Download Greece: Military Expenditure, Economic Growth

“Greece: Military Expenditure, Economic Growth
and the Opportunity Cost of Defence”
Emmanuel Athanassiou
Centre of Planning and Economic Research &
Department of Economics, University of Athens
Christos Kollias
Department of Business Administration,
Technological Education Institute of Larissa
Eftychia Nikolaidou
Department of Economics,
Middlesex University Business School
Stavros Zografakis
Department of Economics, University of Athens
Abstract: Greece yearly allocates a substantial share of its national income to defence.
Expressed as a share of GDP Greek military spending is about twice as high as the
European Union and NATO averages. The paper examines empirically the impact of
defence spending on the Greek economy using two distinctly different methodologies.
The first methodology uses an augmented Feder-type model to assess the military
expenditure – growth relationship in the case of Greece. The second methodology
uses a CGE model to estimate the effects on the economy on the basis of the
hypothesis that Greek defence spending followed an equivalent downward path in the
post Cold War era as on average it did in NATO. The obtained results from the
augmented Feder-type model indicate that the defence sector in Greece has had no
positive effect on growth while the CGE estimations suggest that a shift of
expenditure from defence into non-defence public spending would be beneficial for
growth.
Keywords: Greece, defence spending, economic growth, opportunity cost, Feder-type
model, CGE model
1
1. Introduction
Assessing the impact of military expenditure on economic performance is an
area that has been growing fast following Benoit’s seminal contribution that provided
a strong impetus for research, both theoretical and applied, on the subject during the
past two and a half decades approximately. The diversity of the results that have been
reported by the numerous empirical studies is such that not surprisingly no consensus
appears to exist as to the nature and extent of the economic effects of defence outlays
(Dunne, 1996; Ram, 1995; Fontanel, 1990; Gleditsch et al, 1996).
This paper addresses the issue of the economic effects of military expenditure
in the case of Greece. Greece, a member of NATO and the European Union, presents
a particularly suitable vehicle for empirical investigation since for many years it has
been allocating a relatively high proportion of its national income to defence.
Compared to other countries in NATO and the EU its defence burden (military
expenditure as a share of GDP) is often twice as high. At the same time, the Greek
economy has gone through periods of high economic growth as well as periods of
stagnation and fiscal imbalances. Without necessarily implying any correlation or
indeed causation, it is interesting to note that in the past couple of years its growth
rates are among the highest in the EU while its defence burden is twice as high as that
of the EU average.
In the sections that follow we address the impact of defence spending on the
Greek economy using two distinctly different methodologies. The aim is twofold.
First we want to examine the issue of the impact of military spending on the growth
performance of the economy. The results obtained, namely that defence expenditure is
largely neutral vis-à-vis growth, contradicts earlier findings using the same method,
which indicate a negative relationship. Secondly, given the fact that a) the end of the
2
Cold War brought about significant reductions in the defence budgets of many
countries and b) Greek military spending did not exhibit similar downward trends; we
want to estimate the effects on the economy had the reduction in defence spending
been equal to the NATO average. This scenario is used to illustrate the fact that the
defence expenditure growth relationship is affected by changes in regime, but also by
changes in the opportunity cost of defence.
The structure of the paper is as follows. Section two is a brief overview of the
main issues associated with Greek defence spending and the economy with a short
literature review. In section three the defence-growth relationship is estimated for the
period 1960-96 by employing an augmented Feder-type model consisting of the
civilian, military, government and export sector. Then, in section four a Computable
General Equilibrium (CGE) model is used to quantify the forgone benefits had
Greece’s defence spending exhibited an equivalent downward trend to the NATO
average. Finally section five summarises and concludes the paper.
2. Defence and the economy in Greece
With a per capita GDP in 1997 of $13912 Greece, in terms of this traditional
indicator of development, is placed among the developed nations of Western Europe
albeit probably in the lower half of the group being in fact one of the poorest members
of NATO and the European Union. The country is situated at a strategically volatile
region in the crossroads of three continents. The Balkans have traditionally been an
area of friction, tension and conflict and with the collapse of bipolarity the region
entered a period of protracted instability as events over the past decade indicate
(Glenny, 1995; Larrabee, 1992).
3
In the pursuit of national security, Greece has over the years allocated
substantial human and material resources to defence. In comparative terms, Greece is
the most militarised country in NATO and the EU (Kollias, 1995a). Expressed as a
share of GDP Greek military spending has invariably been higher than the EU and
NATO averages. In fact, over the last decade or so, the Greek milex/GDP ratio has
been about twice as high compared to the EU and NATO averages. For example in
the period 1985-97 Greece on average allocated 5.1% of GDP to defence while the
corresponding averages for NATO and the EU were 2.9% and 2.4% respectively
(Figure 1). Furthermore, in a period when most defence budgets have been shrinking
in real terms, Greek defence spending has increased. For example, according to SIPRI
data, in the ten-year period 1989-98 military spending in Greece increased by about
24.2% from $5001 mil. in 1989 to $6211 mil. in 1998 (constant prices). By
comparison, total NATO spending fell by about 26.3% from $601 bil. to $443 bil. and
total EU defence spending fell from $209 bil. to $183 bil. a reduction of about 12.4%
for the same period (constant prices).
4
Figure 1: Military expenditure as a share of GDP in Greece, NATO and the EU
Greece
ΝΑΤΟ
European Union
7
6
5
4
3
2
1
0
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
In many respects, when compared to other EU and NATO members, Greek
security concerns present a unique case that is reflected in the level of resources –
both human and material - the country yearly allocates to defence. The Greek defence
effort that the various indices reflect cannot be explained only in terms of the broader
western security priorities as they have evolved during the bipolar era as well as in the
post-bipolar period. Thus, during the Cold War the country’s external security
concerns were not only the WTO countries but also its neighbour and NATO ally,
Turkey. In fact, Turkey has long been regarded as the main and most imminent source
of external threat to Greek sovereignty and national interests. The Greek-Turkish
conflict is well documented in the international relations literature (Kurop, 1998;
Gurel, 1993; Constas, 1991; Larrabee, 1992). It has also attracted considerable
attention in the defence economics literature. A number of studies that have estimated
demand functions for Greek military expenditure have reported results indicating that
the ongoing conflict with Turkey appears to be an important determinant of Greek
military expenditure (Avramides, 1997; Kollias, 1996; 1995a; Kapopoulos and
5
Lazaretou, 1993). Using causality analysis other studies have empirically investigated
the hypothesis of a Greek-Turkish arms race with mixed findings. Depending on the
methodology, the data and the time period of the study, unidirectional, bidirectional
and no causality have been reported (Majeski, 1985; Majeski and Jones, 1981;
Georgiou, 1990; Kollias, 1991; Stavrinos, 1992; Georgiou et al, 1996; Kollias and
Makrydakis, 1997; Dunne et al, 2000; Smith et al, 2000).
Given the well-documented ongoing Greek-Turkish conflict as well as the
broader security environment of the region, Greece is forced to allocate a substantial
part of its national income to defence at a period when the defence budgets of many
European countries are shrinking. This undoubtedly hinders Greece’s efforts to
achieve economic convergence with other EU members. In fact, Greece was the only
member of the EU that wanted but was not eligible to join the Euro currency area. It
hopes to do so in 2000 when its application will be reassessed.
On a broader level, despite some impressive rates of growth during the first
two and a half post-war decades (Figure 2), the Greek economy has since about the
mid-70s been facing serious structural and fiscal problems (Alogoskoufis, 1995;
Jouganatos, 1992). Only after a number of successive stabilization programs has the
country been set on the road to economic recovery in the last few years, correcting its
fiscal imbalances and reducing the rate of inflation. Currently the economy is
exhibiting relatively high growth rates 3.2% in 1996-97 and 3% in 1997-98 compared
to the EU averages of 2.7% and 2.9% for the same periods.
Figure 2: Military expenditure as a share of GDP and GDP growth rates
12
10
8
6
4
2
6
Given the relatively high defence burden, the economic effects of Greece's
high defence burden is a subject that has been attracting attention in the relevant
literature. The results from the available studies that have attempted to empirically
evaluate and quantify these effects are not conclusive. Again, depending on the
methodology used - adopting single equation or multi-equation models - and the time
period covered by the relevant tests the reported findings are mixed, as Brauer (1999)
notes in a comprehensive survey of the studies. A number of them have shown that
growth has been retarded through various channels while others have reported
stimulative effects through aggregate demand generation (Kollias and Makrydakis,
2000; Antonakis, 1997; 1997a; 1999; Chletsos and Kollias, 1995; Kollias, 1994;
1995; Dunne and Nikolaidou, 1999, Dunne et al, 1999; Nikolaidou, 1999). Studies
have also estimated the potential benefits from hypothesised defence spending
reductions (Balfousias and Stavrinos, 1996; Kollias and Refenes, 1996).
In next sections we address the impact of defence spending on the Greek
economy using, as already pointed out, two distinctly different methodologies. The
defence-growth relationship is estimated for the period 1960-96 by employing an
augmented Feder-type model consisting of the civilian, military, government and
export sector. Following that, a CGE model is used to estimate the forgone benefits to
the economy had Greece’s defence spending exhibited an equivalent downward trend
7
to the NATO average, illustrating the dependence of the defence expenditure growth
relationship on changes of regime and changes in the opportunity cost of defence.
3. Military spending and growth: a multi sector analysis
This part of the paper investigates the defence-growth relationship in Greece
over the period 1960-1996 by employing a commonly used supply-side model, the
Feder-type. Although the Feder-type model can be employed in either the “overall”
form or the “augmented” form in which case externality effects and productivity
differences of each sector are separated from the total (overall) effects, this is not the
concern of the present paper1. In this paper we account for as many economic
linkages as possible by decomposing the economy in four sectors, the civilian, the
military, the government and the export sector while the common approach is to
decompose the economy in two sectors - the civilian and the military). The aim is to
examine the sensitivity of the model to the inclusion of the extra sectors.
1
For an empirical application of the “augmented” Feder-type model for Greece see Antonakis (1997,
1999) and Nikolaidou (1999).
8
3.1. Model Specification
In the Neoclassical framework, supply side models for the defence-growth
relationship have developed from Biswas and Ram (1986) who adopted Feder’s
(1982) model of the role of exports in economic growth, as a two-sector framework
(military and civilian) to assess the externality effect of the military sector and the
factor productivity variation between the two sectors. Since then, many versions of
the Feder model have been developed (ie. assuming different sets of externalities or
more sectors) with most of the studies employing cross-sectional methodologies.2 The
form of the model used here assumes that the economy consists of four sectors
mutually exclusive and exhaustive: the civilian sector (C), the non-military
government sector (G), the export sector (X) and the military sector (M) so that total
output of the economy is the sum of the civilian output, the non-military government
output, the export output and the military output. That is:
Y=C+G+X+M
(1)
Capital and labour are allocated among the four sectors at each point in time. So, that:
K = KC + KG + KX +KM and
L = LC + LG + LX + LM
(2 a, b)
where uppercase subscripts denote the civilian sector (C), the non-military
government sector (G), the export sector (X), and the defence sector (M).
Each of the M, G and X sectors has an externality effect on the civilian (C) sector.
For this approach the production functions for the four sectors are:
G = G (KG , LG)
M = M (KM , LM )
(3 a, b, c, d)
2
For example Ward et al (1991), Ward & Davis (1992), Huang & Mintz (1991), Mintz & Huang
(1990), Atesoglou & Mueller (1990), Mueller & Atesoglou (1993), Ward, Davis and Chan (1993).
9
X = X (KX, LX )
C = C ( KC , LC , G, X, M)
where subscripts C, M, G, X denote sectoral inputs.
Allowing for relative productivity differences between the “base” sector (civilian) and
the other three sectors, ie by (1+∗s24 ), the ratios of the marginal productivities for the
sectors are:
ML /CL = MK / CK = (1 + ∗m)
GL /CL = GK/ CK = (1 + ∗g)
(4 a, b, c)
XL /CL = XK /CK = (1+ ∗x)
where the uppercase subscripts on M, G, X, C denote partial derivatives (or marginal
products) of labour and capital (ie. ML=θM/θLm and MK = θM/θKm). Also, the size of
M, G, X may act as “externality” factors for the civilian sector (C). In other words,
the model also identifies marginal externality effects of each of the three sectors (M,
G, X) on the civilian sector (C). So, we will have:
GK =(1+∗g) CK
and
GL = (1+ ∗g) CL
XK = (1+∗X) CK
and
XL = (1+∗X) CL
MK = (1 + ∗m) CK
and
(5 a, b, c)
ML = (1 + ∗m ) CL
24 where ∗i is the relative factor productivity between the “base” sector and the other
three sectors. If for example the productivity index for defence δm is positive then the
defence sector is more productive than the civilian sector. A zero value for δm would
indicate the absence of a productivity difference while a negative value for δm would
indicate that the civilian sector is more productive. Due to unavailability of sectoral
input data the model is reformulated in terms of aggregate inputs. The equation for
10
this approach that gives the “overall” effect can be derived by manipulating the
production functions (Feder, 1982; Ram,1986;1989; 1995) :
•
Y =α
• é
• G
éæ δ ö
ù • M
éæ δ ö
ù• X
δg
I
÷÷ + CG ú Gæç ö÷ + êç x ÷ + CX ú X æç ö÷ + êç m ÷ + CM ú M æç ö÷ (6)
+ β L+ êçç
Y
êë 1 + δ g
ú è Y ø êëè 1 + δ x ø
ú èYø
ú è Y ø êëè 1 + δ m ø
where α is the marginal product of capital in sector C (civilian), β is the elasticity type
measure equal to CL(L/Y), and [δi/1+δi +Ci] is the sum of the externality and factor
productivity differences (the overall effect of sector i on economic performance).
(δi/1+δi ) alone is the relative productivity effect of the i sector on economic growth
while Ci alone represents the marginal externality effect of the i sector on the civilian
sector (Huang and Mintz 1991). Adding an intercept and a disturbance term, gives the
equation to be estimated for Greece over the period 1960-96 to get the size (total)
effect of each of the sectors on economic growth. Note that in order to estimate
equation 6 we replace the instantaneous change rate of variables with their discrete
•
equivalents (ie. Y =∆Y/Y-1 ). The four-sector Feder-type model for Greece specified
above, defines simple externalities from each of the defence, export and non-military
government sector only on the base sector (civilian sector), derives values for nonmilitary government spending by deducting military spending from total government
spending, provides estimates for a two, three and four-sector model in order to
compare the results each time a sector is added. As such, it overcomes a number of
shortcomings3 associated with this type of model.
3
For a discussion of the shortcomings and advantages of the Feder-type model, see Nikolaidou (1999).
11
3.2. Data and Empirical Findings
The military expenditure data used here is drawn from SIPRI Yearbooks.
Usually labour force data is not available and its growth is proxied by the population
growth rate (Ram, 1986; Ward et al, 1991; Alexander, 1990). For Greece, labour
force data was in fact available from 1970 onwards, with labour force growth proxied
by population growth prior to 1970.
Data for GDP, investment, government
expenditure, exports and labour force were taken from the OECD database. All
figures were first deflated in constant 1990 million drachmas and then converted to
1990 million US $ by means of exchange rates. To measure non-military government
spending military expenditure was subtracted from government expenditure. This
overcomes the problem exemplified by Alexander (1990) where government
consumption was used as a proxy for non-military government consumption leading
to an overvaluing of government consumption by the amount of military expenditure
(the variables are described in the Appendix).
Table 1 shows the results for the total (overall) effects of each sector (equation
6). The first column gives results from the two-sector model (military and civilian),
the second column from the three-sector model (civilian, military and government)
and the third column gives results from the four-sector model (civilian, military,
government and export).
12
Table 1. Estimation Results (1961-1996)
Dependent GDP Growth
Indep.Var.
Eq. 6
Eq. 6
Eq. 6
Intercept
-0.03 (1.23)
-0.03 (1.15)
-0.03 (1.29)
I / Y−1
0.33 (3.06)***
0.27 (2.51)**
0.26 (2.39)**
∆L / L−1
-0.92 (2.50)**
-0.95 (2.71)***
-0.83 (2.21)**
∆M / M −1 ( M / Y−1 )
1.59 (1.84)*
0.43 (0.42)
0.37 (0.36)
∆G / G−1 ( G / Y−1 )
------
1.04 (1.92)*
1.14 (2.05)**
∆X / X −1 ( X / Y−1 )
------
------
0.49 (0.90)
R2
0.41
0.47
0.48
SE
0.03
0.02
0.03
DW
1.66
1.74
1.54
F-stat
F(3,32)=7.3***
F(4,31)=6.85***
F(5,30)=5.61***
X2(1)=0.06 [.801]
X2(1)=0.01 [.918]
X2(2)=9.74 [.008]
X2(1)=1.08 [.299]
X2(1)=0.03 [.861]
X2(1)=0.05 [.830]
X2(2)=10.94[.004]
X2(1)=0.18[.675]
X2(1)=0.63 [.427]
X2(1)=0.05 [.825]
X2(2)=8.80 [.012]
X2(1)=0.45 [.501]
Diagnostic Tests
Serial Cor.
Funct.Form
Normality
Heterosc.
The top columns give the coefficients estimates followed by the t-ratios (in parentheses) while the
bottom columns give the X2 tests for Serial Correlation, Functional Form, Normality and
Heteroskedasticity followed by the probabilities (in brackets). ***: significant at 1% level of
significance,**: significant at 5% level of significance, *: significant at 10% level of significance. For
all estimations Microfit 4.0 was used.
Looking at equation 6, that describes the total (size) effects of each sector on
economic growth, it is obvious that the overall performance of the model in terms of
explanatory power is not very satisfactory with the R2 being 0.41, 0.47, 0.48 for the
two, three and four sector models respectively. But another consideration when
evaluating the overall performance of this model, concerns the coefficients on the
investment and labour variables since as Mintz and Stevenson (1995) point out “In
general, one would be more confident in the specification of the model if the
coefficients on these variables conform to the standard predictions of the economic
13
theory”. If for example investment is found significantly negatively related to
economic growth, the validity of the model should be questioned. But this is not the
case for any of the three specifications of the model for Greece in all of which
investment is positive and highly significant. Labour force growth on the other hand
has an unexpected significantly negative effect, which is problematic as it does not
conform to the standard predictions of the economic theory although the theories
underlying the impact of labour on the economy are less conclusive than that of
investment. Its significance might suggest that, in Greece, increases in the workforce
do not necessarily imply a more productive workforce. In fact, this is not an unusual
finding as Ward et al. (1991) argue. Furthermore, Antonakis (1997) in his two-sector
model for Greece found a negative but insignificant effect for the labour force4
variable justifying it on the grounds that in labour surplus economies like Greece, the
natural rate of growth is not a binding constraint. As far as the total effect of the
military sector is concerned (in the 2 sector model) it is positive and significant at
10% supporting the modernisation and spin off arguments for defence spending.
On the second column of Table 1, the government sector (excluding the
military) enters the equation with a positive sign, significant at 10%. As for the effect
of the military sector, it is still positive but now insignificant. The intercept,
investment and labour force growth continue to have the same signs as before with
their significance slightly altered. Finally, by adding the export sector (see column 3
of Table 1), all of the variables’ signs remain the same, the significance of the
government sector increases (from 10% to 5%) and the significance of investment and
labour force growth drops slightly to 5%. The effect of the export sector is positive
4
Antonakis (1997) as most other studies have done, proxied labour force growth by population growth.
But this can cause the impact of labour growth to be underestimated, especially in cases where the size
of the labour force changes significantly while population remains stable (almost stable population is a
fact for Greece).
14
but not significant, which is not surprising for a country like Greece that mainly
exports agricultural products. As for the constant term, which measures an average
trend rate of technological progress, it is insignificant in all three specifications.
Non-nested tests indicated that the three and four-sector models are preferred
to the two-sector one, while no clear cut preference could be made among the three
and four-sector models (as the Akaike’s information criterion was in favour of the
four-sector model while the Schwarz Bayesian criterion was in favour of the threesector one). Given this, plus the fact that results are very similar for the three and
four-sector models, reliance on either model is acceptable. It would appear that the
military sector in Greece does not have a significant impact on economic growth, and
that the same applies to the export sector. Only the non-military government sector
seems to be growth promoting in Greece. This is in contrast with earlier findings5
although this may be due to differences in the time periods used or the difference in
outcome may be due to a change in regime and a resulting change in the opportunity
cost of defence (Murdoch et al. 1997).
4. Change of Regime and the Opportunity Cost of Defence:
As pointed out, the results concerning the effect of defence expenditure on
growth of the first part of the paper conflict with earlier findings using the same
methodology. It has been suggested that the relationship between public defence,
public non-defence and non-public expenditure on the one hand, and growth on the
other, may be affected by changes in regime, that will in turn affect the opportunity
cost when undertaking different types of expenditure. In this section we attempt to
illustrate this possibility through the use of a CGE model, trying to capture what the
5
Antonakis (1997) reports a strong negative effect
15
change of regime represented by the end of the Cold War would have had on the
Greek economy, had no other threat to security existed.
In order to circumvent the numerous problems associated with public good
pricing, we make the simplifying assumption that decisions are taken in such a way so
that the change in the slope of the production possibilities frontier reflects the relative
marginal costs in undertaking a particular pattern if expenditure. Thus a measure of
the opportunity cost would be the difference in the growth rate owing to a change in
the pattern of exogenous expenditure. In this part of the paper, we attempt an analysis
of the structural incidence of a hypothetical reduction of current defence spending, as
opposed to investment and labour expenditures, to the average NATO level.
Furthermore we obtain results for alternative substitute expenditure patterns. Given
that there has been practically no reduction in the level of current defence expenditure
in Greece since 1988, the calculations consist in the calculation of a set of opportunity
costs cum regime change effects reflecting the particularity of Greek defence needs.
The tool of analysis is a Computable General Equilibrium model for Greece,
(Zografakis, 1997; Sarris and Zografakis, 1997; Dervis et al, 1982) which
incorporates the national accounting relationships of the 1988 Social Accounting
Matrix (SAM) for Greece. The choice of base year for the analysis although dictated
by the availability of data is fortuitous since it corresponds to the beginning of the
dissolution of the command economies of eastern Europe, and the resultant defusion
of East-West tensions. The SAM framework permits the inclusion in the analysis of
multiplier effects that work through the distribution of income as well as through
transfers between institutions. This is particularly important for countries like Greece
16
where the bulk of armaments procurements is of foreign provenance, thus the direct
structural effect of such spending is minimal.
The SAM adopted here, disaggregates the national accounts into fifteen
sectors6. The labour input is divided into four categories according to skill,
agricultural, unskilled, skilled and highly skilled. There are two classes of
employment possible, salaried and self employed. Labour is assumed to be perfectly
mobile between sectors of economic activity. Capital is sector specific and due to the
comparative static nature of the analysis is kept fixed at the initial level. In what
concerns the aims of this paper the government sector net of health and education
expenditures was broken into two components, defence and non-defence. The model
comprises the real sectors of the economy only, excluding the financial sector. Thus
interest rates enter as a parameter into the analysis. The functioning of three types of
market regulates production and exchange, that for goods, inputs and the foreign
sector, so that prices and quantities are endogenous to the model. Commodities
offered in the domestic market are composites of domestically produced and imported
commodities. (see Appendix IIa). Aggregation is made through use of a CET
function, the domestic and imported goods being considered imperfect substitutes
according to the Armington specification (Armington, 1969). The same specification
is adopted in order to determine the composition of production between exports and
domestically offered goods.
Domestic prices and wages are determined endogenously, and are determined
by market clearing equations. The exchange rate is adopted as the numeraire, and thus
set to unity. Prices of imports are taken as given, following the small country
6
See Athanasiou et al (1998)
17
assumption. Labour is offered as a function of the real wage rate, private investment is
determined by the return on capital while public investment is a policy variable.
Demand of commodities is formed by private consumption, investment in dwellings,
government expenditure and investment, private investment, stock variations and
exports. These demand elements are allocated over the set of commodities.
Production technology in the model is of the CES type. A multi-layered nested
CES format allows the consistent aggregation of the large number of disaggregate
inputs by stages, per sector of economic activity and type of input (see Appendix II).
Thus, at the highest level of aggregation the production function accepts two inputs, a
quantity of labour - intermediate goods and a quantity of capital - energy, each being
the result of aggregation of the different specific types of inputs. There are three
closure rules that characterise the model and encompass the assumptions about the
working of the economy. The first concerns employment and is of the neoclassical
variety, thus full employment of labour is assumed. That for investment is of the
Keynesian variety, with the investment rate determining the savings rate. Finally, the
external sector is dominated by exogenous transfers from abroad reflecting the
importance of EU and other non-market transfers to the Greek economy.
In summary the model exhibits the following characteristics:
1.
It considers explicitly market clearing mechanisms, and related price formation,
in the economy; prices are computed by the model as a result of supply and
demand interactions in the markets, in which economic agents are price takers.
2.
It formulates separately the supply or demand behaviour of the economic agents
in the individual optimisation of their objectives, within markets that are cleared by
prices that achieve global equilibrium.
18
3.
The model exhibits a large degree of disaggregate detail concerning sectors,
social groups, structural features and policy oriented instruments (e.g. taxation).
The figures for defence spending were taken from the 1989 budget reporting
realised expenditures for 1988, and were subtracted from the government
consumption figures in the original SAM matrix. Elasticities used in the model are
based on estimates undertaken for the HERMES7 econometric model and GEM-E3
(Computable General Equilibrium Model for studying Economy-Energy-Environment
Interactions for Europe8)
4.1 Results of Simulations
The question we seek to answer is what would the effects on the Greek
economy be if the reduction in current defence expenditure (i.e. excluding labour and
investment expenditure) were equal with the average decrease for NATO countries
over the period 1988 - 1996. In fact, since current defence expenditure in Greece has
practically remained constant for the period in question, the question is what the peace
dividend would be if a considerable reduction in this type of expenditure had taken
place.
We examined five different counterfactual cases. In the first, current defence
expenditure was reduced by about 25%, the NATO average for the period under
examination. Since the interest rate in our model is exogenously given, the private
sector cannot compensate for the reduction in public spending, so this case can be
viewed as an initial impact scenario. In fact we use it as a benchmark case to assess
7
Harmonised Econometric Research for Modelling Economic Systems, Edited by: Commission of the
European Communities, North-Holland, 1993.
19
the influence of this type of defence spending on the structure of the economy, in
order to better be able to interpret the results of the other scenarios. In the second case
we consider the effect of a decrease in current defence spending by the same amount
as above, which however is compensated by expenditure on public consumption in
such a way as to leave the shares of the various non-military consumption items
constant. Non military public consumption is divided into three categories, education,
health and other government. The three last cases consist in exactly compensating the
reduction in military expenditure by increases in expenditure of each of these
categories alone.
Given that the composition and share in GDP of government current
expenditure in Greece has remained practically constant over the period in question,
we are in fact considering a two tier “what - if” situation. First, we estimate the peace
dividend, as if Greece faced a level of national security similar to that of the average
NATO country. Second we examine alternative uses of this peace dividend, and hence
the opportunity cost of defence spending, always remaining within the constraint that
current public expenditure remains constant.
Table 2:
Effects on National Accounting Aggregates (percent deviation from reference)
A
Defence
-24.9%
B
Defence
-24.9%
Gov.Cons.
(non def.)
7.156%
C
Defence
-24.9%
Gov.
Education
D
Defence
-24.9%
Gov. Health
+40.05%
E
Defence
-24.9%
Other
Gov. Cons.
+13.14%
GDP f. p.
-0,45%
0,19%
+26.4%
0,26%
-0,10%
0,24%
Private Investment
-1,51%
0,67%
0,84%
1,48%
0,39%
Total Exports
0,81%
-0,40%
-0,45%
-0,81%
-0,29%
8
The GEM-E3 model was built under the auspices of European Commission (DG-XII, co-ordinator
P.Valette) by a consortium involving CORE, NTUA, KUL, Univ. Mannheim, Univ. Strathclyde and
CEA.
20
Total Imports
-3,27%
0,39%
0,53%
1,75%
0,02%
Private Consumption
-0,53%
0,37%
0,50%
0,43%
0,33%
Consumer Price Index (CPI)
-4,71%
2,59%
2,87%
5,74%
1,79%
Real Wage Rate
-1,07%
1,06%
1,17%
1,28%
1,01%
Employment (Thousands)
-14,67
11,55
13,29
12,73
11,16
In Table 2, column A, we see the impact of a once for all reduction in current
defence spending to the NATO average level. GDP would decrease by about half a
percentage point. The GDP components would all decrease except exports which
would benefit. Consumer prices would decrease by close to 5%, a figure that may be
compared to the actual increase of the CPI for the period ranging from 20% at the
beginning and 10% at the end of the period. Real wages would drop by about a
quarter of the drop in the CPI. The drop in real wages would be higher the higher the
skill level of the work force. Employment would decrease, the lower skill categories
decreasing relatively more.
Comparing the four alternative cases, we note that if the compensation is such
that the shares of the non-military expenditure categories are kept constant at the 1996
composition, the results fall in between the cases where the totality of the
compensation is imputed to one category alone. This is not surprising due to the
homogeneity assumptions made in the construction of the model. Generally speaking,
expenditure on education seems to be the best alternative, followed by expenditure in
other government. Expenditure on health is the worst alternative, and in many cases is
worse than military expenditure itself. Of course both these types of expenditure
affect welfare by reducing two important types of risk, and this is not taken into
consideration by the static, non-stochastic framework adopted here.
21
The effect of compensating the decrease in current defence expenditure would
add 0.2 of a percentage point to GDP from 1996 onwards. This is about 10% on the
actual growth rate for that year. This growth could increase to 0.26% if the
compensation was accomplished by increase in education alone, or become a 0.1%
decrease if health alone takes up the slack. Expenditure in other government services
would result in almost as high a gain in GDP as education, but with a lower increase
in private consumption and private investment. An increase in GDP similar to that of
education is achieved by the relative smaller impact that this combination of
expenditure will have on both exports and imports. This scenario will also be kindest
to price increases but also will result in the lowest increase in real wages.
Expenditure on health comes off badly in growth enhancing terms due to the
increased requirements for imports, and the relatively important decrease in exports.
Private investment would increase the growth rate the most, while the rise in private
consumption would be slightly less than that seen in the case of education. The effect
on prices would be considerable, while that on wages would also be greater than that
of the case of education. A significant result is that current defence spending seems to
be directed more towards the domestic economy than any other type of expenditure. It
is the only category of government expense that is positively related to exports and
negatively to imports. Given that Greece is not a major arms exporter, it would seem
that current defence spending is directed to domestic goods, which are complementary
in their factor requirements to exportables.
5. Conclusions
The results obtained from the application of the Feder methodology for the
case of Greece for the period 1960-95 indicate that the effect of defence spending on
22
growth is neutral, in contrast to non-defence public spending which has a positive
effect on growth. This is in contrast to earlier findings, which showed that defence
spending has a negative effect on growth. The change in the estimated coefficients
may well depend on a change in regime (or set of regimes) due to the different time
spans covered. While the difference in the results is interesting in itself, the policy
implications do not change. A shift of expenditure out of defence and into nondefence public spending would seem to be beneficial for growth. Using a CGE model
we examined such a case. In contrast to the two cases mentioned above we used data
for defence expenditures excluding expenditures relating to weapons systems
procurement. Our base year is 1988, which coincides with the beginning of the
change in the security regime as represented by the collapse of the Warsaw Pact
threat. The results indicate that there is a positive relationship between transfers of
expenditure from defence to non-defence public expenditure and growth. This
supports the findings of the Feder type applications to the case of Greece.
Furthermore, while transfers from defence to non-defence expenditure may be
favourable to growth, it is less so than when the transfer is exclusively to general
government expenditure, or expenditure on education, while when expenditure on
health substitutes for defence the effect is negatively related to growth. It follows then
that policy regimes, as represented by priorities between types of public expenditure
would also be operative in determining the relationship of defence expenditure on
growth.
References
Alogoskoufis, G. (1995). “The two faces of Janus: institutions, policy regimes and
macroeconomic performance in Greece”. Economic Policy, No 20, pp. 149192.
23
Alexander, R. (1990) "The impact of defence spending on economic growth",
Defence Economics, Vol. 2, pp. 39-55.
Alexander, R. (1995) "Defence spending: burden or growth promoting?" Defence and
Peace Economics, Vol. 6, pp. 13-25
Antonakis, N. (1999), ‘Guns Versus Butter: A Multisectoral Approach to Military
Expenditure and Growth with Evidence from Greece, 1960-1993’, Journal of
Conflict Resolution, Vol 43 (4), pp.501-520
Antonakis, N. (1997a), "Defence spending and growth in Greece: a comment and
further empirical evidence", Applied Economics Letters, Vol. 4, pp. 651-655.
Antonakis, N. (1997), "Military expenditure and economic growth in Greece 196090", Journal of Peace Research, Vol. 34 (1), pp. 89-100
Armington, P., (1969), "A Theory of Demand for Products Distinguished by Place of
Production", IMF Staff Papers 16 (1). 159-78.
Atesoglou, H.S and M.J. Mueller (1990) "Defence spending and economic growth",
Defence Economics, Vol. 2 (1), pp. 19-28
Athanassiou, E., C. Kollias, and S. Zografakis (1998) “A computable general
equilibrium estimation of the opportunity cost of defence in the case of Greece”,
mimeograph
Avramides, C. (1997) “Alternative models of Greek defence expenditure”, Defence
and Peace Economics, Vol 8 (2), pp. 145-187
Balfousias, A. and V. Stavrinos (1996) "The Greek military sector and
macroeconomic effects of military spending in Greece", pp. 191-213 in N.P.
Gleditsch et al (eds), The Peace Dividend, North Holland Publications
Biswas, B. and Ram, R. (1986) ‘Military Expenditure and Economic Growth in
LDCs: an Augmented Model and further Evidence’, Economic Development
and Cultural Change, Vol 34 (2), pp.361-372
Chletsos, M. and C. Kollias (1995) "Defence spending and growth in Greece 1974-90:
some preliminary econometric results" Applied Economics, Vol. 27, pp. 883890
Constas, D. [ed] (1991) The Greek-Turkish conflict in the 1990s, Macmillan
Dunne, P., Nikolaidou, E. & Smith, R. (2000), ‘Arms Race Models and Econometric
Applications’, Forthcoming in Levine, P., Sen, S. & Smith, R. (eds) (2000), The
Arms Trade, Security and Conflict, Harwood Academic Publishers
Dunne, P. & Nikolaidou, E. (1999), ‘Military Spending and Economic Growth in
Greece: A Demand and Supply Model, 1960-1996’, Defence and Peace
Economics (forthcoming).
Dunne, P., Nikolaidou, E. and Vougas, D. (1999), ‘Military Spending and Economic
Growth in Greece and Turkey: A Causal Analysis, 1960-1996’, Defence and
Peace Economics (forthcoming).
Dunne, P. (1996) “ Economic effects of military expenditure in developing countries:
a survey”, pp. 439-464, in in N.P. Gleditsch et al (eds), The Peace Dividend,
North Holland Publications
Feder, G. (1982) ‘On Exports and Economic Growth’, Journal of Development
Economics, Vol 12, pp.59-73
Fontanel, J. (1990) "The economic effects of military expenditure in Third World
countries", Journal of Peace Research, Vol. 27, pp. 461-466
Georgiou, G. (1990) “Is There an Arms Race between Greece and Turkey? Some
Preliminary Econometric Results”, Cyprus Journal of Economics, Vo 3(1), pp.
58-73.
24
Georgiou, G., P. Kapopoulos and S. Lazaretou (1996) “Modelling Greek-Turkish
Rivalry: An Empirical Investigation of Defence Spending Dynamics”, Journal
of Peace Research, Vol. 33(2), pp. 229-239.
Gleditsch, N.P., O. Bjerkholt, A. Cappelen, R. Smith and P. Dunne (1996), The Peace
Dividend, North Holland Publications
Glenny, M. (1995) “Heading off war in the Southern Balkans”, Foreign Affairs, Vol
74(3), pp. 98-108)
Gurel, S. (1993) “Turkey and Greece: a difficult Aegean relationship”, in A. Balkir
and C. Williams [eds], Turkey and Europe, Pinter Publishers
Huang, C. and A. Mintz (1991) "Defence expenditures and economic growth: the
externality effect", Defence Economics, Vol. 3 (1), pp. 35-40
Jouganatos, G. (1992). The Development of the Greek economy 1950-1991,
Greenwood Press.
Kapopoulos, P. and S. Lazaretou (1993) “Modelling the demand for Greek defence
expenditure: an error correction approach”, Cyprus Journal of Economics, Vol.
6 (1), pp. 73-86.
Kollias, C. and A. Refenes (1996) "Modelling the effects of defence spending
reductions using neural networks: evidence from Greece", Peace Economics,
Peace Science and Public Policy, Vol. 3, pp. 1-12
Kollias, C. and S. Makrydakis (1997) "Is there a Greek-Turkish arms race? Evidence
from cointegration and causality tests", Defence and Peace Economics, Vol. 8,
pp. 335-379
Kollias, C and S. Makrydakis (2000) “A note on the causal relationship between
defence spending and growth in Greece: 1955-93”, Defence and Peace
Economics, (forthcoming)
Kollias, C. (1996) “The Greek-Turkish conflict and Greek military expenditure 196092”, Journal of Peace Research, Vol. 33(2), pp. 217-228.
Kollias, C. (1994) "The economic effects of defence spending in Greece 1963-90:
some preliminary econometric findings", SPOUDAI, Vol. 44, pp. 114-130
Kollias, C. (1995) "Preliminary findings on the economic effects of Greek military
expenditure", Applied Economic Letters Vol. 2, pp. 16-18
Kollias, C. (1995a) "Country Survey VII: Military spending in Greece" Defence and
Peace Economics, Vol. 6, pp. 305-319
Kurop, M. C. (1998) “Greece and Turkey. Can they mend fences”, Foreign Affairs,
Vol 77 (1), pp. 7-12
Larrabee, S.F. (1992) "Instability and change in the Balkans", Survival, Vol 34(2), pp.
31-49.
Majeski, S. (1985) "Expectations and arms races", American Journal of Political
Science, Vol. 29(2), pp. 217-245
Majeski, S. and D. Jones (1981) "Arms race modeling. Causality analysis and model
specification", Journal of Conflict Resolution, Vol. 25(2), pp. 259-288
Mintz, A. & Stevenson, R. (1995), ‘Defence Expenditures, Economic Growth, and the
“Peace Dividend”. A Longitudinal Analysis of 103 Countries’, Journal of
Conflict Resolution, Vol 39 (2), June, pp.283-305
Mintz, A. & Huang, C. (1990), ‘Defence Expenditures, Economic Growth and the
Peace Dividend’, American Political Science Review, Vol 84, pp.1283-1293
Mueller, M.J. & Atesoglou, H.S. (1993), ‘Defence Spending, Technological Change,
and Economic Growth in the United States’, Defence Economics, Vol 4 (3),
pp.259-269
25
Murdoch J.C., Sandler, J., (1997), “The Impact of Defence and Non-Defence Public
Spending on Growth in Asia and Latin America”, Defence and Peace
Economics, Vol. 8 (2,) pp 205-225
Nikolaidou, E. (1999), ‘Military Spending and Economic Growth in Greece, A Multi
Sector Analysis’, Economics Discussion Paper, Middlesex University Business
School, No. 72, May
Ram, R. (1995), ‘Defence Expenditure and Economic Growth’, Ch.10 in Hartley, K.
& Sandler, T. (eds), Handbook of Defence Economics, Elsevier: Amsterdam
Ram, R. (1986), ‘Government Size and Economic Growth: A New Framework and
some Evidence from Cross-Section and Time-Series Data’, American Economic
Review, 76 (1), pp.191-203
Sarris A.H., S. Zografakis (1999), “A Computable General Equilibrium Assessment
of the Impact of Illegal Immigration on the Greek Economy”, Journal of
Population Economics, Volume 12, Number 1, pp.155-182.
Smith, R., Dunne, P. & Nikolaidou, E. (2000), ‘The Econometrics of Arms Races’,
Defence and Peace Economics (forthcoming)
Smith, R. (1980) "Military expenditure and investment in OECD countries 1954-73",
Journal of Comparative Economics, Vol. 4, pp. 19-32
Stavrinos, V. (1992) “Defence Expenditure in Arms Competition: Modelling and
Causality Analysis”, Greek Economic Review, Vol. 14(1), pp. 115-128.
Ward, M., Davis, D. & Chan, S. (1993), ‘Military Spending and Economic Growth in
Taiwan’, Armed Forces and Society, Vol 19 (4), pp.533-550
Ward, M. & Davis, D. (1992), ‘Sizing up the Peace Dividend: Economic Growth and
Military Spending in the US, 1948-1996’, American Political Science Review,
Vol (3), pp.748-755
Ward, M., Davis, D., Penubarti, M., Rajmaira, S. & Cochran, M. (1991), ‘Country
Survey I: Military Spending in India’, Defence Economics, Vol 3 (1), pp.41-63
Zografakis, S. (1997) "Economic Policy and Impacts on the Evolution of Income
Distribution in Greece: An Analytical Approach based on a Computable
General Equilibrium Model", unpublished PhD Thesis, Department of
Economics, University of Athens, Greece.
26
APPENDIX I
Description of variables in the Feder-type model
•
Y = ∆s24 Y/Y-1 = GDP growth
I / Y = I/Y-1 = Share of investment in GDP
•
ng1032
L = ∆L/ L-1 = Labour force growth
•
M
÷ = ∆M/ M-1 ( M/Y-1
Y
• G
Gç ÷
Y = ∆G/ G ( G/Y ) = Total effect of non-military government sector
Mç
4 )= Total effect of defence sector
-1
•
Xç
-1
X
÷ = ∆X/ X-1 ( X/Y-1) = Total effect of export sector
Y
•
C
M ç ÷ = ∆M/ M-1 (C/Y-1)= Externality effect of military sector
Y
• C
æ ö
Gç ÷ = ∆G/ G-1 ( C/Y-1) = Externality effect of non-military government sector
èY
• C
X ç ÷ = ∆X/ X-1 (C/Y-1) = Externality effect of export sector
Y
where C=Y-M in the two-sector model
C=Y-M-G in the three-sector model
C=Y-M-G-X in the four-sector model
δm= productivity difference of the military sector with respect to the civilian
sector
δg= productivity difference of the government sector with respect to the b
civilian sector
δx= productivity difference of the export sector with respect to the civilian
sector
and Y=GDP in constant 1990 mn US $
M=Military Expenditure in constant 1990 mn US $
G=Government Expenditure (excluding military) in constant 1990 mn US $
X=Exports in constant 1990 mn US $
tab I=Private Investment in 1990 mn US $
L=Labour Force in ‘000s
27
APPENDIX IIa
The structure of Consumption
Disposable Income
Consumption
Investment in Dw ellings
Saving
expend. for cars
Durable Goods
expend. for domestic appliances
Non-Durable Goods
Food,
beverages,
tobacco
Medical care and health expenses
Products
Services
Education and Culture
Energy
Clothing,
footwear
Recreation, Entertainment
Heating Systems
Others goods
and services
Communication Services
Electricity
Transport Services (purchased transport)
Vehicle (Petrol for cars)
28
APPENDIX IIb
The structure of production
Production function
product
Aggregate:
Capital - Energy
CES function
Capital
Aggregate:
Labor - Materials
CES function
Aggregate
Labor
Energy
Employed Labor
CES aggregate function
skilled labor - Urban
Materials
(Factor demand Interm.
Consump. of non-Energy
Products)
Self Employed
CES aggregate function
skilled labor - Urban
semi-skilled labor - Urban
semi-skilled labor - Urban
unskilled labor - Urban
unskilled labor - Urban
in agricultural sector
29
in agricultural sector