ArticlePDF Available

Does optimal size of government spending exist?

Authors:

Abstract

Growth theory is an important part of modern macroeconomics. The analysis of growth has long been based on the Solow (1956) "growth accounting" approach, also termed as neo- classical growth theory, which has two important predictions about growth in the long run. These predictions are that economic growth occurs as a result of exogenous technological change, and that income per capita of countries will converge. Since it is presumed that all determinants of growth are exogenous, it is obvious that government policy cannot affect growth rates, except temporarily during the transition of economies to their steady state. Consequently, the role of government in growth process was usually not investigated in
DOES OPTIMAL SIZE OF GOVERNMENT SPENDING
EXIST?
Primož Pevcin
University of Ljubljana
Faculty of Administration
Gosarjeva ulica 5,
SI-1000 Ljubljana, Slovenia
E-mail: primoz.pevcin@fu.uni-lj.si
1. Introduction
Growth theory is an important part of modern macroeconomics. The analysis of growth has
long been based on the Solow (1956) growth accounting approach, also termed as neo-
classical growth theory, which has two important predictions about growth in the long run.
These predictions are that economic growth occurs as a result of exogenous technological
change, and that income per capita of countries will converge. Since it is presumed that all
determinants of growth are exogenous, it is obvious that government policy cannot affect
growth rates, except temporarily during the transition of economies to their steady state.
Consequently, the role of government in growth process was usually not investigated in
standard neo-classical growth models.
The advent of the class of growth models developed by Romer (1986), Lucas (1988), Barro
(1990) and Rebelo (1991), which in essence constitute a new, endogenous growth theory, has
caused that the view on the role of government in growth process changed. According to this
theory, both transition and steady state growth rates are endogenous, implying that also long-
run economic growth rates are endogenous. There are several factors that should be important
for determining long run growth, although in all endogenous growth models, government can
influence growth, either directly or indirectly (see Brons, de Groot and Nijkamp, 1999). As a
result, long-term growth rates can differ across nations, and there is no necessity that
convergence in income per capita should occur. More significantly, as Dar and
AmirKhalkhali (2002) report, a major implication of endogenous growth models is that
government policy can have wide-ranging implications for a country's long-run growth
performance. Namely, the three main fiscal instruments, being taxation, expenditure, and the
aggregate budgetary balance, affect long-term growth through their effects on the efficiency
of resource use, the rate of factor accumulation and the pace of technological progress.
2. Government spending and economic growth
Empirical work on the determinants of economic growth seems to present strong evidence
that a large government sector negatively affects economic growth. This result has been
confirmed in numerous studies (e.g., Barro (1991), Engen and Skinner (1992), Hansson and
Henrekson (1994), Gwartney, Holcombe and Lawson (1998), Fölster and Henrekson (2001)).
More specifically, in recent studies, the negative impact of government size on factor
productivity and capital formation has been stressed, resulting also in lower economic growth.
For instance, Dar and AmirKhalkhali (2002) argue that this adverse impact appears to reflect
the lower productivity of the capital input in countries with a large government size.
Accordingly, the advantage of a small government sector is that, in general, should reflect the
1
greater efficiencies resulting from fewer policy induced distortions and lower tax burden,
more efficient resource use due to the existing market forces, and the absence of crowding-out
effects that impair the incentives for capital creation.1
In order to investigate the relationship between government spending and economic growth,
data for 12 European countries (Austria, Belgium, Denmark, Finland, France, Germany,
Ireland, Italy, the Netherlands, Norway, Sweden and the United Kingdom) were obtained for
1951-1995 time period.2 According to the panel data regression analysis using five-year
arithmetic averages, there is clearly observable negative relationship between the size of
government and economic growth.
Table 1: Relations between government spending and GDP growth, restricted fixed
effect approach3
Dependent variable
Explanatory variables Real growth of GDP (RGR) Change in real growth
of GDP (D(RGR))
Regression constant
(CONST)
7.9391
(0.000)
8.6378
(0.000)
0.1789
(0.179)
Government expenditures
in GDP (EXP)
-0.1076
(0.000)
-0.1155
(0.000)
/
Change in government
expenditures (D(EXP))
/ -0.0678
(0.024)
-0.1428
(0.000)
N 96 84 84
R2adj. 0.5096 0.5531 0.2636
se 1.02 0.95 1.23
d-stat. 2.34 2.08 2.00
F-stat. 50.36 35.23 15.86
P-values are in parentheses.
Source: Own calculations.
The slope coefficient of variable EXP in the first column indicates that a country with a 10
percentage point larger general government expenditures as a share of GDP has, on average,
approximately 1.1 percentage point lower economic growth rate. It seems that government
spending alone would explain more than 50 percent of the differences in economic growth
rates among 12 European countries during the sample period. Moreover, economic growth
also seems to be negatively correlated with the change of government expenditures in last
period, reinforcing the negative effect of government expansion on economic growth. The
above evidence indicates that big government imposes a heavy penalty in the form of a lower
rate of economic growth. Moreover, it can be concluded that the growth rate reductions are
1 Heitger (2001) also investigated this phenomenon and discovered that growth in government size has negative
impact on physical capital formation through the "investment channel", implying that government also crowds
out private investment in physical capital.
2 Data are based on GGDC (2003) (GDP growth) and Cusack (1991/98) (government expenditure ratio).
3 Regression estimates include White’s heteroscedasticity-consistent covariance matrix estimators and first-order
autoregressive adjustment. The results of restricted F test indicate that the restricted fixed effect regression
model should not be invalid.
2
substantially greater for the countries with the largest expansion in the size of government.4
Namely, the regression results add some precision to these findings – there is a strong
negative relationship between the change in government expenditures and change in GDP
growth. For example, the results indicate that a one-percentage point change (increase) in
government spending is associated with an approximately 0.15 percentage point change
(reduction) in real economic growth rate. This provides additional support for the hypothesis
that also growth of government sector reduces economic growth rates.
3. Non-linear relationship between government spending and economic output (the
concept of optimally sized government)
Although it may seem plausible that negative relationship between government size and
economic growth exists, mainstream theory, such as in Barro (1990) or Dar and
AmirKhalkhali (2002), predicts that the negative effect should be expected in countries where
the size of government sector exceeds a certain threshold. The fact that a complex and non-
linear relationship between government spending and growth exists has been first empirically
verified in endogenous growth models. For instance, Barro (1990) pointed out that different
sizes of government have two effects on growth rate. Namely, an increase in taxes reduces
growth rate through disincentive effects, but an increase in government spending raises
marginal productivity of capital, which raises growth rate. He argues that the second force
dominates when the government is small, and the first force dominates when the government
is large. Consequently, the effect of increased government spending on economic growth
should be non-monotonic and some optimal size of government should exist. He showed that
the government services are ”optimally” provided when their marginal product equals unity
(so-called Barro rule). Interestingly, based on empirical findings Barro plotted an inverted U-
shaped curve showing the relationship between growth rate and government expenditure ratio.
The notion of the existence of systematic non-linear relationship between government
spending and economic growth has been later reformulated and popularised in several studies.
For instance, Heitger (2001) views increases in government size arising from increased
consumption as constraints on growth, while increases in size that arise from government
investment should be positive in their effect on growth. His central hypothesis is that
government expenditures on core public goods (such as on the rule of law, internal and
external security, etc.) have a positive impact on economic growth5, but this positive impact
of government tends to decline or even reverse if government further increases expenditures
in a way that it also provides private goods. He stresses that two important reasons for a
negative impact of excessive government spending on economic growth are the fact that the
necessary taxes reduce the incentives to work, to invest and to innovate, and the fact that
government crowds out more efficient private suppliers.
Similarly, Yavas (1998) has shown that an increase in government size will increase the
steady-state level of output if the economy is at low steady-state (i.e., underdeveloped), and
will decrease the steady-state level of output if the economy is at a high steady-state (i.e.,
4 Because the sample is based on countries with relatively similar political structures, incomes, and levels of
development, the potential impact of differences in such things as culture, natural resources, and motivation of
the people is minimised.
5 For instance, Brumm (1997) argues that military spending positively and statistically significantly impacts
economic growth rate by increasing the security of property rights. This calls into some question the notion that
defence spending is unproductive and Kennedy’s (1987) suggestion that increased military spending, which
results in imperial overreach, drains resources out of nations and causes their decline.
3
developed). He argues that in the underdeveloped countries a significant portion of the
government expenditures is directed to the building of the infrastructure of the economy and
this type of government expenditure will have a stimulating effect on private sector
production. In contrast, the developed countries already have most of their infrastructures
built and a major part of their government spending is on welfare programmes and various
social services. Accordingly, the positive effect on spending on these programmes on private
output will not be as great as that of expenditures on infrastructure.
The notion of optimal size of government has been popularised by Armey (1995), who
developed so-called Armey Curve. He argues that non-existence of government causes a state
of anarchy and low levels of output per capita, because there is no rule of law, and no
protection of property rights. Consequently, there is little incentive to save and invest, because
the threat of expropriation exists. Similarly, where all input and output decisions are made by
government, output per capita is also low.6 However, where there is a mix of private and
government decisions on the allocation of resources, output should be larger. Accordingly, the
output-enhancing features of government should dominate when government is very small,
and expansions in governmental size should be associated with expansions in output.
Nevertheless, at some point growth-enhancing features of government should diminish and
further expansion of government should no longer lead to output expansion. Namely, as
spending rises, additional projects financed by government become increasingly less
productive and the taxes and borrowing levied to finance government impose increasing
burdens.7 At some point, the marginal benefits from increased government spending become
zero (point E* in Figure 1).
Figure 1: Government spending and the economy (the Armey Curve)
The general welfare (real
GDP) or rate of economic
growth
g*
E*
Size of government (as a share of GDP)
Source: Armey, 1995.
According to Chao and Grubel (1998), several forces shape afore mentioned inverted U curve.
Namely, they stress that the law of diminishing returns to additional government spending
exists and the additional withdrawal of resources from the private sector more and more
6 As Vedder and Gallaway (1998) stress, the monopolisation of the allocation of resources and other economic
decisions by government usually does not lead to sustained economic prosperity, as too much government stifles
the spirit of enterprise and consequently lowers economic growth. In this context, the revealing should be the
experience of former socialist and communist countries in Europe.
7 For instance, while the construction of main roads as a part of infrastructure initially assists output expansion,
the construction of secondary roads has less added positive impact.
4
occurs at the cost of projects with ever-higher returns. Besides, in order to raise revenue to
finance government spending, taxes have to be imposed, which reduce the private sector’s
incentives to work, save, invest, and take risks. Nevertheless, some of the spending
programmes can also disincentive effects if they lower the risk of economic life.8 These
effects change economic behaviour of individuals, which decrease the effective supply of
labour and entrepreneurship. As Chao and Grubel point out, all these forces reduce economic
growth.
There have been several attempts to determine the level of government spending at which
growth rate is optimised. For example, Vedder and Gallaway (1998) estimated that the
optimal size of federal government spending based on the Armey Curve in the United States
in the period 1947-1997 was 17.45 percent of gross domestic, meaning that federal spending
of about 22 percent at the beginning of 1990s was roughly 20 percent too large from the
standpoint of growth optimisation. Considering general government spending, Peden (1991)
estimated that the optimal size of U.S. government is at 20 percent of GDP. Similar
conclusion were obtained by Scully (1994) who estimated that optimal growth-maximising
average rate for federal, state and local taxes combined was between 21.5 percent and 22.9
percent of gross national product in the United States in the period 1929-1989. In addition,
Chao and Gruber (1998) estimated that in the period 1929-1996 optimal size of government
spending in Canada was about 27 percent, which is about 20 percentage points less that the
actual government spending in 1996. In somewhat different context, by examining socio-
economic indicators, Tanzi and Schuknecht (1998) and Afonso, Schuknecht and Tanzi (2003)
suggest that general government spending in excess of 30 percent of national output reduces
economic growth and produces practically no additional improvement in social measures of
well being.
Besides, some theoretical studies have advocated the use of an allocative efficiency rule to
establish the optimal size and scope of government.9 As Gupta et.al. (2001) stress, the rule
that the size and scope of government are optimal when the social marginal cost of public
resources is equal to their social marginal benefit is difficult to operationalise, although it may
be intellectually appealing. Consequently, alternative rules need to be used to establish
optimal or ”appropriate” size of government, as for instance Armey Curve relationship.
Nevertheless, it should be mentioned that most economists would probably accept the validity
of the inverted-U shape of the Armey curve as a realistic description of the world. However,
in order to validate the curve and get the precise curvature of the line, empirical analysis
should be performed.
4. Evidence on the existence of Armey Curve phenomenon
Statistical testing has shown that the Armey Curve phenomenon exists in the United States,
although it was tested separately for central and local government spending. However, as
Vedder and Gallaway (1998) admit, it is possible that their results could be spurious, as they
potentially do not fully account for other factors that might affect economic growth10, so they
suggest the replication of the test for other countries as different nations have different
8 For instance, social security programmes protecting workers from unemployment, illness, and retirement often
cause them to change their behaviour and reduce work-effort and savings as they are insured.
9 On the theoretical background of this rule see Bailey (1999).
10 For instance, the possibility exists that economic growth is strongly influenced by cycles of innovation and the
rise in the relative size of government may coincide with a slowdown in the rate of innovation for reasons
unrelated to government, as reported by Schumpeter (1934).
5
political environments, different spending histories, and different patterns of change in non-
observable variables, such as the pace and pattern of innovation.
4.1. Data and methodology
Therefore, the purpose of the article is to empirically verify the existence of Armey Curve in
the sample of selected European countries for the 1950-1996 period, using general
government expenditure ratio11. This was the period without wars and the extensive
development of the welfare state in modern times, resulting in the rise of government
spending. The data on general government expenditure as a share of GDP and real gross
domestic product were, similarly to the analysis in chapter 2, obtained for 12 Western
European industrial countries, for which the data after the end of World War II are
continuously available: Austria, Belgium, Denmark, France, Finland, Federal Republic of
Germany, Republic of Ireland, Italy, Netherlands, Norway, Sweden and the United Kingdom.
Compared to most others countries around the world, the institutional arrangements and
income levels of 12 European countries are relatively similar. These countries were politically
stable democracies throughout the whole period after the World War II and 11 of these
countries are now even members of the European Union, so panel data analysis could be
applied. Nevertheless, despite their similarities in many areas, including economic and
political connections and common institutional framework, the size of government sector as a
share of economy has varied substantially among them and across time periods. The data on
real gross domestic product were obtained from GGDC (2003) and the data on general
government expenditure were obtained from Cusack (1991, 1998).12
Since time series and cross section data on general government expenditure and real gross
domestic product were obtained, both of which are supposed to be relevant to this particular
empirical problem, it is reasonable that the econometric methodology is based on pooling of
time series and cross-sectional observations. Since each cross-sectional unit (12 countries) has
the same number of the time series observations (47), balanced panels is estimated. Therefore,
the Armey Curve can be written as:
GDPit = β1 + β2EXPit + β3EXP2it + uit
i = AUT, BEL,..., UK
t = 1950, 1951,..., 1996
It is expected that the linear term, EXP, would have the positive sign and is designed to show
beneficial effects of government spending on output. On the contrary, the squared term, EXP2,
is expected to have negative sign and should measure any adverse effects associated with
11 It is worth noting that several studies use central government expenditure ratio as a measure of the size of
government, probably because this measure is available for more countries. However, as Gwartney, Holcombe
and Lawson (1998) report, this ratio can be highly misleading, because it can underestimate the size of
government for countries where substantial activities are undertaken at lower levels of government, as for
example in Nordic Countries. Consequently, general government expenditure ratio is used in the analysis in
order to test the impact of all budgetary government activities on economic output.
12 The analysis ends in 1996. Namely, the principle is to use data from one source and not to update them with
data from other sources. Besides, in the late 1990s public sector reform (including welfare state reform) and
reduction of government spending has become an important part of policymaking in European countries.
Therefore, the period until mid 1990s, when growth in government spending occurred, is useful in modelling
optimal size of government from growth perspective.
6
increased governmental size. Put differently, this term should indicate the decreasing
marginal productivity of government spending.
4.2. Panel data regression analysis of the Armey Curve
Estimation of panel data regression models depends on the assumptions that are made about
intercept, the slope coefficients, and the error term. First, the simplest, and possibly naive
approach is to disregard the space and time dimension of the pooled data. The results are as
follows.
Table 2: Panel data modelling of Armey Curve13
Dependent variable
D(GDP), all models
include White’s HC
Fixed effect
(FEM) model
Least-squares
dummy variable
(LSDV) model
Error
components
(ECM) model
CONST -10944.56
(0.3899)
Country-specific
intercepts14
-5619.9115
(0.3709)
EXP 1167.84
(0.0543)
539.5015
(0.0341)
686.8285
(0.0175)
EXP2 -15.9709
(0.0157)
-6.7380
(0.0190)
-8.1528
(0.0155)
AR adjustment
(order)
1 1 /
N 540 540 552
R2adj. 0.4626 0.5605 0.5145 (GLS)
se 8251.03 7461.59 7789.97
d-stat. 2.08 1.89 1.36
F-stat. 155.66 350.72 /
Curve peak (EXP as
a share of GDP)
36.56 40.03 42.12
P-values are in parentheses.
Source: Own calculations.
The results suggest that Armey Curve for 12 European countries peaks where the government
spending equals at 36.56 percent of GDP (FEM model), at 40.03 percent of GDP (LSDV
model), or at 42.12 percent of GDP (ECM model). If these results are correct, the size of
government in countries analysed has been too large from the standpoint of growth or output
13 FEM model suggests that the intercept and slope values are identical for all countries (see Gujarati, 2002). As
already mentioned, each country has its political environment or spending history, which means that
appropriate model should also incorporate the individuality of each country. One proposed methodological way
to take into account the individuality of each country is to let the possibility that intercepts vary for each
country but still assume that the slope coefficients are constant across countries (LSDV and ECM model). Due to
the existing problem of non-stationarity in level values of variable GDP, first-order differenced values of
variable GDP are used as dependent variable.
14 Country-specific values of regression constant are as follows: UK= 6673.01, IRL= -6954.54, NL= -2669.33,
BEL= -4588.66, FRA= 10327.06, GER= 14738.87, AUT= -5462.23, ITA=9840.66, FIN= -6169.01, SWE= -
4268.80, DK= -5277.06, NOR= -6162.97.
15 Random error components for each country are: UK= 5079.00, IRL= -9182.20, NL= -4562.90, BEL= -
6850.86, FRA= 8690.07, GER= 13067.52, AUT= -7652.81, ITA=7999.27, FIN= -8607.15, SWE= -6462.65,
DK= -7565.78, NOR= -8343.45.
7
optimisation, since their government amounted, on average, 52.02 percent of GDP in 1996.
Therefore, potential scope for reduction of government spending is from approximately 19
percent up to approximately 30 percent, depending on the estimation results of various panel
data models employed.
However, given large differences in the size of government across countries included in the
sample, some theoretical as well as methodological considerations about panel data estimation
should be expressed. For instance, the results of the Seemingly Unrelated Regression (SURE)
panel data model show that many differential intercept and the differential slope coefficients
are statistically significant, indicating that the Armey Curves of all 12 analysed countries are
obviously not identical. This might suggest that the data for the 12 countries are not
”poolable”, in which case Armey Curve phenomenon should be estimated for each country
separately. In addition, Gupta et.al. (2001, p. 6) argue, although on more theoretical grounds,
that there is no unique optimum level of government spending applicable to all countries.
According to them, the social cost of raising revenues, as well as their social benefits, can be
expected to vary among countries because of political economy factors, such as differences in
voters’ preferences and in the effectiveness of budgetary institutions. Consequently, a change
in the scope of government would necessarily affect the social marginal benefit of some
programmes, and, following, the overall marginal benefit of public spending.
4.3. Time series modelling of Armey Curve
According to the above results, panel data estimation of common Armey Curve for 12
European countries seems to be invalid. Therefore, Armey Curve should be estimated for
each country separately. Methodology of time-series estimation is based on Autoregressive
Integrated Moving Average (ARIMA) modelling.
8
Table 3: Time series modelling of Armey Curve
D. var. (GDP) Italy France Finland Sweden Germany Ireland Netherlands Belgium
CONST -415560.3
(0.0153)
-163202.8
(0.0878)
-3336.61
(0.1611)
-4131.56
(0.2785)
-211235.0
(0.1449)
-11204.78
(0.0203)
-27559.11
(0.0615)
-14771.23
(0.0073)
EXP 24205.31
(0.0180)
8751.52
(0.0522)
264.91
(0.0315)
317.34
(0.0723)
12603.73
(0.0910)
610.0280
(0.0118)
1541.45
(0.0244)
920.8113
(0.0015)
EXP2 -326.33
(0.0278)
-102.01
(0.0476)
-3.40
(0.0238)
-3.45
(0.0635)
-163.90
(0.0794)
-7.1804
(0.0137)
-17.1825
(0.0237)
-10.9868
(0.0017)
ARIMA Adj. (0,1,0) (0,1,1) (0,1,3) (1,1,3) (0,1,0) (1,1,0) (1,1,0) (0,1,0)
N 46 46 46 45 46 45 45 46
R2adj. 0.1411 0.1516 0.2446 0.3124 0.060 0.4660 0.1703 0.1761
se 35009.29 9466.09 1490.15 1705.16 16223.43 757.62 2794.73 2053.72
d-stat. 1.96 1.97 1.58 1.72 1.50 1.96 1.97 2.24
F-stat. 4.70 3.68 5.86 5.00 2.44 13.80 4.01 5.81
P-values are in parenthesis. Only countries with revealed phenomenon are presented in the table.
Source: Own calculations.
9
The results reported in the above table reveal that Armey Curve can be modelled for
eight out of twelve European countries analysed. Those countries are Italy, France,
Finland, Sweden, Germany, Ireland, Netherlands, and Belgium.16 Based on those
estimates, the proper extension of government spending reduction from the standpoint
of output estimation can be derived for each country. The results presented in the table
below indicate that the potentially huge scope exists for spending reduction,
approximately 19 percent on average, which is a huge reduction, but still not so colossal
to leave the government only the role of ”night-watchman”. These results should prove
useful also for other countries, but nothing more or nothing less, as a benchmark value
and not as possible highly sophisticated prescription.17
Table 4: Possible scope for government spending reduction
Country Size of government
(% of GDP, 1996)
Armey Curve
optimum (% of
GDP)
Percentage change
in spending as a
share of GDP
Italy 44.90 37.09 -17.39
France 54.73 42.90 -21.62
Finland 58.74 38.98 -33.64
Sweden 65.02 45.96 -29.31
Germany 48.72 38.45 -21.08
Ireland 39.60 42.28 +6.77
Netherlands 51.97 44.86 -13.68
Belgium 52.97 41.91 -20.88
Average percentage change -18.85
Source: Own calculations.
5. Concluding remarks
A substantial broadening in the scope of government activities occurred in recent
decades in the majority of developed countries, primarily due to the development of
modern welfare states. However, those welfare states have faced with several problems,
especially in the form of efficiency losses from redistribution and disincentives of high
taxation, which have obviously caused the decline of long-term GDP growth. Although
negative and statistically significant relationship between government size and GDP
growth has been established in this and several other studies, mainstream theory
predicts that the negative effect should be expected in countries where the size of
government sector exceeds a certain threshold. Consequently, optimal size of
16 Yet, statistically significant estimates are not being obtained for United Kingdom, Austria, Denmark
and Norway. Interestingly, results even show that the size of government is obviously “too small” in
Ireland. However, this should prove useful, since it indicates that the modelling of Armey Curve is not
biased toward smaller government.
17 For instance, one of the frequently asked questions in connection with public sector reform in Slovenia
is, what is the optimal size of our government sector. It is extremely difficult to answer this question,
especially if the lack of comparable data prior to 1990s (due to the existence of different social and
economic system) prevents the implementation of serious national study of causes of government sector
growth.
10
government sector from GDP growth perspective should exist. The panel data estimates
of Armey Curve suggest that optimal size of government in the sample of 12 European
countries is approximately between 36 and 42 percent of GDP, indicating that potential
scope for reduction of government spending ratio is from approximately 19 to
approximately 30 percent. However, given the fact that large differences in the size of
government across countries included in the sample exist, some theoretical as well as
methodological considerations about panel data estimation occurred. Consequently,
separate time series data estimations are implemented, implying, on average,
approximately 19 percent reduction in government spending ratio.
6. References
Afonso, A., L. Schuknecht and V. Tanzi (2003), Public Sector Efficiency: An
International Comparison, Working Paper No. 242. Frankfurt: European Central Bank.
Armey, D. (1995), The Freedom Revolution, Washington: Regnery Publishing.
Bailey, S. J. (1999), Local Government Economics, Basingstoke: Macmillan Press.
Barro, Robert J. (1990), Government Spending in a Simple Model of Endogenous
Growth, Journal of Political Economy 98(5), 103-125.
Barro, Robert J. (1991), Economic Growth in a Cross-section of Countries, Quarterly
Journal of Economics 106(2), 407-443.
Brons, M., H.L.F. de Groot and P. Nijkamp (1999), Growth Effects of Fiscal Policies,
Tinbergen Discussion Paper, Amsterdam: Vrije Universiteit.
Brumm, Harold J. (1997), Military Spending, Government Disarray, and Economic
Growth: A Cross-Country Empirical Analysis, Journal of Macroeconomics, 19 (4),
827-838.
Chao, Johnny C. P. and Herbert Grubel (1998), Optimal Levels of Spending and
Taxation in Canada, In Herbert Grubel, eds. How to use the fiscal surplus Vancouver:
The Fraser Institute, 53-68.
Cusack, T. R. (1991/1998), The Changing Contours of Government, WZB Discussion
Paper. Berlin: Wissenschaftszentrum Berlin für Sozialforschung.
Dar, Atul A. and Sal AmirKhalkhali (2002), Government size, factor accumulation,
and economic growth: evidence from OECD countries, Journal of Policy Modeling
24(7-8), 679-692.
Engen, E. M. and J. Skinner (1992), Fiscal policy and economic growth, NBER
Working Paper No. 4223. Cambridge (MA): National Bureau of Economic Research.
Fölster, Stefan and Magnus Henrekson (2001), Growth effects of government
expenditure and taxation in rich countries, European Economic Review 45(8), 1501-
1520.
GGDC (2003), Total Economy Database, Groningen Growth and Development Centre,
{http://www.ggdc.net}.
Gujarati, D.N. (2002), Basic Econometrics, New York: McGraw-Hill.
Gupta, Sanjeev, et.al. (2001), Transition Economies: How Appropriate Is the Size and
Scope of Government, Washington: International Monetary Fund.
Gwartney, James, Randall Holcombe and Robert Lawson (1998), The Scope of
Government and the Wealth of Nations, Cato Journal, 18(2), 163-190.
Hansson, Par and Magnus Henrekson (1994), A new framework for testing the effect
of government spending on growth and productivity”, Public Choice 81, 381-401.
11
Heitger, B. (2001), The Scope of Government and Its Impact on Economic Growth in
OECD Countries, Kiel Working Paper No. 1034. Kiel: Institute of World Economics.
Kennedy, P. (1987), The Rise and Fall of the Great Powers, New York: Random
House.
Lucas, Robert E. (1988), On the Mechanics of Economic Development, Journal of
Monetary Economics, 22, 3-42.
Peden, Edgar A. (1991), Productivity in the United States and Its Relationship to
Government Activity: An Analysis of 57 Years, 1929-1986”, Public Choice 69, 153-
173.
Pindyck, R.S. and D.L. Rubinfeld (1991), Econometric Models and Economic
Forecasts, New York: McGraw-Hill.
Romer, Paul M. (1986), Increasing Returns and Long-Run Growth, Journal of
Political Economy, 94, 1002-1037.
Schumpeter, J. A. (1934), Theory of Economic Development, Cambridge (MA):
Harvard University Press.
Scully, G. W. (1994), What Is the Optimal Size of Government in the United States?,
NCPA Policy Report No. 188. Dallas: National Centre for Policy Analysis.
Solow, Robert M. (1956), A Contribution to the Theory of Economic Growth,
Quaterly Journal of Economics, 70(1), 65-94.
Tanzi, Vito and Ludger Schuknecht (1998), Can Small Governments Secure Economic
and Social Well-being, In Herbert Grubel, ed. How to use the fiscal surplus Vancouver:
The Fraser Institute, 69-92.
Vedder, R. K. and L. Gallaway (1998), Government Size and Economic Growth,
Washington: Joint Economic Committee.
Yavas, Abdullah (1998), Does too much government investment retard economic
development of a country, Journal of Economic Studies, 25(4), 296-308.
12
... The extant literature on the relationship between government size and economic growth is vast and has capitulated extensive conflicting outcomes. Some authors found nonlinearity, claiming that the relationship between government expenditure and economic growth is explained by the BARS inverted U-shape (Pevcin, 2004;Rana and Hasan, 2016;Kimaro et al., 2017;Zungu et al., 2020;Nouira and Kouni, 2021;Jain et al., 2021), and others the U-shape (Pelin and Taner, 2017;Phan, 2019). Some authors found it to be a positive relationship (Ghali, 1999;Kimaro et al., 2017;Santos, 2018;Akaakohol et al., 2019), some a negative relationship (Bader and Qarn, 2003;Afonso and Furceri, 2010;Makhoba et al., 2019;Nuru and Gereziher, 2020) and still others found the relationship to be inconclusive (Barro, 1990). ...
... The empirical literature on the relationship between government spending and economic growth is diverse. Some authors found nonlinearity, claiming that the relationship between government expenditure and economic growth is explained by the BARS inverted U-shape (Pevcin, 2004;Chen and Lee, 2005;Rana and Hasan, 2016;Kimaro et al., 2017;Zungu et al., 2020;Nouira and Kouni, 2021;Jain et al., 2021), and others by the U-shape (Pelin and Taner, 2017;Phan, 2019). For some authors it was a positive relationship (Ghali, 1999;Haque and Osborn, 2007;Wu et al., 2010;Kimaro et al., 2017;Santos, 2018;Akaakohol et al., 2019), for others a negative relationship (Bader and Qarn, 2003;Afonso and Furceri, 2010;Makhoba et al., 2019;Nuru and Gereziher, 2020) and some even found it to be inconclusive (Barro, 1990). ...
... Going as far back as the study by Pevcin (2004), the same subject matter was studied using data in 12 European countries over the period 1995-1996. Government expenditure, government gross fixed-capital formation and budget deficits were included in their regression. ...
Article
Full-text available
Purpose –This study is aimed at testing the validity of the BARS theory and determining the threshold level at which excessive government expenditure hampers growth. The data from 10 African emerging economies from 1988 to 2019 were used. Design/methodology/approach–The methodology comprises several different stages. In the first stage, an Lagrange Multiplier (LM) type test was employed to find the appropriate transition variable among all the candidate variables to assess the linearity between government expenditure and economic growth and to find the sequence for selecting the order m of the transition function. The linearity test helped to identify the nature of the relationships between government expenditure and economic growth. In the second stage, the model evaluation was tested using the wild cluster bootstrap-Lagrange Multiplier (WCL-LM) test to assess appropriateness of the model. Thirdly, the Panel smooth transition regression (PSTR) model with one regime was estimated to test the validity of the BARS curve. Findings –The results revealed evidence of nonlinear effects between government expenditure and economic growth, where the size of the government spending was found to be a 27.84% share of GDP, above which government expenditure caused growth to decline in African emerging economies. The findings combined into an inverted U-shape relationship, in line with the BARS theory. Originality/value –This study proposes that policy-makers ought to formulate prudent fiscal policies that encourage expenditure, which would improve growth for selected countries as their current level of spending is below the threshold. This might be done through: (1) a suitable investment portfolio and (2) spending more on infrastructure
... The third group of studies proposes the hypothesis that there is a non-linear nexus between government spending and economic growth, and numerous studies have illustrated a non-monotonic relation between government spending and economic growth (Pevcin 2004;Chen and Lee 2005;Mavrov 2007;Aydin et al. 2016;Iyidogan and Turan 2017;. However, there is a strong paradox in explaining the nature of government spending's influence on economic growth. ...
Article
Full-text available
This study rigorously investigates the non-monotonic phenomenon of the government spending–growth nexus in the EECA. Using panel data from 19 countries over the period 1995–2019, a nonlinear quadratic estimator and cubic nonlinear estimator were applied to quantile regressions. The preliminary findings revealed a negative linear nexus of government spending and economic growth using a linear model, while the nonlinear models (i.e., quadratic and cubic nonlinear estimators) indicated evidence of nonlinearity in the nexus in the EECA over the study period. Furthermore, the study found strong evidence of the existence of an inverted “N-shaped” (nonlinear cubic) nexus between government spending and growth, which can be interpreted as the typical credit-driven boom-and-bust cycle in most EECA countries. Finally, in elucidating the nexus between government spending and economic growth, the study found that most macroeconomic and governance variables are relaxing in explaining GS.
... In the United Kingdom, it is 20.97%, and in the European Union, 41.2%. In the same vein, Pevcin (2004) found that the level of public spending was on average 19% higher than the optimal level of spending in eight developed European countries. ...
Article
This study focuses on the sustainability of public finances in relation to economic growth in Morocco for the period from 1987 to 2019. We set out to explore therewith the non-linear relationship between government size, the level of fiscal discipline and economic growth. This issue at hand has attracted broad public interest and decision-makers' attention in Morocco, especially after the financial crisis of 2008 and during the COVID-19 pandemic. In order to determine government optimal size, we apply the Hansen's approach which postulates the coexistence of different fiscal regimes conditioned by the public debt, government expenditures, and tax revenues in the form of a non-linear inverted-U curve. These regimes are separated by an optimal threshold maximizing economic growth below which the impact is positive and above which the impact becomes negative, as the rising side of the curve is interpreted as consequence of higher taxes providing more resources for public investment, which in turn promotes growth. Once the economy reaches the slippery side of the curve, more taxes and excessive public debt become more distortionary and negatively correlated with economic growth. Our findings indicate that Morocco is relatively in a prudential fiscal stance with recessive effects on growth. ARTICLE HISTORY
... The third group of studies proposes the hypothesis that there is a non-linear nexus between government spending and economic growth, which numerous studies have confirmed (Pevcin (2004) S.-T. Chen and Lee (2005) Mavrov (2007) Aydin, Akinci, and Yilmaz (2016) Iyidogan and Turan (2017) Olumide ). ...
Preprint
Full-text available
This study investigates the non-monotonic phenomenon of the government spending-growth nexus in the EECA rigorously using panel data from 19 counties over the period 1995–2019 by applying a nonlinear quadratic estimator and cubic nonlinear estimator to quantile regressions. The preliminary findings revealed a negative linear nexus of government spending and economic growth using a linear model, while the nonlinear models (i.e., quadratic and cubic nonlinear estimators) indicated evidence of nonlinearity in the nexus in the EECA over the study period. Further, the study found strong evidence of the existence of an inverted “N-shaped” (nonlinear cubic) nexus between government spending and growth, which can be interpreted as the typical credit-driven boom-and-bust cycle in most EECA countries. Finally, in elucidating the nexus between government spending and economic growth, the study found that most macro governance and economic variables are relaxing.
... Pawel Swianiewicz & Herbst (2002)described that the measurement of local government consists of three factors i.e. development capacity, provide service delivery in high quality and affordable and functioning local democracy. Many reserachers argued that the estimation technique on minimal cost of local government and provide the difference on size and scale effect (Pevcin, 2004;King & Ma, 2000). ...
Conference Paper
Full-text available
The number of local governments grew significantly due to the New Law allowed and relaxed to form new local government. Based on the recent data in 2016, the number of Regency are 416. On the other hand, Central Government faces limited fiscal space due to the increase fund transfer to the local government. The aim of this paper is to calculate optimal size of regency government in Indonesia. The scale effect can be measured from the spending of budget on specific sector such as, health, education and infrastructure on the population, area and population density. Econometric analysis will be conducted to identify whether there is an optimal size of regency government among local governments in Indonesia. The result of this study shows that there is sub-optimal size regency government for government spending health, education and infrastructure based on population. The result also shows that inefficiency exists in government spending. Therefore, quality spending should be improved by amalgamation and reorganizing regency government in providing public services.
... This implies that there exists an optimal size of government spending. Pevcin (2002) uses data for 44 European countries for the period 1950-1960 to test the relationship between government size and economic growth, using panel data for 12 countries and regressions over time for each country in 8 out of 12 countries. The results show that the 7/8 Governments in the sample have an excessive spending scale compared to the optimal size. ...
Article
This study examines the impact of public spending on human development in Asian countries, considering both investment and current spending. The empirical method is a system-GMM, using a dataset of 35 Asian countries collected from 2005 to 2014 by the Asian Development Bank (ADB) and the United Nations Development Programme (UNDP). The findings indicated that government spending, both investment and current spending, had effects on the human development index, but these effects are not linear. Depending on the type of expenditure, the detected threshold effect is U-shape or inverted U-shape. According to this study, government spending could adversely impact human growth if the optimal expenditure thresholds are broken. These findings have significant implications for enhancing the effectiveness of public expenditure to improve the human development index. This study also provides meaningful lessons that are especially pertinent for Asian countries, including Vietnam.
... .14% of GDP, Italy 22.23% of GDP, Sweden 19.43% of GDP as well as Great Britain 20.97% of GDP. P. Pevcin[60] also dealt with optimization of the government sector in the EU countries. Based on the data for the years 1970-2007, he found that the optimal GGS, in terms of expenditure, looks as follow: for Italy37.09% ...
Article
Full-text available
The issue of optimal size of the general government sector is analyzed by researchers using various methods, most often through the prism of a specific goal. The article is an attempt to determine the optimal size of the general government sector from the perspective of EU economies. To achieve this goal, the innovative decision tree technique - the c5.0 method was used. The study covered data describing 28 EU member states in the years 2000-2017 and 16,632 input data were analyzed. The results of the conducted research showed that despite the fact that there is no single optimal and universal solution, a series of dependencies can be observed. Knowing the impact of individual actions on the economy, you can choose such instruments, as well as such a configuration that will help in a given area without harming others. Thus, the technique used, combined with specific priorities in terms of impact on the economy, may show which values of specific variables in the general government sector level should be pursued in order to model the desired effect.
... The share of public expenditures for E 1 ( and E 2 ( are E 1 = and E 2 = (1-To check the relationship among public spending and growth, Vedder & Gallaway (1998), Pevcin (2004) and Davies (2009) use simple quadratic equation e.g. GDP= α +β (E 1 ) + ...
Article
Full-text available
This paper discovers the asymmetric relation between government size and economic growth in Malaysia, an emerging Asian economy, by utilising a dynamic threshold nonlinear approach. The study finds empiric evidence that the threshold effect linking government size to economic growth in Malaysia exists when two different proxies of government size are set as the threshold variable. On closer inspection, it is found that an inverted U-shaped Armey curve is present when allowing for endogenous government size threshold proxied by government operating/real GDP. However, a U-shaped curve of nonlinear relationship exists instead when the government investment/real GDP serves as the threshold variable. In general, this paper gives policymakers a real insight into the optimal government size in Malaysia especially when designing an effectual fiscal policy to advance sustainable economic growth to materialise its Shared Prosperity Vision by the year 2030.
Article
This study aims to investigate the short-run and long-run effects of government’s social expenditure proxies, namely education, and health spending on economic growth during the period 1985 - 2019 in West African Economic and Monetary Union. Using Auto Regressive Distributed Lag model (ARDL) based on panel data, the results of the study reveal that in short-run, government spending in social sectors has no significant impact on economic growth but in long-run the effects of education and health expenditures on the economic growth are significantly positive.
Article
Full-text available
This paper examines the development of general government finance and employment within fifteen industrialized democracies during the post-World War II era. Annual time series, extending, where possible, from 1950 through 1988, are analyzed (as well as detailed in the appendix to the paper). Using a variety of accounting schemes, the changing contours of government's control over economic resources are described. It is pointed out that while the growth in government spending relative to GDP has slowed down during the 1980's, the general inability or unwillingness to accelerate the growth in revenues has led in many cases to a significant increase in public debt. This has been accompanied by a squeeze on many spending items, particularly in the area of investment, as interest payments on the debt absorb more and more funds. While growth in government employment has continued, though at a slower pace than in the decades of the '60's and 70's, there has been some marked progress in bringing the cost of public employment into line with what is occurring in the private sector. This latter development is symptomatic of the convergence in price movements that has occurred; during the last decade or so governments have not continued to leap ahead of the private sector, particularly households, in terms of rising costs.
Article
Full-text available
The potential interactions among fiscal policies, investments and economic growth are complex and manifold. In this paper, we will perform a systematic comparative analysis of the various economic insights that are currently available on these complex relationships, both theoretically (by a selective literature review) and empirically (by investigating available data from various countries). Despite the wide variety of potential theoretical relationships between government expenditures, taxation and growth, most empirical analyses are restricted to simple linear regressions of growth on some measure of government expenditures. Based on empirical experiments, we will indicate directions for future empirical research that may enrich our knowledge on the complex relationship between fiscal policies and economic growth, not only nationally but also regionally.
Article
I. Introduction, 65. — II. A model of long-run growth, 66. — III. Possible growth patterns, 68. — IV. Examples, 73. — V. Behavior of interest and wage rates, 78. — VI. Extensions, 85. — VII. Qualifications, 91.
Article
Under the classical supply-side paradigm of productivity grouth, raising government economic above rudimentary levels integrates markets and provides useful public goods which enhance growth. Beyond some point however, increasing government activity discourages growth through tax disincentives and by rewarding dependence on government expenditures. Defining government economic activity to be government expenditures as a percentage of GNP, this study looks first at data for the period 1889–1986 and shows that the U.S. economy has experienced what is predicted by the supply-side paradigm. In addition, the study presents an econometric analysis of government's effects on productivity for the period 1929–1986 using a standard neoclassical growth model. This analysis validates the classical supply-side paradigm and shows that maximum productivity growth occurs when government expenditures represent about 20% of GNP, far less than the 35% which existed in 1986.
Article
Pushed along by the failures of socialism, the idea that a market economy provides the foundation for prosperity has gained widespread acceptance in recent years. Many countries have moved toward an environment more consistent with economic freedom and the smooth operation of a market economy. Trade barriers have been reduced, monetary systems have become more stable, marginal tax rates have been lowered, and various price controls—including exchange and interest rate controls—have been liberalized or eliminated. 1 Yet in one critical dimension—the size of government—most nations have moved in the other direction. Over the past several decades, government expenditures as a share of GDP have been rising, resulting in more resource allocation through government. Table 1 illustrates the growth of government in countries that are members of the Organization for Economic Cooperation and Development (OECD). Data for the 1960–96 period are shown for all 23 nations that have been members throughout the period. Measured as a share of GDP, total government expenditures have grown substantially in every one of the OECD countries. In 1960 government
Article
This report is a feasibility study into an international productivity project using the KLEM growth accounting methodology. It could not have been written without the help of many people. Thanks are due to Bart van Ark, Robert McGuckin, Mary O'Mahony and Dirk Pilat who provided me with helpful comments on earlier drafts of this report. During my visit to the Kennedy School of Government, Harvard University, Mun Ho, Kevin Stiroh and Dale Jorgenson provided me with many insights into the theories and empirical applications of the KLEM-methodology. Discussions with Frank Lee, Wulong Gu and Jianmin Tang (Industry Canada), René Durand (Statistics Canada) and Svend E. Hougaard Jensen, Anders Sørensen, Mogens Fosgerau and Steffen Andersen of the Center of Economic and Business Research (CEBR) provided further insights and new ideas. At the meeting of the KLEM-research consortium on 9 and 10 December at the Centre of Economic and Business Research (CEBR), Copenhagen, consortium members provided me with relevant information, references on ongoing projects and helpful suggestions.Their help is gratefully acknowledged. Prasada Rao (University of New England) is thanked for his advice on the section on purchasing power parities. Finally, the Conference Board is acknowledged for financing the preparation