the supply ofcredit due to an increase in private saving. Also,the increase in aggregate demand that results

from a rise in deficit spending will be offset bya decrease in aggregate demand due to the fallin private spending—that is, there is no netchange in aggregate demand.

Because the Ricardian view holds that changesin deficit have no net effect on the excess de-mand for credit or aggregate demand, deficitsshould be uncorrelated with interest rates, thetrade deficit, the price level, output or total sav-ing. Deficits and private saving, however, shouldbe perfectly, positively correlated.

Ehl/II3.IIIICAL EVIIJIENCE

The evidence presented here consists of sim-ple scatter plots between the government deficitand the trade deficit, personal saving, real out-put growth, the price level, the inflation rateand the nominal interest rate for 16 OECDcountries.’4 (A more detailed statistical analysisconsistent with these scatter plots is presentedin the appendix). The currency-denominatedvariables are expressed as a percent of thecountry’s gross domestic product, GDP, to putthem in common units, and all variables are ad-justed for their mean values.’~

Figure 2 shows the relationship between theso-called twin deficits: the trade and the govern-ment deficits. The scatter plots show that there

‘4The countries are: Australia, Austria, Belgium, Canada,Finland, France, Greece, Great Britain, West Germany,Ireland, Japan, Norway, Netherlands, Sweden, Switzerlandand the United States. The interest rate was not availablefor Austria, Canada, Finland or Greece.

15These are pooled time-series cross-sectional data, wherethe average level is allowed to vary each year. See the ap-pendix for details.

10

—15 —It —9 —5 —3 0 3 5Governmenl Delicil

15

10

0

—5

—10

is a weak (but statistically significant) positiveassociation between the two deficits. Thisassociation is consistent with both the conven-tional view and U.S. time-series data.’6 Barro(1987), however, finds that the positive associa-tion exists for U.S. time-series data only whendata for the 1980s are included- This instabilityis evident for these cross-sectional data as well.When the sample includes only observations forthe earlier period, 1975-80, there is no signifi-cant positive relationship between the twindeficits. These data are presented in figures 3,while data for the later period, 1981-86, arepresented in figure 4. Moreover, further analy-sis suggests that the statistically significant posi-tive relationship does not hold up if other factorsare considered (see the appendix).

Figure 5 shows the relationship between per-sonal saving and the government deficit. Thestrictest form of the conventional view arguesthat these series should be unrelated, while theRicardian alternative argues that they should beperfectly, positively correlated. Neither view issupported by these cross-sectional data. Instead,there is a statistically significant negative rela-tionship between these variables. This anomalousresult may stem from the response of both thedeficit and personal saving relative to GDP tocyclical movements in output. For example, it iswell-known that deficits typically rise relative toGDP as output falls and vice versa. If consumersattempt to maintain their living standard in theface of a temporary decline in output, personalsaving would fall relative to output. Consequent-ly, personal saving and government deficits rela-tive to GDP should be negatively correlated overthe business cycle. In any event, the statisticallysignificant negative relationship supports neitherhypothesis.’~

Figures 6-9 show the relationship between thegovernment deficits and inflation, output growth,the price level and interest rates. In all cases,

there is no statistically significant relationshipbetween the deficit and these variables.’~Conse-quently, the cross-sectional data provides littleor no support for the conventional view. On theother hand, support for the Ricardian alternativeis not overwhelming. While the lack of signifi-cant relationships between the deficit and in-terest rates, the trade deficit, output, prices,inflation and economic growth is consistent withthe Ricardian view, the lack of a statisticallysignificant positive relationship between thegovernment deficit and private saving is not. In-deed, if deficits and private saving are perfectly,positively correlated because public and privatesaving are perfect substitutes, a positive rela-tionship between these variables should havebeen apparent. Thus, at best, the evidence pre-sented here should be interpreted as cautiouslyfavoring the Ricardian view.

Whether the evidence presented here is takento support, however cautiously, the Ricardianalternative, the most interesting result is thelack of evidence to support the conventionalview. In particular, the evidence suggests thatcountries with large government deficits do nothave higher interest rates or larger trade defi-cits despite the widespread opinion to the con-trary. Of course, these puzzling results could bean artifact of the simple measures of deficitspending used here. If the relationships betweengovernment deficits and either interest rates orthe trade deficit were as strong as many com-mentators suggest, however, it is odd that theydo not emerge in these annual data across coun-tries. These results, coupled with the fact thatmore sophisticated empirical studies using U.S.time-series data have also failed to uncover theconventional relationships, should perhaps leadadvocates of the conventional view either to re-think their position or present some evidence tosupport their claims.

155ee Barro (1987) and Dewald and Ulan (1990). Dewaldand Ulan argue that the observed positive association bet-ween the twin deficits in U.S. time-series data results fromfailing to account for inflation or changes in the marketvalue of the U.S. federal debt or other major elements ofthe U.S. net external wealth position. This argument,however, requires a different definition of the deficit thanthe one used in the conventional stories of the effects ofdeficits on economic variables.

17Recently, Swamy, Kolluri and Singamsetti (1990) have sug-gested that finding any statistically significant relationshipbetween the deficit and interest rates—although the signof the coefficient is opposite that hypothesized by the con-

ventional view—contradicts the Ricardian equivalenceparadigm. Such an extreme view is unwarranted.

18There are two exceptions when first-differences are used.The one of some importance for the conventional wisdomis the positive and statistically significant relationship be-tween the deficit and output growth, suggesting that largergovernment deficits have at least some initial positive ef-fect on real output. Of course, if the rise in the deficit isdue to an increase in government spending, there is apositive relationship between the deficit and GDP bydefinition. See the appendix for details.

111’ .919 .1’ 14 444 I $9 94 44 WI

1.C

D~

1940

944

g~

3•

~L°’

I’u

~I P $~

C’9 44

a44. 4’

—‘4

‘0

~.Ill

’o

111C)

P~

Ir~

P31

o 9.9!

oh’

~ 2.4%

—.44 I i’ll 94 (4. .1 ii I till 44 •4444 ‘94 :41,

I (91 I 1491 ‘44 44 4)4 dl P3 IN-1

1i4

44

,0

(DC

49a

-l~

944 490

eit

49~.

941Z

944 441,~

I44

40’

~44’4

o‘4

o$~

‘0

~in

—is

~ od

l:~

o41 Vt

09,

-44

(0dl

0 ~a

.—4$

(044.

044°

091 144 99 II,

a4%’ 441 4444 444 $14’ 414% $141’

91 :1 .441 1944 44i4~ 4$) II N—

I:ii

1411:r

~~

4*4:

oC

4141

—Is

444:~

4j:o

49CD .~

44C

’444

; ~441

~44,4

CD4144

CD$4

—414 1144

43

:CD

441 4114—

141

o1:

CD‘144 4, ‘1 ‘cdi 14~ 4’44

Figure 6The Rate of Inflation and the Government DeficitRote of notation Rate of Inflation

It It

to ID

o 5

o 0

5 5

— v “‘ 10—It —42 —9 —6 —2 0 3 6

Government Deficit

Figure 7The Growth Rate of Real GDP and the Government Deficit

Figure 8The Price Level and the Government DeficitPnce Level

it

10

05

00

Price Level

LI

1.0

0.5

0.0

-0.5—it —14 —11 —It —t —6 4 —2 0 2 4 6 6 It

Goterilment Deficit

—It —52 —9 —6 —3Government Deficit

3-3

Figure 9The Interest Rate and the Government Deficitittereot Rete inferett itetet

is to

5 0

0 0

—Ia ‘ if—15 42 —9 6 3 0 3 9

Government Detc,t

144)’ FEB EN{1~ES

Aschauer, David. “The Equilibrium Approach to Fiscal Revisited: Comment,’ American Economic Review (JunePolicy,” Journal of Money Credit and Banking (February 1990), pp. 589-99.1988), pp. 41-62. Friedman, Milton. “The Role of Monetary Policy:’ American_______ “Does Public Capital Crowd Out Private Capital?” Economic Review (March 1968), pp. 1-17.Journal of Monetary Economics (September 1989a), PP Koray, Faik, and R. Carter Hill. “Money, Debt and Economic171-88. Activity:’ Journal of Macroeconomics (Summer 1988), pp.

_______- “Is Public Expenditure Productive?” Journal of 351-70.Monetary Economics (March 1989b), 177-200. Kormendi, Roger, and Philip Meguire. “Government Debt,

Government Spending, and Private Sector Behavior: ReplyBarro, Robert J. The Ricardian Approach to Budget and Update[ American Economic Review (June 1990), pp.Deficits, Henry Thornton Lecture, City University Business 604-17School, London, November 1987.

Laumas, Gurcharan S. “Anticipated Federal Budget Deficits,Bernheim, B. Douglas. “Ricardian Equivalence: An Evalua- Monetary Policy and the Rate of Interest:’ Southern

tion of Theory and Evidence:’ NBER Working Paper No. Economic Journal (October 1989), pp. 375-82.2330 (July 1987). ‘Leiderman, Leonardo, and Assaf Razin. Testing Ricardian

_______- “A Neoclassical Perspective on Budget Deficits,” Neutrality with an Intertemporal Stochastic Model” JournalJournal of Economic Perspectives (Spring 1989), pp. 55-72. of Money, Credit and Banking (February 1988), pp. 1-21.

Chrystal, K. Alec, and Daniel L. Thornton. “The McCallum, Bennett T. “Are Bond-Financed Deficits Infla-Macroeconomic Effects of Deficit Spending: A Review,” this tionary? A Ricardian Analysis:’ Journal of Political EconomyReview (November/December 1988), pp. 48-60. (February 1984), pp. 123-35.

Darrat, Au F. “Fiscal Deficits and Long-Term Interest Rates: Miller; Presten J., and Thomas J. Sargent. ‘A Reply to Dar:Further Evidence from Annual Data:’ Southern Economic by, Federal Reserve Bank of Minneapolis Quarterly ReviewJournal (October 1969), pp. 363-74. (Spring 1984), pp. 21-26.

Modigliani, Franco, and Arlie G. Sterling. “GovernmentDewald, William G., and Michael Ulan. “The Twin-Deficit Illu- Debt, Government Spending, and Private Sector Behavior:

sion[ Working Paper No. 90-17, U.S. Department of State, A Further Comment:’ American Economic Review (JuneBureau of Economic and Business Affairs (February 1990). 1990), ~. 600-03.

Evans, Paul. “Are Consumers Ricardian? Evidence for the Sargent, Thomas J., and Neil Wallace. “Some UnpleasantUnited States,” Journal of Political Economy (October 1988), Monetarist Arithmetic:’ Federal Reserve Bank of Mm-pp. 983-1004. neapolis Quarterly Review (Fall 1981), pp. 1-17.

_______- “Interest Rates and Expected Future Budget Swamy, Paravastur A. V. B., Bharat R. Kolluri and Rao N.Deficits in the United States:’ Journal of Political Economy Singamsetti. “What Do Regressions of Interest Rates on(February 1987), pp. 34-58. Deficits Imply?” Southern Economic Journal (April 1990),

_______- “Is the Dollar High Because of Large Budget pp. 1010-28.Deficits?” Journal of Monetary Economics (November Tatom, John A. “U.S. Investment in the 1980s: The Real19B6), pp. 227-49. Story,” this Review (March/April 1989), pp. 3-15.

Feldstein, Martin. and Douglas W. Elmendorf. “Government Thornton, Daniel L. “Monetizing the Debt,” this ReviewDebt, Government Spending, and Private Sector Behavior (December 1984), pp. 30-43.

34

4/ 44-‘ - /4

1/

/ 4/ ~/4 / -

- ‘V , ‘,~ - ~/V ~

The purpose of this appendix is to see if thereis a statistical association between governmentdeficits and some important macroeconomicvariables as hypothesized by both the conven-tional view and the Ricardian alternative.

The analysis begins with the following generalequation:

(A1) DV, a,, + j35DEF,, + E~,, i=1,..., K, and

where DV,, and DEF,, denote the t” observationfor the P” country of the dependent variableand the deficit measure, respectively, a,,, and /3,,denote fixed parameters and s,, denotes a ran-dom error.

Equation A.1 cannot be estimated because thenumber of parameters exceeds the number ofobservations. This problem can be circumventedby obtaining time-series and/or cross-sectionrepresentations of equation Al. The time-seriesrepresentation is obtained by imposing therestrictions a5 = a, and /3,, /3, for all t. Thecross-sectional representation is obtained by im-posing the restrictions a,, = a0 and /3,, = /3, forall i. These specifications are:

(A.2) DV,, = a, + /3,DEF,I + ~,0

and

(A.3) Dy,0 = a0 + J3,DEF,, + E,~.

A pooled time-series/cross-section representa-tion can be obtained by imposing the restric-

tions a,, = a and /3,, = /3 for all i and t, toobtain

(A.4) DV,, = a + /3DEF,, +

(This is equivalent to imposing the restrictionsa, = a and /3, = /3 for all i on the time-seriesmodel or a, = a and /3, = /3 for all t on thecross-sectional model).

Equations A.2-A,4 were estimated with annualobservations on the government deficits, nomi-nal interest rate, the trade deficit, the pricelevel (1980 = 100), the inflation rate, real outputgrowth and private saving for 16 OECD coun-tries (K=16) for which the relevant annual dataare available.1 The period is from 1975 to 1986(T= 12). Because the currency-denominatedvariables are expressed in the respective coun-try’s currency, it is necessary to put these datain common units. This was done by measuringthese variables as a percent of gross domesticproduct, GDP.2

The equations were estimated both in levelsand first-differences. The latter form wasestimated because some of the data showed atendency to trend with time. The sample wastoo small, however, to perform the usual testsfor stationarity. Estimates of first-order autocor-relation from equation A,2 suggest that first-differencing may result in over-differencing inmany cases. Because some of these data ap-peared to have trends, the pooled time-series!cross-section equation, A.4, was estimated byallowing the intercept term to vary with time.3

Because the estimates of equations A.2 and A.3

4The interest rate is the three-month money market rate.2This was also done by measuring these variables relativeto their mean value for the period and indexing them to1975. When the variables were normalized in this way, ameasure of the level of GDP was also used as a depen-dent variable. In nearly all cases, estimates of equationsA.2 and A.3 resulted in insignificant coefficients on thedeficit measure, so the results are not reported here.

Also, the OECD data are based on the system of na-tional accounts which differs in a number of respects fromour system of national income and product accounts. Oneof these is that capital expenditures by government arenot treated as current expenditures, so that the deficit iscurrent expenditures plus consumption of fixed capital less

~Thereported intercepts in tables A.1 and A.2 are for 1986.The equations were also estimated using a time trend. Theresults with this variable were not qualitatively differentfrom those that allowed for a time-varying intercept. Con-sequently, they are not reported here. Also, the equationswere estimated using the alternative dependent variablesas regressors. Except for the interest rate, however, thequalitative results were unaffected, so these results arenot presented here.

t=1,..., T,

revenue.

are qualitatively the same as those of equationA.4, only estimates of equation A.4 arepresented.

.44444 fl.44-$-/U44-~

Estimates of equation A.4 for both levels andfirst-differences are reported in table Al.4 Theonly coefficients that are statistically significantin both the level and first-difference specifica-tions occur when the trade deficit or privatesaving is the dependent variable. In the lattercase, however, the coefficient was negative,which suggests that increases in deficit spendingare associated with decreases in personal sav-ing: when public saving decreases, so doesprivate saving. This result is not consistent withthe Ricardian view that public and private sav-ing are substitutes; however, it might reflect anendogenous response of both governmentdeficits and personal saving to cyclical variationsin output. It is well-known that deficits tend torise relative to GDP when output falls. Likewiseit might be that personal saving also rises whenoutput falls, as individuals try to maintain theirlevel of consumption in the wake of decliningincome levels. In this case, personal savingwould decline relative to GDP. The simultaneousresponse of both government deficits and per-sonal saving to cyclical variation in incomecould account for the statistically significantnegative coefficient on the deficit when per-sonal saving is the dependent variable.

Barro (1987) also found a positive and statisti-cally significant relationship between the twodeficits using U.S. data, but stated that this re-lationship emerges only when data for the 1980swere included. When the sample was partitionedinto 1975-80 and 1981-86 periods, a positiverelationship emerged between the twin deficitsduring both periods; however, consistent withthe findings of Barro, the relationship was notstatistically significant during the first period.~

Dewald and Ulan (1990) argue that the posi-tive relationship between the twin deficits inthe U.S. results from failing to account for infla-tion or changes in the market value of the netU.S. federal debt or other major elements of the

U.S. external wealth position. When adjustmentsfor these factors are made, they find no statisti-cally significant relationship between the twodeficits. Because this argument might apply tocross-sectional data as well, caution should beused in interpreting these results as evidence ofthe conventional wisdom on the twin deficits.This is especially true because the estimatedcoefficient on the deficit is not large: a 1percentage-point rise in the government deficitrelative to GDP would result in only about a .20percentage point rise in the trade deficit relativeto GDP. Moreover, as shown later, the result isnot robust.

The statistical significance of the coefficienton the deficit when inflation or the price levelis the dependent variable depends on whetherthe equation is estimated in levels or first-differences, being insignificant in the formercase and significant in the latter. In either case,however, the coefficient has a sign opposite thatof the conventional story. Consequently, theseresults are not evidence for or against either ofthese views.

The same result occurs when the growth rateof GDP is the dependent variable. In this case,however, the conventional wisdom does notstate the direction of the expected change. Froma short-run perspective, output growth shouldbe positively related to deficits. From a long-runperspective, they should be negatively related.The results in table A.1 suggest a positive rela-tionship if first differences are used, with a 1percentage-point increase in the deficit beingassociated with a quarter of a percent increasein the rate of growth of GUP. Of course, if theincrease in the deficit is due to an increase ingovernment spending, there is a positive rela-tionship between deficits and GDP by definition,if crowding out is not complete.

Estimates of the effect of the deficit on in-terest rates also are sensitive to the specificationof the equation. When levels are used, there isa negative, but statistically insignificant, relation-ship between interest rates and the deficit.When first-differences are used, however, thereis a positive and statistically significant (a one-

~ltwas assumed that variance was constant over time, butdiffered over cross-sections, that is, E(r./)= c’, for i=jand t=s and 0 otherwise. Estimates of equation A.2 in-dicated considerable cross-sectional heteroskedasticity.

5The estimated coefficient on the deficit variable during thefirst period for the level specification was .181, with at-

statistic of 1.42. The comparable statistics for the secondperiod are .307 and 3.52, respectively. The same statisticsfor the two periods for the first-difference specification are.094 and 1.37 and .226 and 3.01, respectively.

Table A.1Pooled Time-Series, Cross-Sectional Estimates of Equation A.4

Levels First difference

Dependent Deficit Deficit

variab constant - measure .9?~!!,nt - measure

Trade deficit — 1.396’ 0.195’ -0.453 0127’

(2.43) (2.42) (1.43) (2.26)

Personal saving 9.163’ —0.44V 0.197 —0.170’

(14 25) (6.59) (0.77) (3.43)

Output growth 2.34r —0.060 —1.206’ 0.244

~5.67~ ft34~ ~287~ ~3.73~

Inflation 4.444’ —0.061 0.327 —0.154’(5.76) (0.77) (1.70) (4.40)

Price level 1 517’ —0.002 0.07 —0.004’(75.44) (1.17) (12.11) (3.58)

Interest rate 8.120 —0.020 —0.886 0.246’(10.89) (0.28) (1.96) (1.95)

- Indicates statistical significance at the 5 percent level.

tailed test at the 5 percent significance level) attributable to two countries, Greece and Nor-relationship between the two variables. Even if way. Also, the reported results could be due toone takes the results that are most supportive the failure to account for other factors that af-of the conventional wisdom, however, the effect fect the relevant dependent variable. Both issuesof deficits on interest rates is fairly weak: a 1 are investigated in table AZ, which reports thepercentage-point rise in the deficit relative to results of estimates of the first-difference speci-GDP is estimated to produce about a 25 basis- fication when Greece and Norway are deletedpoint rise in nominal interest rates. and, alternatively, when all other variables are

included as regressors.6 As the table shows, the

results are sensitive to both of these changes.- 4 ~ Deficits are not significantly related to the price

While the results for the first-difference specifi- level or interest rates when the two countriescation appear to indicate statistically significant are deleted and no longer have a statisticallyrelationships between the government deficit and significant effect on the trade deficit, the priceall of the dependent variables, these relationships level or the interest rate when the other van-are not as strong as the results in table A.1 would ables are included. The only significant effectssuggest. Scatter plots of the first-differenced data, that are robust to these changes are those formean-adjusted, are presented in figures A.1-A.6. the effect of the deficit on private saving, out-These data suggest that much of the reported put growth and inflation. In the first and laststatistically significant relationship could be due to case, however, the sign of the coefficient is op-a relatively few outliers. In most cases, these are posite that suggested by the conventional view.

CAll other variables save the interest rate, since the interestrate was unavailable for Austria, Canada, Finland andGreece.

Figure AlChange in The Trade Deficit and the Government DeficitTrade Dei,c,t Trade Def,dt

Ia - - - ‘0

O - - - - — 0

4- .4- - -

a -- - -- - - - 0

-4 - 4 4 4 — -4—04\ - - - - --4 - - - -4---- —5

4 -4 -4 — -

10 4 —- -- —10

—30 —27 —IS —0.2 0.0 21 33 4.5 57

Government Dcccii

-4/4 / /4 - /4 - / ‘V / ~ / 4V/

Figure A2Change in Personal Saving and the Government DeficitPertonti Savine Pertonsi Se,tng

6 a

4/ ~-4/4 -- ~ 1V~-, ‘V’V ~ —-~—~.

4 — — 4- ___/ — — — 4

‘V -~- ~ -4~A-444 / 4-4/- /4/

2 ‘V - ~-V/4 ~\45 /4 -~- 4-4444 2V - - - ‘/4- 4- - - - 1~-

0 - - - 4 - - -- - 0

-- 44 — ,, 4-4 ----4 -4- —- 44 - ‘44 - - - -4 - - - - -

- - 44 - / / - -4—2/ -4- -\ - 4/-

-4-—4 - - _ / 4/ 4 / - - - —4

- - - - — -4—e - 4 - 44 - - -4 —e44

—a - - —0—30 —27 IS —a oe 21 32 45 57

Government Oet’dt

- ~~—‘V/ --4 ~/4 / / - / / -/4/4 / _/// /4/4

Figure ASChange in the Inflation Rate and the Government DeficitRare of In4tatinn Rate or tnnat’on

O — — 0

o .4- 4, 0

0 * —5

—10 —10

1S

—20 20—20 —27 —IS —03 09 21 33 4.5 07

Government Deficit

- /4 ~ -4/4/44 -4/4 //~// / — - - / / ---/4 -/44/ - / /

(:4/44U:4>40

4/4/44//444/4/

C

:0

a.44/4

4(0

1444—

4/;0

U~

l~~1-U

2-4

/:It

UH2°

/4~-

a’E

oE

44C

0

Lt~

11.4/

-44a

-r

—

/*

//

44/

*-44

044

~/

/

440

/0

/

44—

/

C*

/

4190E

//

C

-4/I-0

-a>

/

0/

/

0/

//

V/

*

1/

/*

I-~

//

//

,4

lJs~-

0/

*41,

‘4.9

/4

144’/

/--4

-

4a

/.4

,4-

2:

/

‘1944/—

//

44C

//

//

/4

0/~

//4

-a

4‘74/

~/

-/

4/19~

,,-a

--4-4

4,

LL

0~

I44-i4

Vi14/0

44440

44:4/—

V-4~V0

144E

N

~‘74

44

£0

1-4/4.lV44~.—

14/,~

44/-:

44/-LLO

U‘/44/—NUN44/il:11414111191494-(119

am

Table A. 2Estimates of Two Alternative First-Difference Specifications

Greece and All alternative dependentNorway deleted variables included

Dependent Deficit Deticitvariable constant variable constant variable

Trade deficit —0611 e 0 128’~ 0.072 0.013

(2.03) (222) (0 21j (0.23)Personal savng 0.312 —0.143 0400 —0.213’

(1 28) (2 97) (1 63) (4.55)

Output growth —1 .056~ 0342’ — 0.930’ 0.330(2.31) (4.01) (226) (4.66)

Inflation 0378 —0126’ 0112 —0093(1.94) (3 65) (0.54) (2.63)

Price level 0.051 - —0.002 0.062 —0.001(7.91) (1.73) (6.81) (095)

Interest rate —1.093’ 0 128 — 0.934 0.090(2.29) (0.78) (1 95) (0.61)

indicates statistical significance at the 5 percent level