Large Changes in Fiscal Policy: Taxes Versus Spending

NBER WORKING PAPER SERIES

LARGE CHANGES IN FISCAL POLICY:

TAXES VERSUS SPENDING

Alberto F. Alesina

Silvia Ardagna

Working Paper 15438

http://www.nber.org/papers/w15438

NATIONAL BUREAU OF ECONOMIC RESEARCH

1050 Massachusetts Avenue

Cambridge, MA 02138

October 2009

Prepared for Tax Policy and the Economy 2009. We thank Jeffrey Brown, Roberto Perotti, Matthew

Shapiro and other conference participants for useful comments and discussions. The views expressed

herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of

Economic Research.

NBER working papers are circulated for discussion and comment purposes. They have not been peer-

reviewed or been subject to the review by the NBER Board of Directors that accompanies official

NBER publications.

exceed two paragraphs, may be quoted without explicit permission provided that full credit, including

Large Changes in Fiscal Policy: Taxes Versus Spending

Alberto F. Alesina and Silvia Ardagna

NBER Working Paper No. 15438

October 2009, Revised January 2010

JEL No. H2,H3

ABSTRACT

We examine the evidence on episodes of large stances in fiscal policy, both in cases of fiscal stimuli

and in that of fiscal adjustments in OECD countries from 1970 to 2007. Fiscal stimuli based upon

tax cuts are more likely to increase growth than those based upon spending increases. As for fiscal

adjustments, those based upon spending cuts and no tax increases are more likely to reduce deficits

and debt over GDP ratios than those based upon tax increases. In addition, adjustments on the spending

side rather than on the tax side are less likely to create recessions. We confirm these results with simple

regression analysis.

Alberto F. Alesina

Department of Economics

Harvard University

Littauer Center 210

Cambridge, MA 02138

and NBER

[email protected]

Silvia Ardagna

Department of Economics

Harvard University

Littauer Center

Cambridge, MA 02138

[email protected]

1 Intr o duction

As a result of the ﬁscal response to the ﬁnancial crisis of 2007-2009 the US will

experience the largest increases in deﬁcits and debt accumulation in peacetime.

Virtually all other OECD countries will also face ﬁscal imbalances of various sizes.

After the large reduction in government deﬁcits of the nineties and early new cen-

tury, public ﬁnances in the OECD are back in the deep red.

Only a few months ago the ke y policy question was whether tax cuts or spend-

ing increases were a better recipe for the stimulus plan in the US and other countries

as well. By and large these decisions have been taken, and w e are in the process

of observing the results. The next question which governments all over the world

will face next year, assuming, as it seems likely, that a recovery next year will be

under way, is how to stop the gro wth of debt and return to more “normal” public

ﬁnances.

The ﬁrst question, namely whether tax cuts or spending increases are more

expansionary is a critical one, and economists strongly disagree about the answer.

It is fair to say that we know relatively little about the ef fect of ﬁscal policy on

growth and in particular about the so called ﬁscal multipliers, namely how much

one dollar of tax cuts or spending increases translates in terms of GDP. The issue is

very politically charged as well, since right of center economists and policymakers

believe in tax cuts and the left of center ones believe in spending increases. While

the differences are often rooted in different views about the role of government and

inequality, not so much about the size of ﬁscal multipliers, both sides also wish

to "sell" their prescription as growth enhancing and more so than the other policy.

Unfortunately both sides can’t be right at the same time!

As far as reduction of large public debts the lesson from history is reasonably

optimistic. Large debt/GDP ratios have been cut relatively rapidly by sustained

growth. This was the case of post WWII public debts in belligerent countries; it

was also the case of the US in the nineties when without virtually any increase in

tax rates or signiﬁcant spending cuts, a large deﬁcit turned in a large surplus.

the UK the debt over GDP ratio at the end of WWII was over 200 percent but that

country did not suffer a ﬁnancial crisis due to its historically credible ﬁscal stance

and the debt was gradually and relatively rapidly reduced. However, it would be

probably too optimistic to expect another decade like the nineties ahead of us;

that kind of sustained growth would certainly do a lot to reduce the debt/GDP

ratio but the lower growth which we will most likely experience will do much

less. Inﬂation also has the effect of chipping away the real value of the debt but

it may be a medicine worse than the disease. While a period of controlled and

See Alesina (1998) for a discussion of the budget surplus in the nineties in the US.

moderate inﬂation would have the potential to reduce the real value of outstanding

debt, pursuing such a strategy would run the risk of uncontrolled inﬂation. It took a

sharp recession in the early eighties to eliminate the great inﬂation of the seventies,

and the last thing we need is another major recession in the medium run. The post

WWI hyperinﬂations are certainly not in the horizon, but we should keep them in

the back of our mind as an extreme case of debt induced runaway inﬂation.

If growth alone cannot do it and inﬂating away the debt carries substantial risks,

we are left with the accumulation of budget surpluses to reign in the debt in the next

several years in the post crisis era. But then the same question returns: is raising

taxes or cutting spending more likely to result in a stable ﬁscal outlook?

This is precisely what this papers is about. We focus upon large changes in

ﬁscal policy stance, namely large increase or reduction of budget deﬁcits and we

look at what effects they had on both the economy and the dynamics of the debt.

In particular, for the case of b udget expansions (increase in deﬁcits or reduction of

surpluses) we look at which have been more expansionary on growth. On ﬁscal

adjustments (deﬁcit reductions) we consider their effect on a medium term stabi-

lization/reduction of the debt over GDP level and their cost in terms of a downturn

in the economy. We focus only on large ﬁscal changes because we try to isolate

changes in ﬁscal policy which are policy induced as opposed to cyclical ﬂuctua-

tions of the deﬁcits, which in any event we try to cyclically adjust. Our method-

ology is rather simple. We identify episodes of large changes in ﬁscal policy. Ob-

viously the decision of when to engage is such policy changes is not exogenous to

the state of public ﬁnances and of the economy. But up to a point the decision of

whether to act upon the spending side or the revenue side is largely political and

due to bargaining amongst political and pressure groups. The uncertainty about

the size of ﬁscal multipliers make this discussion even less constrained by solid

economic arguments. Thus we cannot offer new measures of ﬁscal multipliers, but

we can look at what effects have different approaches (spending versus revenue

side) have had during and after large ﬁscal changes.

Our results suggest that tax cuts are more expansionary than spending increases

in the cases of a ﬁscal stimulus. For ﬁscal adjustments we show that spending cuts

are much more effective than tax increases in stabilizing the debt and avoiding

economic downturns. In fact, we uncover several episodes in which spending cuts

adopted to reduce deﬁcits have been associated with economic expansions rather

than recessions. We also investigate which components of taxes and spending af-

fect the economy more in these large episodes and we try uncov er channels running

through private consumption and/or investment.

The present paper is more directly related to several ones written in the early

nineties using a similar approach to ours. Giavazzi and Pagano (1990) were the ﬁrst

to argue that ﬁscal adjustments (deﬁcit reductions) large, decisive and on the spend-

ing side could be expansionary. This was the case of Ireland and Denmark in the

eighties which were the episodes studied by Giavazzi and Pagano (1990), but there

were others as then discussed and analyzed by Alesina and Ardagna (1998). The

same authors and Alesina and Perotti (1997) investigate various episodes of ﬁscal

adjustments reaching conclusions similar to that of the present paper. But in this

paper we hav e many more episodes and we use more compelling techniques. There

is quite a rich literature that studies the determinants and economic outcomes of

large ﬁscal adjustments. A non exhaustive-list includes Ardagna (2004), Giavazzi,

Jappelli and Pagano (2000), Huges and McAdam (1999), Lambertini and Tavares

(2000), McDermott and Wescott (1996), Von Hagen and Strauch (2001), Von Ha-

gen, Hughes, and Strauch (2002), and more recently, OECD (2008) and IMF

(2009). Theoretically, expansionary effects of ﬁscal adjustments can go through

both the demand and the supply side. On the demand side, a ﬁscal adjustment

may be expansionary if agents believe that the ﬁscal tightening generates a change

in regime that “eliminates the need for larger, maybe much more disruptive ad-

justments in the future” (Blanchard (1990)).

Current increases in taxes and/or

spending cuts perceived as permanent, by removing the danger of sharper and more

costly ﬁscal adjustments in the future, generate a positive wealth effect. Consumers

anticipate a permanent increase in their lifetime disposable income and this may

induce an increase in current private consumption and in aggregate demand. The

size of the increase in private consumption would depend, however, on the presence

or absence of “liquidity constrained” consumers. An additional channel through

which current ﬁscal policy can inﬂuence the economy via its effect on agents’ ex-

pectations is the interest rate. If agents believe that the stabilization is credible and

avoids a default on government debt, they can ask for a lower premium on gov-

ernment bonds. Private demand components sensitive to the real interest rate can

increase if the reduction in the interest rate paid on government bonds leads to a

reduction in the real interest rate charged to consumers and ﬁrms. The decrease

in interest rate can also lead to the appreciation of stocks and bonds, increasing

agents’ ﬁnancial wealth, and triggering a consumption/investment boom.

On the supply side, expansionary effects of ﬁscal adjustments work via the la-

bor market and via the effect that tax increases and/or spending cuts have on the

indi vidual labor supply in a neoclassical model, and on the unions’ fall-back posi-

tion in imperfectly competitive labor markets (see Alesina and Ardagna (1998) and

Alesina et al. (2000) for a review of the literature). In the latter context, the compo-

sition of current ﬁscal policy (whether the deﬁcit reduction is achieved through tax

increases or through spending cuts) is critical for its effect on the economy. On the

one hand, a decrease in government employment reduces the probability of ﬁnding

For models that highlight this channel, see Bertola and Drazen (1993) and Sutherland (1997).

a job if not employed in the private sector, and a decrease in government wages

decreases the worker’s income if employed in the public sector. In both cases, the

reservation utility of the union members goes do wn and the wage demanded by

the union for private sector workers decreases, increasing proﬁts, investment and

competitiveness. On the other hand, an increase in income taxes or social secu-

rity contributions that reduces the net wage of the worker leads to an increase in

the pre-tax real wage faced by the employer, squeezing proﬁts,investment,and

competitiveness.

This is not the place to review in detail the large literature on the effect of ﬁs-

cal policy on the economy. It is worth mentioning that Romer and Romer (2007)

also follow an event approach even though they identify events of large discre-

tionary changes in ﬁscal policy in a very different way from ours. Using a variety

of narrative sources, they identify changes in the US federal tax legislation that

are undertaken either to solve an inherited budget deﬁcit problem or to achieve

long-run goals and estimate the effect of such changes on real output in a VAR

framew ork. They ﬁnd that an increase in taxation by 1% of GDP reduces output in

the next three years by a maximum of about 3% and that the effect is smaller when

the only changes in taxes considered are those taken to reduce past budget deﬁcits.

As Romer and Romer (2007), we also ﬁnd that tax increases are contractionary, but

the magnitudes of our results are difﬁcult to compare to theirs. In our estimates, we

ﬁnd that a 1% increase in the cyclically adjusted tax re venue decreases real growth

by less than one-third of a percentage point. However, we estimate a very different

speciﬁcation and, contrary to Romer and Romer (2007), our approach also controls

for changes in government spending undertaken to reduce budget deﬁcits as well

as for changes in taxation.

Blanchard and Perotti (2002) use structurally VAR techniques to identify ex-

ogenous changes in ﬁscal policy and estimate ﬁscal multipliers both on the tax and

on the spending side of the government. They ﬁnd that positive government spend-

ing shocks increase output, consumption and decrease investment, while positive

tax shocks hav e a negative effect on output, consumption and investment. Mount-

ford and Uhlig (2008) use a very different identiﬁcation approach and, while they

also ﬁnd that both taxes and spending increases have a negative effect on private

investment (as previously shown by Alesina et al. (2002)), they show that spending

increases do not generate an increase in consumption and that deﬁcit-ﬁnanced tax

cuts are the most effective way to stimulate the economy. The result of a positive

effect of government spending shocks on private consumption is also challenged

by Ramey (2008). She ﬁnds that, capturing the timing of the news about govern-

ment spending increases with a narrative approach and not with delay as in a VAR

approach, consumption declines after increases in government spending. Our re-

sults on the negative correlation between both spending and tax increases on GDP

growth are clearly consistent with the results of these papers using quite different

methodological approaches than ours.

A substantial literature has investigated political and institutional effects on

ﬁscal policy and in particular on the propensity of different parties in different

institutional settings to prolong ﬁscal imbalances, or to reign them in promptly. On

delayed ﬁscal adjustments see Alesina and Drazen (1999), on politico institutional

effects, like the role of electoral laws, on the occurrence of loose or tight ﬁscal

policy see Persson and Tabellini (2003) and Milesi Ferretti, Perotti and Rostagno

(2002). Alesina Perotti and Tavares (1998) using an approach similar to that of

the present paper and based upon "episodes", investigate which parties are more

or less likely to run in ﬁscal stimuli or ﬁscal adjustments. One criticism that one

could raise to the literature on voting rules and institutions on ﬁscal imbalances is

that rules are not exogenous and third factors may indeed explain both the adoption

of certain voting rule (like proportional representation) and ﬁscal policy, a point

discussed in Alesina and Glaeser (2004) informally and Aghion Alesina and Trebbi

(2007) more formally. We do not pursue in the present paper this politico economic

analysis.

This paper is organized as follows. Section 2 discusses our data and the deﬁn-

ition of episode which we adopt. Section 3 presents basis statistics on the episodes

show ing rather striking results. Section 4 shows some regression analysis, which

although it has no pretence of having solved causality problems reinforces the re-

sults obtained by the simple statistics of Section 3. The last section concludes.

2 Data, Methodology and deﬁnitions

2.1 Methodology

Our approach is very simple. We identify major changes in ﬁscal policy, either

expansionary (deﬁcit increases or surplus reductions) or the opposite. Obviously

the decision about whether to engage in this policy changes is endogenous to the

state of the economy and of the ﬁnances Ho wever we assume that at least up to a

point the decision of whether or not to act on the spending side or the revenue side

of the government is dictated by political preferences and political bargain which

is, at least to a point, exogenous to the economy and generated by ideological or

policy preferences. Looking at the debates proceeding major ﬁscal changes, and

considering the high degree of uncertainty about the size of ﬁscal multipliers this

assumption holds some water. Thus our only emphasis is on the effects of different

composition of ﬁscal stimuli and adjustments. We cannot and do not compute the

size of ﬁscal multipliers. We only compare the effects of different compositions of

major ﬁscal changes.

2.2 Data and Sources

We use a panel of OECD countries for a maximum time period from 1970 to 2007.

The countries included in the sample are: Australia, Austria, Belgium, Canada,

Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, Netherlands,

New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom,

and United States. All ﬁscal and macroeconomic data are from the OECD Eco-

nomic Outlook Database no. 84.

Our approach identiﬁes episodes of large changes in the ﬁscal stance and stud-

ies the behavior of ﬁscal and macroeconomic variables around those episodes to

investigate whether different characteristics of ﬁscal packages are correlated with

dif ferent macroeconomic outcomes. More speciﬁcally, we focus both on the size

of the ﬁscal packages (i.e.: the magnitude of the change of the government deﬁcit)

and on its composition (i.e.: the percentage change of the main government budget

items relative to the total change) and we investigate whether large ﬁscal stimuli

and adjustments that differ in size and composition are associated with booms or

economic recessions (as deﬁned below) and whether governments that implement

dif ferent types of ﬁscal adjustments are successful / unsuccessful in reducing gov-

ernment debt.

We use a cyclically adjusted value of the ﬁscal variables to leave aside vari-

ations of the ﬁscal variables induced by business cycle ﬂuctuations. The cyclical

adjustment is based on the method proposed by Blanchard (1993). It is a simple

method and rather transparent, which corrects various component of the govern-

ment budget for year to year changes in the unemployment rate. More precisely,

the cyclically adjusted value of the change in a ﬁscal variable is the difference be-

tween a measure of the ﬁscal variable in period t computed as if the unemployment

rate were equal to the one in t − 1 and the actual value of the ﬁscal variable in year

t − 1.

We prefer this method to more complicated measures like those produced

by the OECD because the latter are a bit of a black box based upon many assump-

tions about ﬁscal multipliers upon which there is much uncertainty. Based on our

previous work (Alesina and Ardagna (1998)) we are conﬁdent that for the large

episodes which we consider the details of how to adjust for the cycle do not matter

much for the qualitative nature of the results. In fact, even not correcting at all

would give similar results.

To calculate the measure of the ﬁscal variable in period t as if the unemployment rate were

equal to the one in t − 1, we follow the procedure in Alesina and Perotti (1995). Speciﬁcally, for

eachcountryinthesample,weregresstheﬁscal policy variable as share of GDP, on a time trend

and on the unemployment rate. Then, using the coefﬁcients and the residuals from the estimated

regressions, we predict what the value of the ﬁscal v ariable as a share of GDP in period t would have

been if the unemployment rate were the same as in the previous year.

More on this is available from the authors.

2.3 Deﬁnition of the episodes

To identify episodes of ﬁscal adjustments and ﬁscal stimuli we focus on large

changes of ﬁscal policy and use the following rule.

Deﬁnition 1 Fiscal adjustments and stimuli

Aperiodofﬁscal adjustment (stimulus) is a year in which the cyclically ad-

justed primary balance improves (deteriorates) by at least 1.5 per cent of GDP.

These are rather demanding criteria, which rule out small, but prolonged, ad-

justments/stimuli. We have chosen them because we are particularly interested in

episodes which are very sharp and large and clearly indicate a change in the ﬁs-

cal stance. This deﬁnition misses ﬁscal adjustments and stimuli which are small

in each year but prolonged for several years. It would be quite difﬁcult to come

up with a deﬁnition that captured the many possible pattern of multi years small

adjustments. Thus, the study of these episodes gives a clue on what happens with

sharp and brief changes in the ﬁscal stance.

We use the primary deﬁcit, (i.e.: the difference between current and capital

spending, excluding interest rate expenses paid on government debt, and total tax

re venue)

rather than the total deﬁcit, to av oid that episodes selected result from the

ef fect that changes in interest rates have on total government expenditures. Using

these criteria we try to focus as much as possible on episodes that do not result from

the automatic response of ﬁscal variables to economic growth or monetary pol-

icy induced changes on interest rates, but they should reﬂect discretionary policy

choices of ﬁscal authorities. Needless to say, there can still be an endogeneity issue

related to the occurrence of ﬁscal adjustments and expansions, because, in princi-

ple, discretionary policy choices of ﬁscal authorities can be af fected by countries’

macroeconomic conditions. However, note that the budget for the current year is

approved during the second half of the previous year and, even though additional

measures can be taken during the course of the year, they usually become effective

with some delay, generally toward the end of the ﬁscal year.

Deﬁnition 1 selects 107 periods of ﬁscal adjustments (15.1% of the observa-

tions in our sample) and 91 periods of ﬁscal stimuli (12.9% of the observations

in our sample). Table A1 in appendix lists all of them. Of the 107 episodes of

ﬁscal adjustments, 65 last only for one period, while the rest are multiperiods ad-

justments. The majority of the latter (13) last for two consecutive years, 4 are

See the appendix for a detail deﬁnition of each variable used in the empirical analysis.

three years adjustments and the Denmark 1983-1986 ﬁscal stabilization is the only

episode lasting 4 consecutive years. As for ﬁscal stimuli, 52 episodes last one pe-

riod, in 12 cases the stimulus continues in the second year as well, and in 5 cases

deﬁnition 1 selects ﬁscal stimuli that last for 3 consecutive years.

We are interested in two outcomes of very tight and very loose ﬁscal poli-

cies: whether they are associated with an exp ansion in economic activity during

and in their immediate aftermath and whether they are associated with a reduc-

tion in the public debt-to-GDP ratio. Thus, an episode is deﬁned expansionary

accordingtodeﬁnition 2 and successful according to deﬁnition 3; we deﬁne con-

tractionary/unsuccessful all the episodes of ﬁscal stimuli and adjustments that are

not expansionary/successful according to these deﬁnitions.

Deﬁnition 2 Expansionary ﬁscal adjustments and ﬁscal stimuli

An episode of ﬁscal adjustment (ﬁscal stimulus) is expansionary if the average

growth rate of GDP, in difference from the G7 average (weighted by GDP weights),

in the ﬁrst period of the episode and in the two years after is greater than the

value of 75th percentile of the same variable empirical density in all episodes of

ﬁscal adjustments (ﬁscal stimuli). This deﬁnitions selects 26 years of expansionary

periods during ﬁscal adjustments (3.7% of the observations of the entire OECD

sample) and 20 years of exp ansionary periods during ﬁscal stimuli (2.8% of the

observations of the entire OECD sample). See table A2 for a list

Deﬁnition 3 Successful ﬁscal adjustments

Aperiodofﬁscal adjustment is successful if the cumulative reduction of the

debt to GDP ratio three years after the beginning of a ﬁscal adjustment is greater

than 4.5 percentage points (the value of 25th percentile of the change of the debt-to-

GDP ratio empirical density in all episodes of ﬁscal adjustments).

This deﬁnitions

selects 17 periods of successful ﬁscal adjustments (2.7% of the observations of the

entire OECD sample). In Table A3 in Appendix we list all the episodes.

We have experimented with variation of the threshold of these deﬁnitions but

the results are rob ust, that is they do not change signiﬁcantly as result of small

If an episode of tight ﬁscal policy takes place in 2005, the cumulative change of the debt-to-

GDP ratio is computed over a two years horizon, not to loose too many observations at the end of

the sample. If the episode occurs after 2005, we cannot determine whether it is a successful or an

unsuccessful one.

changes of the deﬁnitions. A value of 1.5 change in deﬁcits in a year is sufﬁ-

ciently high to eliminate years of "business as usual" in which ﬂuctuations of the

deﬁcits may just be only cyclical. However it is not so large as to have very fe w

data points. Also, our “horizon” for the deﬁnition of “expansionary” and “success”

is relatively short. Choosing a longer horizon has two problems. First, one looses

many observations at the end of the sample; second, and more importantly, choos-

ing a longer horizon makes the connection between the episodes and economic

outcomes several years later more tenuous, given the extent of intervening factors.

Finally, note that according to deﬁnition 2 and 3, multiyears ﬁscal adjustments and

stimuli are considered as a "single" episode because the length of the time horizon

chosen for the deﬁnition of “expansionary” and “success” starts from the ﬁrst year

of the episode. Alesina and Ardagna (1998), Alesina, Perotti and Tavares (1998),

instead, consider each year of a multiyear period as a single episode. This implies

that, in a multiyear episode, some years can be expansionary, some contractionary,

some can be successful, some unsuccessful. While we have no reason to prefer one

choice over the other, we ﬁnd reassuring that results are robust to these alternative

methods used to select expansionary and successful episodes that last more than

one consecutive year.

3 Basic Statistics

3.1 Fiscal stimuli

Let’s begin by analyzing what happens with ﬁscal stimuli, namely whether we can

detect differences in the effects of ﬁscal packages depending on their composition

on the economy. Table 1 shows the composition in terms of spending components

and revenue components of the 20 years of expansionary ﬁscal stimulus packages

versus the others. In Tables 1-6, the period [T − 2, T − 1] is the two year period

preceding the ﬁrst year of a ﬁscal stimulus/adjustment. The period [T]istheﬁrst

year and the period [T + 1, T + 2] is the two year period following the beginning

of an episode.

All the v ariables in the tables are yearly av erages.

The most striking result of this table is that in expansionary episodes total

spending increases by roughly 1 per cent of GDP while rev enues fall by more

than 2.5 per cent of GDP. In contractionary episodes total spending goes up by

More details on these sensitivity analysis are available from the authors.

The Denmark ﬁscal contractio n is the only episode lasting 4 years. We have included the values

of the variables in 1986 in the column [T + 1, T + 2]. We checked and conﬁrm that the qualitative

nature of the results does not change if the period [T] inlcudes all the years of a tight/expansionary

episode of ﬁscal policy and the period [T + 1, T + 2] is the two year period following the last year

of an episode.

close to 3 per cent of GDP while revenues are roughly constant in terms of GDP.

This correlation seems to suggest that stimulus packages used upon the spending

side do not work or at least not as well as those based upon spending increases.

In terms of components of spending we note that there is no difference between

expansionary and contractionary episodes regarding public investment which goes

up by roughly the same amount in ratios of GDP. All the other components of

primary spending and, in particular transfers, go up much more in contractionary

episodes. This suggests that the non public investment components of the bud-

get are those which explain the different correlation with growth. As for revenues

note the large cut in income taxes in expansionary stimuli and the slight increase

in contractionary ones. Not surprisingly the debt over GDP ratio goes up less in

expansionary episodes since the denominator increases more.

Figure 1 offers a striking visual image of the different compositions in terms

of revenues and spending of expansionary and contractionary episodes. The ﬁrst

two comparison of total spending and rev e nues are rather striking even visually.

In Table 2 we look at the different components of GDP to check whether there

are difference in composition between expansionary and contractionary episodes.

The ﬁrst two lines which refer to GDP growth are somew hat obvious since they

reﬂect the selection criteria of these episodes. All the components of aggregate

demand grow more after the stimulus in expansionary episodes. This result is a

bit different than that reported in Alesina and Ardagna (1998). In that sample the

dif ference between the two types of episodes seemed concentrated on inv estment

rather than consumption.

In this sample both consumption and investment behav e

differently, both increasing in expansionary cases and declining in contractionary

ones. This table also allows us to check whether the state of the economy before

the adjustments was dif ferent in the two groups. In terms of domestic growth and

relative to G7 av erage, expansionary episodes occurred when growth was higher.

As for the other components the only signiﬁcant difference seem to be in the trade

balance. It is obviously cavalier to draw broad conclusions from this but enormous

differences in the preexisting state of the economy do not jump out from this table.

3.2 Fiscal adjustments

Fiscal adjustments can be judged in two ways, as discussed above. One is about

whether they have been successful in signiﬁcantly reducing deﬁcits and the debt

over GDP ratios and second whether they have been associated with a reduction

in growth or not. Obviously, the two criteria are correlated since a growth en-

hancing adjustment is more likely to be successful in reducing the debt-to-GDP

See Also Alesina, Ardagna Perotti and Schiantarelli (2002) for related work on the effect of

ﬁscal policy on investment.

ratio. However, the correlation is not perfect since a ﬁscal adjustment may lead to

a sharp reduction of the debt/GDP ratio because the numerator drops faster than

the denominator. Episodes with this characteristic, that is the ability to reduce the

debt-to-GDP ratio exist, for example Netherlands in 1993, Norway in 1989, and

Sweden in 1986-1987.

Table 3 is organized in the same way as Table 1 above. The ex pansionary

episodes of ﬁscal adjustments are mostly characterized by spending cuts. Primary

spending as a percent of GDP falls by more than 2 per cent. Total revenues in-

stead increase slightly by about 0.34 per cent of GDP. On the other hand, in the

case of contractionary ﬁscal adjustments primary spending is cut by about 0.7 per

cent of GDP, while revenues increase by about 1.2 per cent of GDP. Thus, ﬁscal

adjustments occurring on the spending side have superior effects on growth than

those based upon increases in tax revenues. As far as the composition in compo-

nents probably the most striking difference between the two types of adjustments

has to do with the role of transfers. In contractionary cases transfers continue to

growth as a percentage of GDP of almost half of a percentage point. In expansion-

ary episodes, instead, transfers fall by roughly the same amount. Thus, in between

the two types of episodes there is a very large difference of 1 per cent of GDP in the

share of transfers. Looking at the composition of revenues one is struck by income

taxes: they go down quite signiﬁcantly in expansionary adjustments and go up in

contractionary ones. The difference between the two is almost 1 percentage point

of GDP. This difference is by far the largest among revenue components.

Figure 2 is organized in the same way as ﬁgure 1 and even in this case visually

the contrast between the two types of ﬁscal adjustments is quite obvious. When

we look at the different components of GDP, we ﬁnd that both consumption and

investment grow more during expansionary episodes. We did not uncover any re-

markable composition effects, along the same line a Table 2 displayed for ﬁscal

stimuli. These sample statistics are reported in Table 4 which is organized as Ta-

ble 2. The other interesting observa tion is that at least in terms of GDP growth

and growth of its components the preexisting conditions of expansionary and con-

tractionary episodes look remarkably similar. One rather remarkable observation

comes from comparing the growth performance during expansionary stimuli and

expansionary adjustments: they are quite similar!

Let’s now consider successful versus unsuccessful adjustments as shown in

Table 5. The comparison between the two is especially striking. In successful

episodes total primary spending as a percentage of GDP falls by about 2 per cent

of GDP. Total revenues actually decline of about half of percentage point of GDP.

Thus, successful ﬁscal adjustments are completely based on spending cuts accom-

panied by modest tax cuts! On the contrary, in unsuccessful adjustments total

revenue goes up by almost 1.5 per cent of GDP and primary spending are cut by

about 0.8 of GDP. Once again this comparison points in the direction of spending

cuts as the more successful ways of ﬁxing budget problems.

Regarding the composition of spending and revenue the most striking com-

parison is giv en by the transfers item. In successful adjustments transfers fall by

0.83 per cent of GDP, while in unsuccessful adjustments they grow at about 0.4

per cent, a huge difference between the two episodes of 1.2 percent of GDP. This

comparison points in a clear direction: it is very difﬁcult if not impossible to ﬁx

public ﬁnances when in trouble without solving the question of automatic increases

in entitlements. Regarding the composition of revenues, again as above the most

striking difference is on income taxes. Figure 3, once again, gives a striking visual

image of these results.

4 Some regressions

In this section we present some simple regressions on GDP growth as a function

of changes of ﬁscal policy in the recent past. We should put up-front the fact that

causality issues are all over the place here and we do not claim to have solved them.

These regressions should be viewed as correlations, but we ﬁnd them instructi ve

and the message which they send is on the same line of that emerging from our

descriptive analysis above.

Let’s begin with ﬁscal stimuli. In Table 7, columns 1-4, we regress real GDP

gro wth in a year of ﬁscal stimulus on its one period and two period lagged values,

on the lagged value of the weighted average of the real GDP growth of the G7

countries, on the lagged value of the ratio of public debt to GDP ratio and on a

set of ﬁscal policy variables measuring the size and the composition of the ﬁscal

stimulus. Columns 5-8 are analogous to the previous 4 columns except for the lhs

variable, now equal to the average of real GDP growth in a year of ﬁscal stimulus

andinthetwofollowingones.

We ﬁnd that, controlling for initial conditions, a one percentage point higher

increase in the current spending to GDP ratio is associated with a 0.75 percentage

point lower growth. The effect is statistically signiﬁcant at the 5% level. Instead,

larger increases in spending on capital goods or larger cuts in taxes do not have sta-

tistically signiﬁcant effects on growth (see column 2). When we try to investigate

whether the size of the ﬁscal stimulus or its composition is relevant for economic

growth, we ﬁnd more evidence in favor of the composition. We measure the size of

the ﬁscal stimulus with the change in the cyclically adjusted primary balance. We

measure the composition of ﬁscal stimuli with two different variables: (i) the ratio

between the change in current spending to GDP ratio and the change in the primary

balance (columns 3 and 7), and (ii) the sum of the change in current spending and

tax revenue to GDP ratios (column 4 and 8) to account for the fact that both current

spending increases and tax increases can be negatively associated with growth.

Both measures of composition are statistically signiﬁcantatthe5%levelinall

speciﬁcations. In column 3, the sign of the ratio between the change in current

spending to GDP ratio and the change in the primary balance indicates that the

larger the share of the worsening in the primary balance due to spending increases

the lower GDP growth. On average, during years of ﬁscal stimuli about 54% of

the deterioration in the primary balance is due to increases in current spending

items. A one standard deviation increase in this variable (equal to 51%, undoubt-

edly a very large number) would reduce growth by 1 percentage point. Finally, a

larger increase in the primary deﬁcittoGDPratioisassociatedwithlowergrowth,

howev er, the effect is statistically signiﬁcant only in column 3.

Table 8 is very similar to Table 7 but we replace the change in current spend-

ing and taxes with their respective components. Consistent with the evidence in

Table 7, our regressions show that ﬁscal stimuli more heavily based on increases in

current spending items (government wage and non-wage components, subsidies)

are associated with lower growth, while ﬁscal stimulus packages based on cuts in

income, business and indirect taxes are more likely to be expansionary.

When we turn to the sample of ﬁscal adjustments (Tables 9 and 10), our results

still point in the same direction: namely, the composition of the ﬁscal adjustment,

more than its size, matters for growth and ﬁscal adjustments associated with higher

GDP growth are those in which a larger share of the reduction of the primary

deﬁcit-to-GDP ratio is due to cuts in current spending, to the government wage

and non-wage components, and to subsidies. All this evidence is consistent with

the previous literature on ﬁscal stabilizations and is robust if we introduce among

the regressors the change in the short-term interest rate as a control for the stance

of monetary policy or the rate of change of the nominal exchange rate to control

for exchange rate devaluations that can occur at the same time of large changes in

the ﬁscal stance (results are not shown but are available upon request).

Finally, we have estimated the same speciﬁcationsas in Tables 7 and 9, columns

1, 2, and 4 for the entire sample of OECD data that, hence, includes episodes of

ﬁscal adjustments, stimuli and years in which the cyclically adjusted primary bal-

ance changes between -1.5% and 1.5%. We have also checked whether there are

non-linearities associated with times of large ﬁscal adjustments and stimuli. Table

11 shows the results.

Results are in line with the evidence shown so far: we

ﬁnd that larger reductions in current spending and in taxation are associated with

higher GDP growth, while changes in capital spending do not show any signiﬁcant

ef fect on growth. Moreover, the speciﬁcations in columns 4-9, do not support any

Regressions in Table 11 include country and year dummies among the rhs varia bles.

evidence of non-linearities in episodes of ﬁscal adjustments or stimuli. Both the

coefﬁcients of the dummy variables Tight and Loose and the coefﬁcients between

the interaction terms of these variables and the ﬁscal policy indicators are not sta-

tistically signiﬁcant. As suggested by Alesina, Ardagna, Perotti, and Schiantarelli

(2002), there seems to be nothing special around such episodes that can explain the

behavior of growth relatively to normal times.

5 Conclusions

Rather than reviewing again our result it is worth elaborating, or perhaps speculat-

ing on the current and future ﬁscal stance in the US. As we well know a very large

portion of the current astronomical 12 percent of GDP deﬁcit is the result of bailout

of various types of the ﬁnancial sector. This is an issue on which this paper has

nothing to say. But part of the deﬁcit is the result of the stimulus package that was

passed to lift the economy out of the recession. About two third of this ﬁscal pack-

age is constituted by increases in spending, including public investment, transfers

and government consumption. According to our results ﬁscal stimuli based upon

tax cut are much more likely to be gro wth enhancing than those on the spending

side. Needless to say when considering a single episode many other factors jump

to mind, factors which are difﬁcult to capture in a multi country regressions. For

instance, American families were saving too little before the crisis. An income

tax cut might have just simply been saved and might have had not a big impact on

aggregate consumption. Howev er, more saving might have reinforced the ﬁnan-

cial sector, think of the credit card crisis for instance. In addition, one could have

though of tax cuts that stimulate investment. The beneﬁt of infrastructure projects

which have "long and variable lags" is much more questionable.

After the "perfect storm" of this current crisis the US will emerge with an

unprecedented (for peace time) increase in government debt. As we argued in the

introduction it is unlikely that these deﬁcits and debt will disappear simply because

growth will resume at very rapid pace very soon. Primary suppresses would be

needed since interest rates cannot go other than up from the close to zero actual

levels. The analysis of the present paper suggests that unless primary spending is

cut, it is difﬁcult to acheive ﬁscal stability because spending may rise faster than

tax revenue. But what can be cut? Hopefully improvements in the peace process in

Afghanistan and Iraq might allow a reduction of military expenditure, b ut given the

instability in the region one cannot count on that for sure. Health care reforms seem

to imply large increases in spending, the retirement of the baby boomers is not too

far, and in the pressing time of the crisis the issue of Social Security has been in

the background, but it has not disappeared A relatively high unemployment for a

couple of more years will require spending on subsidies. The budget outlook looks

rather grim on the spending side. The Congressional Budget Ofﬁce predicts deﬁcit

of 7 per cent of GDP up to 2020. This is not a rosy scenario.

References

[1] Aghion Philippe, Alberto Alesina and Francesco Trebbi (2004) "Endogenous

Political Institutions" Quarterly Journal of Economics, 119, 565-612

[2] Alesina Alberto (1988) “The End of Large Public Debts” in F. Giavazzi and

L. Spaventa (eds.), High Public Debt: The Italian Experience, Cambridge,

Cambridge University Press.

[3] Alesina A., and S. Ardagna, 1998, Tales of Fiscal Adjustments, Economic

Policy, October 1998, 489-545.

[4] Alesina A., S. Ardagna, R. Perotti, and F. Schiantarelli, 2002, Fiscal Policy,

Proﬁts, and Investment, American Economic Review, vol. 92, no. 3, June

2002, 571-589.

[5] Alesina, Alberto and Allan Drazen (1991) “Why are Stabilizations De-

layed?,” American Economic Review, 81, 1170-1188.

[6] Alesina, Alberto and Edward Glaeser (2004) Fighting Poverty in the US and

Europe: A World of Difference (Oxford University Press, Oxford, UK).

[7] Alesina A., R. Perotti, and J. Tavares, 1998, The Political Economy of Fiscal

Adjustments, Brookings Papers on Economic Activity, Spring 1998.

[8] Alesina A., and R. Perotti, 1997, The Welfare State and Competitiv eness,

American Economic Review, 1997, 87, 921-939.

[9] Alesina A., and R. Perotti, 1995, Fiscal Expansions and Adjustments in

OECD Countries, Economic Policy, n.21, 207-247.

[10] Ardagna Silvia (2004), “Fiscal Stabilizations: When Do They Work and

Why”, European Economic Review, vol. 48, No. 5, October 2004, pp. 1047-

1074.

[11] Blanchard O., 1993, Suggestion for a New Set of Fiscal Indicators, OECD

Working paper.

[12] Blanchard O., 1990, Comment on Giavazzi and Pagano, NBER Macroeco-

nomics Annual, MIT Press, Cambridge, MA, 1990.

[13] Blanchard, O.J.. and R. Perotti, (2002), “An Empirical Investigation of the

Dynamic Effects of Changes in Government Spending and Revenues on Out-

put”, Quarterly Journal of Economics, November, pp. 1329-1368.

[14] Bertola, G., Drazen, A., 1993. Trigger Points and Budget Cuts: Explaining

the effects of Fiscal Austerity. American Economic Review 83 (1), 11–26.

[15] Giavazzi, F., T. Jappelli, and M. Pagano, 2000, Searching for Non-Linear

Effects of Fiscal Policy: Evidence from Industrial and Developing Countries,

European Economic Review, 2000, vol. 44, n.7, 1259-1289.

[16] Giavazzi F., and M. Pagano, 1996, Non-Keynesian Effects of Fiscal Policy

Changes: International Evidence and the Swedish Experience, Swedish Eco-

nomic Policy Review, vol. 3, n.1, Spring, 67-112.

[17] Giavazzi F., and M. Pagano, 1990, Can Sev ere Fiscal Contractions Be Expan-

sionary? Tales of Two Small European Countries, NBER Macroeconomics

Annual, MIT Press, (Cambridge, MA), 1990, 95-122.

[18] Lambertini L., and J. Tavares, 2001, Exchange Rates and Fiscal Adjustments:

Evidence from the ECD and Implications for EMU.

[19] McDermott J., and R. Wescott, 1996, An Empirical Analysis of Fiscal Ad-

justments, IMF Staff papers, vol. 43, n.4, 723-753.

[20] Milesi-Ferretti Gian Maria, Roberto Perotti, and Massimo Rostagno (2002)

“Electoral Systems And Public Spending”, The Quarterly Journal of Eco-

nomics, vol. 117(2), 609-657.

[21] Mountford A., H. Uhlig, (2008), “What Are the Effects of Fiscal Plicy

Shocks?” NBER working paper 14551.

[22] Perotti, Roberto, 1999, Fiscal Policy When Things are Going Badly, Quar-

terly Journal of Economics, 114, November 1999, 1399-1436.

[23] Persson Torsten and Guido Tabellini, (2003), “The Economic Effects of Con-

stitutions”, MIT Press, Munich Lectures in Economics.

[24] Ramey V., (2008), “Identifying Government Spending Shocks: It’s All in the

Timing”

[25] Romer Christina and David Romer, (2007), The Macroeconomics Effects of

Tax Change: Estimates Based on a New Measure of Fiscal Shocks”, NBER

working paper 13264.

[26] Sutherland, A., 1997. Fiscal Crises and Aggregate Demand: Can High Public

Debt Reverse the Effects of Fiscal Policy? Journal of Public Economics 65,

147–162.

[27] von Hagen J., A. H. Hallett, R. Strauch, (2002), “Budgetary Consolidation

in Europe: Quality, Economic Conditions, and Persistence”, Journal of the

Japanese and International Economics, vol. 16, pp. 512-535.

[28] von Hagen J., R. Strauch, (2001), “Fiscal Consolidations: Quality, Economic

Conditions, and Success, Public Choice, vol. 109, no.3-4, pp 327-346

6 Data Appendix

• Debt.= government gross debt as a share of GDP

• Total deﬁcit = cyclically adjusted total deﬁcit as a share of GDP = primary

deﬁcit + (interest expenses on government debt/GDP)

• Primary deﬁcit = cyclically adjusted primary deﬁcit as a share of GDP =

Primary expenses - Total revenue

• Primary expenses = cyclically adjusted primary expenditure as a share of

GDP = Transfers + ((Government wage expenditures + Government non

wage expenditures + Subsidies + Government investment)/GDP)

• Curr. G = cyclically adjusted current expenditure as a share of GDP = Trans-

fers + ((Government wage expenditures + Government non wage expendi-

tures + Subsidies)/GDP)

• Transfers = cyclically adjusted transfers as a share of GDP

• Government wage expenditures = government wage bill expenditures

• Government non wage expenditures = government non wage bill expendi-

tures

• Subsidies = subsidies to ﬁrms

• Government investment = gross government consumption on ﬁxed capital

• Total revenue = Tax = cyclically adjusted total re venue as a share of GDP =

Income taxes + Business taxes + Indirect taxes + Social security contribu-

tions + (Other taxes/GDP)

• Income taxes = cyclically adjusted income taxes as a share of GDP = cycli-

cally adjusted direct taxes on household as a share of GDP

• Business taxes = cyclically ad justed business taxes as a share of GDP =

cyclically adjusted direct taxes on businesses as a share of GDP

• Indirect taxes = cyclically adjusted indirect taxes as a share of GDP = cycli-

cally adjusted indirect taxes as a share of GDP

• Social security contributions = cyclically adjusted social security contribu-

tions paid by employers and employees as a share of GDP

• Curr.G/Pr.Deﬁcit; Gov.Inv/Pr.Deﬁcit; Spending item/Pr.Deﬁcit;

= an increase in these variable s means that a larger share of the increase

(reduction) of the primary deﬁcit is obtained by increasing (cutting) current

spending/gov. investment/spending item

• Tax Revenue Item/Pr.Deﬁcit = an increase in these variables means that

a larger share of the increase (reduction) of the primary deﬁcit is obtained by

cutting (increasing) a revenue item of the government budget

• Curr.G+Tax is actually equal to the negative of this v ariable. If both

taxes and spending are cut during the episode of loose or tight ﬁscal policy,

the variable has the “highest positive” value. If, instead, both spending and

taxes increase the variable has the “highest negative value”.

• G7 GDP Growth = average growth rate of real GDP (with GDP weights) of

the seven major industrial countries

• GDP Growth = gro wth rate of real capita GDP

• Trade Balance = T r ade balance as a share of GDP = (Exports of goods and

services - Imports of goods and services)/GDP.

60.00%

80.00%

100.00%

120.00%

70.99%

51.38%

45.58%

116.87%

50.21%

Figure 1

Contribution of expenditure and revenue items to the fiscal stimuli

-20.00%

0.00%

20.00%

40.00%

Primary

expenditures

Total revenue Transfers Government

wage

expenditures

Government

non wage

expenditures

Subsidies Government

investment

Income taxes Business taxes Indirect taxes Social security

contributions

29.01%

4.97%

8.84%

4.42%

5.80%

8.84%

20.17%

-17.96%

-17.28%

32.92%

18.52%

4.53%

10.70%

-9.88%

11.93%

-3.70%

-11.52%

Expansionary episodes Contractionary episodes

Note: Figure 1 shows the percentage of the increase (reduction) in the primary deficit (surplus) due to changes in spending and revenue items of the government budget. Positive values indicate that

expenditure items increase and revenue items decrease, contributing to a worsening of the primary balance. Negative values indicate that expenditure items decrease and revenue items increase,

contributing to an improvement of the primary balance.

20.00%

40.00%

60.00%

80.00%

100.00%

86.22%

13 39%

22.83%

15.75%

12 60%

30.31%

25.98%

34.59%

65.41%

16.22%

37.84%

25.95%

18.92%

11 35%

Figure 2

Contribution of expenditure and revenue items to fiscal adjustments

-40.00%

-20.00%

0.00%

Primary

expenditures

Total

revenue

Transfers Government

wage

expenditures

Government

non wage

expenditures

Subsidies Government

investment

Income taxes Business

taxes

Indirect taxes Social

security

contributions

39%

5.12%

60%

-10.63%

0.39%

-2.76%

-25.41%

-1.08%

4.86%

35%

-3.24%

Expansionary episodes Contractionary episodes

Note: Figure 2 shows the percentage of the decrease (increase) in the primary deficit (surplus) due to changes in spending and revenue items of the government budget. Positive values indicate that

expenditure items decrease and revenue items increase, contributing to an improvement of the primary balance. Negative values indicate that expenditure items increase and revenue items decrease,

contributing to a worsening of the primary balance.

40.00%

60.00%

80.00%

100.00%

120.00%

140.00%

135.42%

57.64%

36.11%

26.39%

56.94%

33.80%

66.20%

35.68%

24.88%

19 25%

Figure 3

Contribution of expenditure and revenue items to fiscal adjustments

-60.00%

-40.00%

-20.00%

0.00%

20.00%

Primary

expenditures

Total

revenue

Transfers Government

wage

expenditures

Government

non wage

expenditures

Subsidies Government

investment

Income taxes Business

taxes

Indirect taxes Social

security

contributions

-35.42%

1.39%

16.67%

-47.92%

-21.53%

-6.25%

-20.19%

15.02%

-1.88%

5.63%

25%

14.55%

5.16%

Successful episodes Unsuccessful episodes

Note: Figure 3 shows the percentage of the decrease (increase) in the primary deficit (surplus) due to changes in spending and revenue items of the government budget. Positive values indicate that

expenditure items decrease and revenue items increase, contributing to an improvement of the primary balance. Negative values indicate that expenditure items increase and revenue items decrease,

contributing to a worsening of the primary balance

(0 33)

(0 32)

(0 37)

(0 13)

(0 14)

(0 15)

Table 1: Fiscal stimuli: size and composition

Expansionary Contractionary

[T-2 - T-1] T [T+1 - T+2] (c) - (a) [T-2 - T-1] T [T+1 - T+2] (c) - (a)

(a) (b) (c) (a) (b) (c)

Debt 50.28 50.52 51.1 0.82 60.79 62.38 63.3 2.51

(9.03) (9.09) (9.48) (5.18) (5.18) (4.46)

Change in debt -1.02 0.48 0.53 1.55 -0.29 2.24 2.21 2.50

(1.47) (1.12) (1.24) (0.59) (0.67) (0.68)

Total deficit -1.04 2.19 3.27 4.31 1.5 3.79 3.97 2.47

(1.62) (1.65) (1.24) (0.72) (0.74) (0.71)

Primary deficit -2.01 1.16 1.61 3.62 -0.3 1.99 2.13 2.43

(0.82) (0.92) (0.91) (0.45) (0.48) (0.41)

Primary expenditures 36.79 37.72 37.84 1.05 40.08 42.22 42.92 2.84

(1.73) (1.64) (1.66) (0.94) (0.94) (1.00)

Transfers 14.93 14.88 15.11 0.18 16.83 17.28 18.05 1.22

(1.03) (1.01) (1.04) (0.60) (0.58) (0.57)

Government wage expenditures 10.62 10.74 10.94 0.32 11.78 12.2 12.58 0.80

(0.52) (0.47) (0.50) (0.41) (0.43) (0.46)

Government non wage expenditures 6.81 6.96 6.97 0.16 7.73 8.15 8.18 0.45

(0.49) (0.49) (0.55)

(0.29) (0.28) (0.31)

Subsidies 2.03 2.09 2.24 0.21 1.82 1.93 1.93 0.11

33)

(0 32)

. .

(0 37)

(0 13)

(0 14)

(0 15)

. .

Government investment 2.26 3.05 2.58 0.32 1.95 2.67 2.21 0.26

(0.37) (0.38) (0.37) (0.19) (0.27) (0.21)

Total revenue 38.8 36.56 36.23 -2.57 40.38 40.23 40.8 0.42

(1.90) (1.83) (2.00) (1.15) (1.12) (1.07)

Income taxes 10.89 9.2 9.03 -1.86 11.02 11.21 11.26 0.24

(1.10) (0.98) (1.08) (0.74) (0.71) (0.67)

Business taxes 4.25 3.37 2.6 -1.65 3.03 2.78 2.74 -0.29

(0.83) (0.63) (0.33) (0.25) (0.20) (0.22)

Indirect taxes 13.33 12.57 12.6 -0.73 12.67 12.5 12.76 0.09

(0.61) (0.61) (0.69) (0.39) (0.40) (0.36)

Social security contributions 8.7 8.93 9.35 0.65 11.08 11.17 11.36 0.28

(0.94) (0.82) (0.89) (0.69) (0.68) (0.70)

Source: OECD. Variables are in share of GDP. Total deficit, Primary deficit, Primary expenditures, Transfers, Total revenues, and all revenue items are cyclically adjusted variables.

Standard deviations of the means in parenthesis. See the Data Appendix for the exact definition of the variables

Table 2: Fiscal stimuli and growth

[T-2 - T-1] T [T+1 - T+2] (c) - (a) [T-2 - T-1] T [T+1 - T+2] (c) - (a)

(a) (b) (c) (a) (b) (c)

G7 GDP Growth 0.39 1.6 2.03 1.64 0.2 -0.7 -0.74 -0.94

(0.66) (0.53) (0.32) (0.23) (0.23) (0.19)

GDP Growth 3.9 3.77 4.37 0.47 2.89 0.93 1.79 -1.1

(0.65) (0.35) (0.32) (0.22) (0.26) (0.27)

Private Consumption Growth 3.49 3.47 3.72 0.23 3.08 1.54 1.86 -1.22

(0.7) (0.61) (0.32) (0.22) (0.3) (0.29)

Total Investment Growth 3.44 2.58 6.55 3.11 2.9 -1.39 0.04 -2.86

(1.81) (1.63) (1.00) (0.59) (0.82) (0.64)

Private Investment Growth 3.5 1.14 7.49 3.99 3.36 -1.9 0.07 -3.29

(2.05) (2.04) (1.25) (0.73) (1.02) (0.82)

Business Investment Growth 5.51 2.5 7.64 2.13 6.73 -0.34 -0.78 -7.51

(2.06) (3.21) (1.53) (1.44) (1.34) (1.07)

Trade Balance 0.53 0.61 -1.9 -2.43 0.19 -0.2 0.14 -0.05

(2.07) (2.2) (2.11) (0.7) (0.65) (0.69)

Expansionary Contractionary

(0 33)

(0 30)

(0 28)

(0 13)

(0 14)

Table 3: Expansionary and contractionary fiscal adjustments: size and composition

Expansionary Contractionary

[T-2 - T-1] T [T+1 - T+2] (c) - (a) [T-2 - T-1] T [T+1 - T+2] (c) - (a)

(a) (b) (c) (a) (b) (c)

Debt 59.86 57.53 54.1 -5.76 69.15 71.8 69.52 0.37

(5.52) (5.22) (5.07) (4.04) (4.23) (4.25)

Change in debt -1.46 -2.42 -2.3 -0.84 3.28 1.97 1.28 -2.00

(1.03) (1.14) (0.54) (0.62) (0.54) (0.52)

Total deficit 3.61 1.33 0.56 -3.05 5.67 3.89 4.14 -1.53

(1.09) (1.18) (0.98) (0.63) (0.70) (0.72)

Primary deficit 1.31 -0.84 -1.23 -2.54 2.7 0.74 0.85 -1.85

(0.77) (0.74) (0.60) (0.40) (0.43) (0.39)

Primary expenditures 41.32 39.71 39.13 -2.19 43.22 42.47 42.58 -0.64

(2.04) (1.80) (1.59) (0.98) (0.95) (0.93)

Transfers 18.1 17.66 17.52 -0.58 17.95 18.21 18.42 0.47

(1.37) (1.21) (1.08) (0.54) (0.54) (0.54)

Government wage expenditures 11.65 11.41 11.25 -0.40 12.46 12.25 12.16 -0.30

(0.53) (0.51) (0.46) (0.40) (0.38) (0.36)

Government non wage expenditures 7.03 6.91 6.9 -0.13 8.09 8.1 8.11 0.02

(0.53) (0.49)

(0.48) (0.27) (0.28) (0.28)

Subsidies 2.17 1.95 1.85 -0.32 2.07 1.98 1.98 -0.09

33)

(0 30)

(0 28)

(0 13)

(0 14)

. .

Government investment 2.38 1.77 1.61 -0.77 2.66 1.95 1.96 -0.70

(0.29) (0.27) (0.25) (0.18) (0.17) (0.14)

Total revenue 40.02 40.56 40.36 0.34 40.52 41.73 41.73 1.21

(1.99) (1.90) (1.84) (0.98) (0.94) (0.96)

Income taxes 10.62 10.59 10.35 -0.27 11.45 11.79 11.93 0.48

(1.04) (1.07) (0.97) (0.63) (0.63) (0.63)

Business taxes 2.92 3.49 3.58 0.66 2.55 2.88 2.9 0.35

(0.41) (0.51) (0.50) (0.17) (0.22) (0.25)

Indirect taxes 13.52 13.61 13.53 0.01 12.44 12.69 12.65 0.21

(0.46) (0.40) (0.40) (0.31) (0.30) (0.30)

Social security contributions 9.63 9.52 9.56 -0.07 11.44 11.52 11.38 -0.06

(1.01) (0.92) (0.89) (0.61) (0.62) (0.66)

Source: OECD. Variables are in share of GDP. Total deficit, Primary deficit, Primary expenditures, Transfers, Total revenues, and all revenue items are cyclically adjusted variables.

Standard deviations of the means in parenthesis. See the Data Appendix for the exact definition of the variables

Table 4: Expansionary and contractionary fiscal adjustments and growth

[T-2 - T-1] T [T+1 - T+2] (c) - (a) [T-2 - T-1] T [T+1 - T+2] (c) - (a)

(a) (b) (c) (a) (b) (c)

G7 GDP Growth 0.57 1.49 1.98 1.41 -0.32 -0.42 -0.49 -0.17

(0.55) (0.37) (0.24) (0.2) (0.2) (0.17)

GDP Growth 3.14 4.73 4.68 1.54 2.03 2.36 2.25 0.22

(0.56) (0.39) (0.33) (0.2) (0.18) (0.18)

Private Consumption Growth 2.82 4.12 4.34 1.52 1.94 2.27 2.27 0.33

(0.49) (0.47) (0.42) (0.26) (0.24) (0.19)

Total Investment Growth 1.44 7.72 7.91 6.47 1 1.91 2.5 1.5

(1.68) (0.98) (1.12) (0.61) (0.54) (0.72)

Private Investment Growth 1.41 9.6 7.81 6.4 1.04 2.92 3.15 2.11

(1.86) (1.22) (1.33) (0.75) (0.69) (0.89)

Business Investment Growth 2.23 10.88 4.98 2.75 2.97 3.23 5.17 2.2

(1.9) (1.76) (2.62) (1) (1.18) (1)

Trade Balance 0.71 1.85 1.56 0.85 -0.54 0.15 0.95 1.49

(1.58) (1.61) (1.81) (0.58) (0.64) (0.65)

Expansionary Contractionary

(0 36)

(0 35)

(0 34)

(0 14)

(0 15)

Table 5: Successful and unsuccessful fiscal adjustments: size and composition

Successful Unsuccessful

[T-2 - T-1] T [T+1 - T+2] (c) - (a) [T-2 - T-1] T [T+1 - T+2] (c) - (a)

(a) (b) (c) (a) (b) (c)

Debt 61.92 59.63 53.18 -8.74 68.29 71.4 72.06 3.77

(4.32) (4.50) (4.16) (4.32) (4.53) (4.48)

Change in debt -1.6 -1.97 -3.88 -2.28 3.68 2.29 2.14 -1.54

(0.72) (1.14) (0.34) (0.64) (0.53) (0.43)

Total deficit 2.5 0.29 0.66 -1.84 5.6 3.77 3.69 -1.91

(1.00) (1.06) (1.09) (0.71) (0.83) (0.85)

Primary deficit 0.8 -1.2 -0.64 -1.44 2.7 0.71 0.57 -2.13

(0.68) (0.64) (0.69) (0.45) (0.51) (0.46)

Primary expenditures 45.78 43.67 43.83 -1.95 43.46 42.68 42.74 -0.72

(1.76) (1.60) (1.46) (1.10) (1.10) (1.03)

Transfers 19.86 19.07 19.03 -0.83 18.38 18.59 18.81 0.43

(1.11) (0.94) (0.89) (0.63) (0.64) (0.61)

Government wage expenditures 12.82 12.5 12.3 -0.52 12.51 12.3 12.19 -0.32

(0.69) (0.67) (0.63) (0.44) (0.42) (0.40)

Government non wage expenditures 8.73 8.62 8.71 -0.02 7.96 8.01 8 0.04

(0.49) (0.47)

(0.45) (0.30) (0.31) (0.30)

Subsidies 2.29 2.14 2.05 -0.24 2.05 1.94 1.93 -0.12

36)

(0 35)

(0 34)

(0 14)

(0 15)

. .

Government investment 2.12 1.34 1.74 -0.38 2.57 1.85 1.81 -0.76

(0.38) (0.34) (0.27) (0.19) (0.18) (0.16)

Total revenue 44.98 44.86 44.47 -0.51 40.76 41.97 42.17 1.41

(1.61) (1.57) (1.67) (1.04) (1.04) (1.03)

Income taxes 13.69 13.43 13 -0.69 11.02 11.35 11.55 0.53

(1.18) (1.17) (1.16) (0.64) (0.65) (0.64)

Business taxes 2.77 3.37 3.59 0.82 2.69 3.08 3.1 0.41

(0.26) (0.31) (0.35) (0.22) (0.28) (0.31)

Indirect taxes 13.77 13.6 13.46 -0.31 12.32 12.51 12.63 0.31

(0.68) (0.61) (0.62) (0.33) (0.32) (0.33)

Social security contributions 10.82 10.73 10.73 -0.09 12.04 12.25 12.15 0.11

(1.26) (1.15) (1.20) (0.62) (0.62) (0.64)

Source: OECD. Variables are in share of GDP. Total deficit, Primary deficit, Primary expenditures, Transfers, Total revenues, and all revenue items are cyclically adjusted variables.

Standard deviations of the means in parenthesis. See the Data Appendix for the exact definition of the variables

Table 6: Successful and unsuccessful fiscal adjustments and growth

[T-2 - T-1] T [T+1 - T+2] (c) - (a) [T-2 - T-1] T [T+1 - T+2] (c) - (a)

(a) (b) (c) (a) (b) (c)

G7 GDP Growth 0.4 0.8 0.85 0.45 -0.18 -0.22 -0.12 0.06

(0.53) (0.46) (0.37) (0.23) (0.22) (0.18)

GDP Growth 2.99 3.61 3.45 0.46 2.07 2.56 2.52 0.45

(0.58) (0.5) (0.28) (0.25) (0.2) (0.21)

Private Consumption Growth 2.75 3.74 3.02 0.27 2.01 2.28 2.42 0.41

(0.6) (0.67) (0.3) (0.26) (0.23) (0.2)

Total Investment Growth 2.95 4.11 4.78 1.83 1.02 2.55 3.52 2.5

(1.37) (1.54) (1.24) (0.69) (0.56) (0.73)

Private Investment Growth 3.45 5.6 5.07 1.62 1.18 3.43 4.23 3.05

(1.46) (1.85) (1.43) (0.81) (0.73) (0.9)

Business Investment Growth 3.2 5.46 6.06 2.86 3.23 5.17 5.84 2.61

(1.79) (2.06) (1.42) (1.07) (0.97) (1.08)

Trade Balance 2.72 3.99 4.31 1.59 -0.19 0.48 1.15 1.34

(1.1) (1.03) (1.51) (0.71) (0.77) (0.84)

Successful Unsuccessful

51)

29)

57)

68)

30)

64)

Table 7: GDP growth during and in the aftermath of a fiscal stimulus

(1) (2) (3) (4) (5) (6) (7) (8)

GDP growth GDP growth GDP growth GDP growth

Avg. GDP gr. Avg. GDP gr. Avg. GDP gr. Avg. GDP gr.

GDP growth (-1) 0.467*** 0.484*** 0.51***7 0.48*** 0.217* 0.236** 0.266** 0.237**

(3.18) (3.62) (3.76) (3.66) (1.84) (2.15) (2.40) (2.17)

GDP growth (-2) -0.16 -0.08 -0.10 -0.08 -0.08 -0.02 -0.04 -0.028

(-1.16) (-0.60) (-0.78) (-0.68) (-0.74) (-0.19) (-0.39) (-0.27)

G7 GDP growth (-1) 0.36* 0.27 0.25 0.27 -0.164 -0.23 -0.244 -0.228

(1.80) (1.47) (1.34) (1.49) (-1.03) (-1.53) (-1.61) (-1.53)

Debt (-1) -0.004 -0.007 -0.009 -0.0068 -0.003 -0.006 -0.0061 -0.005

(-0.54) (-0.90) (-1.10) (-0.93) (-0.37) (-0.78) (-0.74) (-0.77)

Δ Curr. G

-0.75*** -0.44**

(-2.87) (-2.02)

Δ Gov. Inv

-0.256 -0.076

(-1.38) (-0.50)

Δ Tax

-0.177 -0.199

(-0.62) (-0.85)

Δ Pr. Deficit

-0.283 -0.428** -0.264 -0.102 -0.197 -0.089

(

1.51)

(-1

(

2.29)

(-2

(

1.57)

(

0.68)

(-1

(-0

(

1.30)

(

0.64)

(-1

(-0

Δ Curr. G/Δ Pr. Deficit

-0.02*** -0.016***

(-3.43) (-3.37)

Δ Gov. Inv/Δ Pr. Deficit

-0.003 -0.005

(-0.39) (-0.73)

Δ Curr. G + Δ Tax

0.466*** 0.323***

(4.07) (3.44)

Constant 0.008 0.012 0.026*** 0.012 0.023*** 0.026*** 0.037*** 0.026***

(0.90) (1.38) (2.66) (1.45) (3.13) (3.52) (4.57) (3.78)

Observations 72 72 72 72 69 69 69 69

R-squared 0.28 0.43 0.40 0.43 0.06 0.21 0.21 0.21

Notes: OLS regressions. Dependent variables: real GDP growth rate during the fiscal stimulus in columns 1-4; average real GDP growth rate during the fiscal stimulus and in the following two years

in columns 5-8. T-statistics in parenthesis. See the Data Appendix for the exact definition of the variables.

Table 8: GDP growth and the composition of a fiscal stimulus

(1) (2) (3) (4) (5) (6)

GDP growth GDP growth GDP growth Avg. GDP gr. Avg. GDP gr. Avg. GDP gr.

GDP growth (-1) 0.36** 0.53** 0.48*** 0.26** 0.38*** 0.25**

(2.61) (3.81) (3.29) (2.25) (3.41) (2.11)

GDP growth (-2) 0.05 -0.09 -0.23 0.046 -0.08 -0.11

(0.42) (-0.71) (-1.64) (0.41) (-0.81) (-0.94)

G7 GDP growth (-1) 0.14 0.08 0.26 -0.343** -0.357** -0.27

(0.78) (0.44) (1.29) (-2.28) (-2.42) (-1.64)

Debt (-1) -0.0127* -0.01 -0.0003 -0.008 -0.0004 -0.0017

(-1.69) (-1.41) (-0.04) (-1.08) (-0.05) (-0.22)

Δ Tran -0.23 -0.345

(-0.50) (-0.87)

Δ Gov. non wage exp. -3.10*** -3.01***

(-3.57) (-4.06)

Δ Gov. wage exp. -1.32** -0.034

(-2.43) (-0.07)

Δ Subsidies -1.50 -0.623

(-1.49) (-0.73)

Δ Gov. Inv -0.22 -0.059

(-1.35) (-0.42)

Δ Income taxes 0.12 -0.281

(0.30) (-0.85)

Δ Bus. taxes -0.23 -0.121

(-0.73) (-0.45)

Δ Soc. security contr. 0.248 0.186

(0.56) (0.50)

Δ Indirect taxes -0.181 -0.167

(-0.37)

()

(-0.40)

()

Δ other taxes -3.001*** -2.022**

(-2.80) (-2.24)

Δ Pr. Deficit -0.493*** -0.32* -0.276* -0.077

(-2.76) (-1.80) (-1.93) (-0.53)

Δ Tran/Δ Pr. Deficit 0.002 -0.007

(0.26) (-1.06)

Δ Gov. non wage exp./Δ Pr. Deficit -0.053*** -0.065***

(-3.05) (-4.72)

Δ Gov. wage exp/Δ Pr. Deficit -0.032** 0.0019

(-2.07) (0.16)

Δ Subsidies/Δ Pr. Deficit -0.062** -0.04*

(-2.14) (-1.75)

Δ Gov. Inv/Δ Pr. Deficit -0.0016 -0.007

(-0.20) (-1.10)

Δ Income taxes/Δ Pr. Deficit 0.016* 0.015**

(1.93) (2.29)

Δ Bus. taxes/Δ Pr. Deficit 0.029*** 0.017*

(2.83) (1.88)

Δ Soc. security contr./Δ Pr. Deficit 0.01 0.008

(0.99) (1.00)

Δ Indirect taxes/Δ Pr. Deficit 0.030** 0.023**

(2.50) (2.42)

Δ other taxes/Δ Pr. Deficit 0.032** 0.014

(2.48) (1.34)

Constant 0.027*** 0.035*** 0.006 0.035*** 0.041*** 0.019***

(3.12) (3.65) (0.70) (4.72) (5.28) (2.73)

Observations 67 69 70 64 66 67

R-squared 0.63 0.51 0.43 0.47 0.39 0.21

Notes: OLS regressions. Dependent variables: real GDP growth rate during the fiscal stimulus in columns 1-3; average real GDP growth

rate during the fiscal stimulus and in the following two years in columns 4-6. T-statistics in parenthesis.

S h A di f h d fi i i f h i bl

See the Data Appendix for the exact definition of the variables.

0 017***

0 015***

Table 9: GDP growth during and in the aftermath of a fiscal adjustment

(1) (2) (3) (4) (5) (6) (7) (8)

GDP growth GDP growth GDP growth GDP growth

Avg. GDP gr. Avg. GDP gr. Avg. GDP gr. Avg. GDP gr.

GDP growth (-1) 0.296*** 0.288*** 0.269*** 0.30*** 0.198** 0.197** 0.182** 0.202***

(2.99) (3.12) (3.04) (3.29) (2.41) (2.56) (2.48) (2.66)

GDP growth (-2) -0.0013 0.08 0.123 0.07 -0.059 0.01 0.045 0.007

(-0.01) (0.98) (1.50) (0.86) (-0.80) (0.14) (0.66) (0.10)

G7 GDP growth (-1) 0.116 0.038 0.018 0.025 0.005 -0.068 -0.08 -0.07

(0.76) (0.27) (0.13) (0.18) (0.04) (-0.58) (-0.72) (-0.63)

Debt (-1) -0.011* -0.006 -0.007 -0.009 -0.008 -0.006 -0.006 -0.006

(-1.84) (-1.11) (-1.33) (-1.54) (-1.42) (-1.05) (-1.22) (-1.20)

Δ Curr. G

-0.433** -0.296**

(-2.55) (-2.10)

Δ Gov. In

0.082 0.046

(0.60) (0.41)

Δ Tax

-0.22 -0.26

(-1.09) (-1.56)

Δ Pr. Defici

-0.044 -0.023 0.016 -0.027 0.006 0.024

(-0.33) (-0.19) (0.13) (-0.24) (0.06) (0.23)

Curr G/

Pr Deficit

urr.

e c

0 017***

0 015***

(4.70) (4.81)

Δ Gov. Inv

Δ Pr. Defici

0.0013 0.004

(0.28) (0.96)

Δ Curr. G + Δ Tax

0.34*** 0.284***

(3.80) (3.84)

Constant 0.027*** 0.024*** 0.019*** 0.027*** 0.029*** 0.029*** 0.024*** 0.03***

(3.85) (3.44) (2.97) (4.23) (4.90) (4.87) (4.28) (5.41)

Observations 88 88 88 88 83 83 83 83

R-squared 0.22 0.35 0.40 0.34 0.12 0.27 0.34 0.27

Notes: OLS regressions. Dependent variables: real GDP growth rate during the fiscal adjustment in columns 1-4; average real GDP growth rate during the fiscal adjustment and in the following two years

in columns 5-8. T-statistics in parenthesis. See the Data Appendix for the exact definition of the variables.

Table 10: GDP growth and the composition of a fiscal adjustment

(1) (2) (3) (4) (5) (6)

GDP growth GDP growth GDP growth Avg. GDP gr. Avg. GDP gr. Avg. GDP gr.

GDP growth (-1) 0.208** 0.26*** 0.276** 0.127 0.187** 0.155*

(2.11) (2.99) (2.54) (1.58) (2.56) (1.80)

GDP growth (-2) 0.112 0.13 0.072 0.079 0.06 0.036

(1.26) (1.59) (0.74) (1.09) (0.88) (0.47)

G7 GDP growth (-1) 0.068 -0.05 0.108 -0.048 -0.15 -0.04

(0.44) (-0.37) (0.61) (-0.39) (-1.33) (-0.30)

Debt (-1) -0.013** -0.010* -0.013* -0.014** -0.008* -0.012*

(-2.08) (-1.74) (-1.95) (-2.43) (-1.69) (-1.99)

Δ Tran -0.057 -0.30

(-0.20) (-1.27)

Δ Gov. non wage exp. -1.53** -0.46

(-2.59) (-0.94)

Δ Gov. wage exp. -1.18*** -1.05***

(-2.66) (-2.85)

Δ Subsidies -1.98** -1.84***

(-2.61) (-2.93)

Δ Gov. Inv 0.044 -0.002

(0.32) (-0.02)

Δ Income taxes -0.016 0.04

(-0.06) (0.18)

Δ Bus. taxes -0.57* -0.79***

(-1.92) (-3.19)

Δ Soc. security contr. -0.04 -0.24

(-0.10) (-0.64)

Δ Indirect taxes -0.19 -0.37

(-0.43)

()

(-1.03)

()

Δ other taxes -0.27 0.106

(-0.50) (0.24)

Δ Pr. Deficit -0.084 0.051 -0.022 0.077

(-0.70) (0.35) (-0.22) (0.67)

Δ Tran/Δ Pr. Deficit 0.006 0.009*

(1.01) (1.93)

Δ Gov. non wage exp./Δ Pr. Deficit 0.025** 0.005

(2.23) (0.56)

Δ Gov. wage exp/Δ Pr. Deficit 0.026*** 0.022***

(3.16) (3.04)

Δ Subsidies/Δ Pr. Deficit 0.043*** 0.036***

(2.69) (2.66)

Δ Gov. Inv/Δ Pr. Deficit -0.0004 0.0026

(-0.08) (0.66)

Δ Income taxes/Δ Pr. Deficit -0.009 -0.005

(-1.42) (-1.11)

Δ Bus. taxes/Δ Pr. Deficit -0.011 -0.015*

(-1.16) (-1.84)

Δ Soc. security contr./Δ Pr. Deficit -0.01 -0.015*

(-1.04) (-1.93)

Δ Indirect taxes/Δ Pr. Deficit -0.015 -0.02***

(-1.62) (-2.68)

Δ other taxes/Δ Pr. Deficit 0.0012 0.0001

(0.11) (0.01)

Constant 0.024*** 0.019*** 0.033*** 0.03*** 0.025*** 0.04***

(3.29) (3.11) (3.99) (4.92) (4.56) (5.78)

Observations 81 88 80 77 83 76

R-squared 0.47 0.46 0.28 0.41 0.38 0.26

Notes: OLS regressions. Dependent variables: real GDP growth rate during the fiscal adjustment in columns 1-3; average real GDP growth

rate during the fiscal adjustment and in the following two years in columns 4-6. T-statistics in parenthesis.

S h A di f h d fi i i f h i bl

See the Data Appendix for the exact definition of the variables.

Inv

0.036

Table 11: GDP growth and fiscal policy

(1) (2) (3) (4) (5) (6) (7) (8) 9

GDP growth GDP growth GDP growth GDP growth GDP growth GDP growth GDP growth GDP growth GDP growth

GDP growth (-1) 0.35*** 0.37*** 0.37*** 0.34*** 0.36*** 0.37*** 0.34*** 0.35*** 0.36***

(8.37) (9.24) (9.45) (8.14) (8.98) (9.21) (7.99) (8.77) (8.93)

GDP growth (-2) -0.038 0.016 0.014 -0.035 0.02 0.017 -0.03 0.026 0.025

(-0.91) (0.41) (0.36) (-0.84) (0.51) (0.43) (-0.71) (0.65) (0.62)

Debt (-1) -0.004 -0.005 -0.005 -0.004 -0.005 -0.005 -0.004 -0.005 -0.005

(-1.09) (-1.31) (-1.34) (-1.04) (-1.22) (-1.27) (-1.10) (-1.28) (-1.39)

Δ Pr. Deficit -0.154*** -0.145*** -0.131** -0.136** -0.11 -0.14

(-3.98) (-3.99) (-1.99) (-2.19) (-1.08) (-1.58)

Tight*Δ Pr. Deficit 0.12 0.16

(0.74) (1.05)

Loose*Δ Pr. Deficit -0.213 -0.126

(-1.27) (-0.80)

Δ Curr. G -0.51*** -0.51*** -0.51***

(-8.37) (-6.20) (-4.59)

Tight*Δ Curr. G 0.29

(1.48)

Loose*Δ Curr. G -0.25

(-1.21)

Δ Gov. Inv -0.07 -0.067 0.006

(-1.16) (-0.89) (0.04)

ht*Δ Gov. Inv

Tight

Gov.

0.036

(0.19)

Loose*Δ Gov. Inv -0.246

(-1.22)

Δ Ta

-0.12** -0.12 -0.134

(-1.97) (-1.45) (-1.29)

Tight*Δ Ta

0.025

(0.11)

Loose*Δ Ta

0.038

(0.17)

Δ Curr. G + Δ Ta

0.314*** 0.312*** 0.315***

(8.36) (8.32) (7.05)

Tight*Δ Curr. G + Δ Ta

-0.145

(-1.52)

Loose*Δ Curr. G + Δ Ta

0.124

(1.34)

Tight -0.0016 -0.0024 -0.0020 0.0019 0.0006 0.0015

(-0.69) (-1.05) (-0.87) (0.52) (0.16) (0.44)

Loose -0.0038 -0.0027 -0.003 0.0014 0.002 0.0017

(-1.45) (-1.11) (-1.23) (0.35) (0.53) (0.44)

Observations 569 569 569 569 569 569 569 569 569

R-squared 0.58 0.64 0.63 0.59 0.64 0.64 0.59 0.64 0.64

Notes: OLS regressions. Dep. var.: real GDP growth rate. Tight=1 in period of a fiscal adjustment, 0 otherwise. Loose = 1 in period of a fiscal stimulus, 0 otherwise. Country and year dummies included. T-stat in (). See Data Appendix.

sca

Adj

men s

Table A1: Episodes of fiscal stimuli and adjustments

Fiscal Stimuli

Australia 1990 1991

Austria 1975 2004

Belgium 1975 1981 2005

Canada 1975 1982 1991 2001

Denmark 1974 1975 1980 1981 1982

Finland 1978 1982 1983 1987 1990 1991 1992 2001 2003

France 1975 1981 1992 1993 2002

Germany 1995 2001

Greece 1981 1985 1989 1995 2001

Ireland 1974 1975 1978 2001 2007

Italy 1972 1975 1981 2001

Japan 1975 1993 1998 2005 2007

Netherlands 1975 1980 1995 2001 2002

New Zealand 1988

Norway 1974 1976 1977 1986 1987 1991 1998 2002 2007

Portugal 1978 1985 1993 2005

Spain 1981 1982 1993

Sweden 1974 1977 1979 1980 1991 1992 2001 2002

United Kingdom 1971 1972 1973 1990 1991 1992 2001 2002 2003

United States 2002

Fiscal Adjustments

t t

Australia 1987 1988

Austria 1984 1996 1997 2005

Belgium 1982 1984 1987 2006

Canada 1981 1986 1987 1995 1996 1997

Denmark 1983 1984 1985 1986 2005

Finland 1973 1976 1981 1984 1988 1994 1996 1998 2000

France 1979 1996

Germany 1996 2000

Greece 1976 1986 1991 1994 1996 2005 2006

Ireland 1976 1984 1987 1988 1989 2000

Italy 1976 1980 1982 1990 1991 1992 1997 2007

Japan 1984 1999 2001 2006

Netherlands 1972 1973 1983 1988 1991 1993 1996

New Zealand 1987 1989 1993 1994 2000

Norway 1979 1980 1983 1989 1996 2000 2004 2005

Portugal 1982 1983 1986 1988 1992 1995 2002 2006

Spain 1986 1987 1994 1996

Sweden 1981 1983 1984 1986 1987 1994 1996 1997 2004

United Kingdom 1977 1982 1988 1996 1997 1998 2000

uga

1986

1988

1995

Table A2

Expansionary Fiscal Stimuli

Canada 2001

Finland 1978 1987

Greece 2001

Ireland 1974 1975 1978 2001 2007

Italy 1972

Japan 1975

Netherlands 1995

Norway 1974 1991 2007

Portugal 1978 1985

United Kingdom 2001 2002 2003

Expansionary Fiscal Adjustments

Finland 1973 1996 1998 2000

Greece 1976 2005 2006

Ireland 1976 1987 1988 1989 2000

Netherlands 1996

New Zealand 1993 1994 2000

Norway 1979 1980 1983 1996

Portugal

t l

1986

1988

1995

Spain 1986 1987

Sweden 2004

Successful Fiscal Adjustments

Austria 2005

Denmark 2005

Finland 1998

Ireland 2000

Italy 1982

Netherlands 1972 1973 1993 1996

New Zealand 1993 1994

Norway 1979 1980 1989 1996

Sweden 1986 1987 2004

United Kingdom 1977 1988 2000