New evidence on fiscal adjustment and growth in transition economies.
Segura-Ubiergo, Alex ; Simone, Alejandro ; Gupta, Sanjeev 等
INTRODUCTION
There is consensus in the literature on the crucial role played by
initial conditions, macroeconomic stabilization, and structural reforms
on growth patterns in transition economies. (1) The consensus on
macroeconomic stabilization is reflected in that most studies of growth
performance in transition economies include variables that measure the
impact of inflation and generally find a negative relationship between
inflation and growth. (2) However, there has been no conclusive evidence
on fiscal adjustment and growth in transition economies. As Havrylyshyn
(2001) noted, 'the empirical literature is nearly unanimous on the
negative impact of inflation on growth but has not been able to
disentangle the separate effects of fiscal deficits and inflation on
growth'. In fact, the fiscal balance is much less widely used in
cross-country studies of growth performance (eg Fischer et al. (1996,
1997); Wolf, 1999; Fischer and Sahay, 2000; Berg et al., 2006). (3) A
recent study (Purfield, 2003) does not find a clear relationship between
fiscal adjustment and growth in transition economies.
This empirical evidence seems at odds with the widely held view
that fiscal adjustment is a necessary condition for growth when fiscal
deficits threaten macroeconomic stability. Agenor and Montiel (1999)
survey the theoretical literature and conclude that fiscal adjustment is
crucial for attaining macroeconomic stability. Excessive fiscal deficits
may lead to inflation, balance of payments difficulties, external debt
crises, and high real interest rates, outcomes that tend to reduce
economic growth. Case studies in the literature also show that permanent
fiscal adjustment is critical to uproot inflation that hurts growth.
Moreover, Havrylyshyn and van Rooden (2005) and World Bank (2002) argue
that growth during economic transition is likely to be driven by vast
resource reallocation, efficiency improvements, and increased
international trade and investment, and thus policy instruments that
facilitate these changes, such as fiscal adjustment, can strongly affect
growth even after accounting for the initial conditions and
institutions.
The empirical literature on fiscal policy and growth also shows
that healthy budget balances are good for growth over the long run.
While the effect of fiscal adjustment in the short run remains open to
question, a large body of empirical research finds salutary long-term
effects of fiscal adjustment. (4) However, a number of studies in
industrial countries have found that improving fiscal positions can
stimulate growth even in the short run. A central theme in these works
is that the composition of fiscal adjustment plays a key role in
determining whether fiscal contractions sustain and support higher
growth over time. (5) Gupta and et al. (2005a, b) show that these
results can largely be extended to emerging market and low-income
countries.
This paper focuses on whether fiscal adjustment affects growth in
transition economies. This issue is analyzed using a sample of 26
transition economies covering the period 1992-2001. Building on the
existing literature discussed above, this paper also contributes to the
discussion of whether the relationship between fiscal adjustment and
growth in transition economies is different from other country groups.
The results suggest that the impact of fiscal adjustment on growth
in transition economies is robust and not qualitatively different from
what has been observed in industrial, emerging market, and low-income
economies. The positive correlation between fiscal adjustment and growth
is stronger for countries that need to achieve macroeconomic stability.
In particular, the sustained fiscal consolidation efforts in the
Commonwealth of Independent States (CIS) have been accompanied by
accelerations in growth in the sample. (6) However, for countries that
have already achieved macroeconomic stability, the relationship between
fiscal adjustment and growth is less clear. (7) A fixed-effects panel
error-correction model that controls for initial conditions, structural
reforms, and inflation confirms the strong positive correlation of
fiscal adjustment and growth. In addition, the positive effect of fiscal
adjustment holds in the short and long run.
The remainder of the paper is organized as follows: the next
section describes the empirical methodology and the data. The following
section discusses some stylized facts on growth and fiscal adjustment in
transition economies. Finally, we discuss the main econometric results.
The last section concludes.
EMPIRICAL METHODOLOGY AND SAMPLE
Overall approach
The overall approach in conducting the empirical test starts with a
general framework that covers all key factors that affect growth in
transition economies. These factors are based on the existing
literature, which includes variables that measure initial conditions,
institutions, and degree of macroeconomic stabilization. As shown in
Havrylyshyn et al. (1998) and Havrylyshyn and van Rooden (2003), these
are the three main categories of factors in explaining growth in
transition economies. For example, Krueger and Ciolko (1998) and Popov
(2000) find that initial conditions explain substantial variations in
growth performance during transition, and institutional capacity is also
a key determinant of growth; Havrylyshyn and van Rooden (2003) argue
that policies matter in addition to institutions in affecting growth;
while Fischer et al. (1996, 1997) focus on stabilization policies and
find that combating inflation is an effective policy instrument to
promote growth. Therefore, controlling for key indicators in all three
dimensions is important for uncovering the potential impact of fiscal
adjustment on growth.
The analysis of the relationship between fiscal adjustment and
growth in this paper is carried out in two stages. The first stage
provides a simple characterization of the data of fiscal adjustment and
growth in transition economies to establish some stylized facts. The
second stage relies on panel estimators to test the effect of fiscal
adjustment on growth and motivates the results by discussing briefly
their consistency with country experiences. Panel estimators are the
most appropriate choices for growth regressions as they can explore data
variations across countries and over time (eg Baltagi, 2001). The
regressions control for other factors that could also affect the
relationship between fiscal adjustment and growth to minimize omitted
variable bias. For example, in countries rich in natural resources, most
notably oil or gas, it is conceivable that favorable terms of trade
could be simultaneously associated with a deficit-reducing increase in
fiscal revenues and a large positive impact on growth. In these cases,
the observed relationship could be spurious as in Azerbaijan,
Kazakhstan, and Russia.
Furthermore, the panel-data approach used in the second stage aims
to minimize several other econometric issues that arise in the analysis
of transition economies. (8) These problems are the endogeneity between
GDP growth and fiscal balance, multicollinearity and parameter
heterogeneity. In particular, this paper mitigates the endogeneity
problem by using several estimators with increased control on
endogeneity, addresses multicollinearity by examining exclusion effects
with the relevant test statistics, and tackles parameter heterogeneity
by focusing on samples and sub samples of countries that share some
theoretical similarities. (9, 10)
The scope and quality of data is also a challenge in transition
country analysis. As noted in many papers, GDP series may have been
underestimated in some countries, especially at the beginning of the
transition. At the same time, the definition of the fiscal deficit may
vary from country to country (central government, general government,
nonfinancial public sector, etc). Hence, the meaning of 'fiscal
adjustment' may not be the same in all countries. This paper uses
consistent data for general government in the World Economic Outlook
(WEO) database to minimize this problem. Additional consistent data
series were obtained from the World Bank's World Development
Indicators (WDI) database and from the European Bank for Reconstruction
and Development (EBRD) economic database.
THE ECONOMETRIC MODEL
Preliminary considerations
Many previous papers had used either cross-sectional data (without
examining variations over time) or a relatively short time series with
fairly simple estimation techniques, and often without including fixed
effects. In contrast, in this paper panel regressions with fixed effects
are used to obtain unbiased and consistent estimates. (11) Transition
economies are characterized by rapid changes whose very nature is likely
to have varied significantly from one country to the next. The exclusion
of fixed effects ignores these country-specific differences and
increases the risk of omitted-variables bias. In addition to these
theoretical reasons, this is ultimately an empirical question. If
regression tests show that fixed effects belong in the model, they
should be included. (12) Given the focus of this paper on the impact of
fiscal adjustment on growth, as long as the error term is not correlated
with the fiscal and other control variables, it is not a problem if the
impact of other variables is negated by the inclusion of fixed effects.
The results from multiple panel estimators that provide different
controls for the endogeneity between growth and fiscal balance indicate
that, once fixed effects are controlled for, endogeneity is no longer a
problem. This is consistent with the good small sample property of
fixed-effect panel ordinary least squares (OLS) as shown in Islam
(2000).
Model and sample
The paper analyzes the relationship between fiscal deficits and
real GDP growth using a sample of 26 transition countries in 1992-2001.
(13) The basic regression framework is described by a one-equation
error-correction model. (14)
[DELTA][Y.sub.i,t] = [alpha] + [DELTA][X.sub.i,t-1][[beta].sub.k] +
[phi]([Y.sub.i,t-1] - [X.sub.i,t-1][gamma]) + [[epsilon].sub.i,t] (1)
where, [Y.sub.i,t] is real GDP in country i during year t, [DELTA]
is the first differences operator, X is a vector of independent
variables and [[epsilon].sub.i,t] is a white noise error term. Although
a number of different specifications have been used, a threevariable
vector accounts for over 70% of the variation in the dependent variable.
The model describes a short-term equilibrium relationship given by
[DELTA][Y.sub.i,t] = [alpha] + [DELTA][X.sub.i,t-1][[beta].sub.k] +
[[epsilon].sub.i,t] and a term [phi] ([Y.sub.i,t-1] - [X.sub.i,t-1]
[gamma]), which measures the deviation from this short-term equilibrium
relationship. Equation 1 shows that a change in [X.sub.i,t-1] produces a
short-term contemporary change in [Y.sub.i,t] that is measured by the
k-dimensional vector of regressors [beta]k. In addition, when the impact
of [X.sub.i,t-1] on [Y.sub.i,t] throws the model off its long-run
equilibrium given by the cointegrating vector [Y.sup.*.sub.i,t-1] =
[X.sup.*.sub.i,t-1][gamma] where the '*' indicates
equilibrium, the discrepancy or 'error' ([Y.sub.i,t-1] -
[X.sub.i,t-1][gamma]) is corrected at a yearly rate of [phi]. (15)
In order to estimate equation 1, it is useful to restate it through
a simple mathematical operation: Let [[beta].sub.j] be defined as -
([phi][gamma]), where both [phi] and [gamma] come from equation 1; then
it follows that [gamma] = [[beta].sub.j]/-[phi]. Equation 1 can
therefore be rewritten as:
[DELTA][Y.sub.i,t] = [alpha] + [Y.sub.i,t-1][phi] +
[DELTA][X.sub.i,t-1][[beta].sub.k] + [X.sub.i,t-1][[beta].sub.j] +
[[epsilon].sub.i,t] (2)
And equation 2 can then be estimated through panel OLS. The
variables used in equation 2 are summarized in Table 1.
The selection of variables included in the empirical analysis
follows the existing literature on growth in transition economies. The
model is relatively simple and includes variables measured both in
'levels' and 'first-differences'. It also includes
the European Bank for Reconstruction and Development (EBRD) structural
reform index, the general government overall fiscal balance, inflation,
and the external current account balance. (16,17) Unlike other studies,
variables that measure differences in initial conditions are not
included, but these country-specific factors will be captured by the
inclusion of fixed effects in most of the models. As noted above, the
inclusion of fixed effects is also key to reduce the possible
omitted-variable bias.
GROWTH AND FISCAL ADJUSTMENT IN TRANSITION COUNTRIES: SOME STYLIZED
FACTS
Both growth and fiscal balances improved significantly in
transition economies over the past decade. For the purposes of this
analysis, the sample can be divided in two 5-year periods: (i)
1992-1996, which can be broadly considered the 'decline'
period; and (ii) 1997-2001, which can be described as the
'recovery' period. (18) Between these two periods, the average
growth rate passed from about -2.0% to 4.5%, while the average fiscal
deficit was cut in half from about 7.0% to about 3.5% of GDP. Although
these averages mask important differences across countries, Table 2
provides compelling evidence that the higher growth trend and lower
fiscal deficit have been quite general. In particular, growth improved
in over 75% of the cases, whereas fiscal deficits were reduced in about
two-thirds of the cases.
There also seems to be a strong correlation between fiscal
adjustment and growth. Before moving to the results of the more
sophisticated econometrics, we examine some simple descriptive patterns
underlying the data. Figure 1 plots the change in the overall fiscal
deficit and the change in the growth rate between the transition and
post-transition periods. It shows that (i) fiscal adjustment was
generally associated with higher growth; and (ii) the effect is greater
the higher the initial level of the deficit. When the adjustment exceeds
about 10% of GDP, the impact on growth begins to decline, as can be
shown by the concavity of the curve.
The positive relationship between fiscal adjustment and growth is
particularly strong for the CIS. Table 2 shows that large fiscal
adjustment cases in the sample have predominantly occurred in CIS
countries. (19) In the Eastern European countries, only Bulgaria had a
fiscal adjustment comparable to that of a CIS country. (20) With the
exceptions of Hungary and the Slovak Republic, most Central European
countries had already achieved relatively low fiscal deficits by 1992.
All the Baltic countries had deficits below 3% of GDP by 1992.
[FIGURE 1 OMITTED]
The correlation between fiscal adjustment and growth is unclear in
countries that had relatively low deficit levels at the beginning of the
period. Table 2 shows that in the Central and Eastern European group,
countries where the fiscal deficit increased from a relatively low
position experienced lower growth, as, for example, in the case of the
Czech Republic. The opposite seems true in the Baltic countries, which
are generally viewed as leaders in the implementation of structural
reforms and macroeconomic stabilization. Other factors including the
composition of spending may have played a role as well.
These results become evident if countries are divided along two
criteria: the magnitude of fiscal adjustment and the degree of growth
acceleration. The matrix diagonal of Table 3 shows cases of either low
adjustment-low growth acceleration or high adjustment-high growth
acceleration. About 75% of the cases fall into this category. But there
is also an important group of countries with low fiscal adjustment and
high growth acceleration. This corresponds mostly to cases in three
situations: (i) 'early adjusters' such as the Baltic
countries, where most of the adjustment had taken place before 1992,
hence preceding the trends captured in the sample; (ii) countries where
growth is affected primarily by changes in the international price of
specific commodities; or (iii) outliers such as Albania. Mongolia is the
only case of high fiscal adjustment and low growth acceleration. (21,22)
Fiscal adjustment was largely sustained through time and appears
not to be driven by cyclical fluctuations. Coefficients of variation of
fiscal deficits of countries were modest, suggesting relative stability
through time. In addition, there does not seem to be strong evidence of
significant cyclicality driving fiscal adjustment as revenue and
expenditure elasticities with respect to GDP are low. (23)
ECONOMETRIC MODEL RESULTS
The use of more sophisticated econometric techniques confirms the
strong relationship between fiscal adjustment and growth. The results
presented in Table 4 show that fiscal adjustment is positively
associated with growth both in the short run as captured by the
first-difference variable and in the long run as captured by the lagged
variable in levels.
The positive association of fiscal adjustment and growth is robust
to a variety of model specifications and estimation techniques. With the
exception of Model 9--the Arellano-Bond estimator without including
inflation as a regressor--all parameter estimates for the lagged level
and change in the fiscal deficit are statistically significant, and in
most cases at the 1% level of significance. (24)
The results suggest that the positive impact of fiscal adjustment
is rather large, both in the short and long run. (25) Over the long run,
controlling for everything else, if country A maintained a fiscal
deficit 1% of GDP lower than country B, long-term growth would be higher
by between 0.5 and 1 percentage points, depending on the estimation
model. (26) The effect of fiscal adjustment on growth in the short term
also seems to be positive. (27) Consistent with the previous section,
this relatively large effect seems to be driven by CIS countries. In
fact, if we split the sample into CIS and non-CIS countries, the
regression coefficient for the CIS sample is much larger than that for
the non-CIS one. (28) Figure 2 captures the different effects in both
samples by plotting changes in the fiscal balance (x-axis) versus GDP
growth (y-axis), controlling for other factors in the model. The
contrast in the slope of the partial regression line is striking.
[FIGURE 2 OMITTED]
The effect of the other variables is similar to that found in other
studies. A reduction in inflation has a positive and highly significant
effect on growth. Progress in reforms as measured by the institutional
reform index of the EBRD also has a strong impact on growth. However,
the inclusion of both inflation and the reform index in the same model
wipes out the statistical significance of the reform index. This is not
surprising, as inflation stabilization is one of the components of the
reform index, and including both in the same equation leads to
multicollinearity. An adjustment in the current account balance is
associated with a significant decline in growth in one of the model
specifications (Model 5), but it loses statistical significance once
inflation is included as a regressor.
The findings from the econometric analysis show strong associations
between fiscal adjustment and growth, particularly in CIS countries.
Such findings are largely consistent with country experiences discussed
in the literature. As shown in World Bank (2002), at the onset of
transition, the highly distorted production structures, the severe terms
of trade shock, and the sudden cutoff of financing from Russia generated
an output collapse. The combination of the output collapse with the cut
in fiscal transfers generated large fiscal deficits in many countries.
Difficulties to cut public spending in a short time period and limited
or non-existent access to domestic or foreign debt financing in the
first years of transition forced most of the deficit to be monetized,
thus leading to high rates of inflation. Countries also resorted to
involuntary financing by running arrears, particularly on the external
front. The large fiscal deficits were quickly associated with growing
and unsustainable current account imbalances (Table 5).
Fiscal adjustment in this context helped anchor agent's
expectations and attract capital inflows. Garibaldi et al. (2001) show
that rising foreign direct investment was the main form of private
capital inflow to transition economies, (29) and these investments were
strongly influenced by sound macroeconomic fundamentals. Similar effects
were also found for remittances. (30) Official flows also rose sharply.
As a result, Helbling et al. (2004) document, for example, a large
increase in external debt after 1992 in the CIS-7, mostly in public and
publicly guaranteed debt (above 80%). (31)
This evidence suggests that fiscal adjustment may have created a
strong domestic demand effect that helped achieve growth recovery in the
CIS. Fiscal adjustment can provide strong signals for sustained and
credible government reform and thus create incentives for private agents
to fully explore new economic opportunities, particularly those enabled
by increases in capital inflows. This is consistent with findings from
Helbling et al. (2004) who suggest that increased capital inflows
financed higher consumption; Loukoianova and Unigovskaya (2004) also
support this result, and argue that only after 1998 did net export start
to drive growth. But structural reforms that fostered better
market-support institutions could have reinforced the effects of fiscal
adjustment on growth. Further research using detailed country
experiences could shed additional light on the causal mechanism through
which fiscal adjustment can boost economic growth in each particular
context.
CONCLUSIONS
The main result of this paper is that the impact of fiscal
adjustment on growth in transition economies is not qualitatively
different from the one found for industrial, emerging markets, and
low-income economies.
A simple bivariate correlation analysis suggests a strong positive
correlation between fiscal adjustment and growth in transition
economies. In particular: (i) the correlation between fiscal adjustment
and growth seems stronger the higher the initial level of the deficit;
and (ii) when fiscal adjustment exceeds about 10% of GDP, the positive
impact on growth begins to decline. A fixed-effects panel
error-correction model that controls for initial conditions, structural
reforms, and inflation confirms the strong positive correlation of
fiscal adjustment and growth. In addition, the positive effect of fiscal
adjustment holds in the short and long run.
The correlation between fiscal adjustment and growth is stronger
for those countries that need to achieve macroeconomic stability. The
largest fiscal adjustments in the sample have been observed
predominantly in the CIS, since these countries had unsustainable fiscal
positions at the beginning of the 1990s. The sustained consolidations
efforts in the CIS have been accompanied by the largest accelerations in
growth in the sample.
However, for countries that have already achieved macroeconomic
stability, the correlation between fiscal adjustment and growth is less
clear. As Adam and Bevan (2005) have shown, this may be because other
factors such as the composition of expenditures or deficit financing
play a bigger role, or because these countries could invest in
structural reforms that have moderate up-front fiscal costs that yield
significant long-term benefits. Further research in this area is
therefore needed. (32)
These results suggest that countries that have managed to credibly
stabilize and achieve fiscal sustainability are unlikely to get large
benefits in terms of growth by pursuing additional fiscal adjustment.
For example, addressing microeconomic distortions with structural
reforms to increase the rate of return to investment to attract capital
flows, improving the composition of government spending while
maintaining a prudent fiscal stance, and strengthening governance are
likely to be more important policies to generate growth in that context.
The above results highlight the importance of a medium term exit
strategy to reverse ongoing expansionary fiscal policies adopted in
response to the current financial crisis. Several countries in the
sample have responded to this crisis by allowing automatic stablizers to
operate and by implementing discretionary fiscal stimulus. This has
caused a significant deterioration in their fiscal balances, endangering
fiscal sustainability and macroeconomic stability. Unless corrective
action is taken by them to restore fiscal sustainability, there are
serious risks to growth over the medium term.
Acknowledgement
The authors thank for their valuable comments two anonymous
referees, Messrs. Flood, Chami, Flickenschild, Saadi-Sedik, Grigorian,
Ding, Abiad, Gueorguiev, Klingen, Mody, and Szekely, and Mmes. Corbacho,
Allard, Farhan, and Ohnsorge.
REFERENCES
Adam, C and Bevan, D. 2005: Fiscal deficits and growth in
developing countries. Journal of Public Economics 89: 571-597.
Agenor, P and Montiel, P. 1999: Development macroeconomics.
Princeton University Press: Princeton, NJ.
Alesina, A and Ardagna, S. 1998: Tales of fiscal adjustment.
Economic Policy 27(October): 487-546.
Alesina, A and Perotti, R. 1997: Fiscal adjustments in OECD
countries: Composition and macroeconomic effects. IMF Staff Papers
44(2): 210-248.
Alesina, A, Perotti, R and Tavares, J. 1998: The political economy
of fiscal adjustments. Brookings Papers on Economic Activity 1: 197-266.
Alesina, A, Ardagna, S, Perotti, R and Schiantarelli, F. 2002:
Fiscal policy, profits, and investment. American Economic Review 92(3):
571-589.
Arellano, M and Bond, S. 1991: Some tests of specification for
panel data: Monte Carlo evidence and an application to employment
equations. Review of Economic Studies 58: 277-297.
Aslund, A, Boone, P and Johnson, S. 1996: How to stabilize: Lessons
from post-communist countries. Brookings Papers on Economic Activity 1:
217-313.
Baltagi, B. 2001: The econometrics of panel data, 2nd Edition. John
Wiley & Sons: New York.
Berg, A, Borensztein, E, Sahay, R and Zettelmeyer, J. 2006: The
evolution of output in transition economies: Explaining the differences.
In: Berglof, E and Gerard, R (eds). The Economics of Transition: The
Fifth Nobel Symposium in Economics. Palgrave Macmillan: Houndmills.
Brunetti, A, Kisunko, G and Weder, B. 1998: Credibility of rules
and economic growth: Evidence from a worldwide survey of the private
sector. World Bank Economic Review 12(September): 353-384.
Buti, M and Sapir, A (eds). 1998: Economic policy in EMU: A study
by the European Commission Services. Oxford University Press: New York.
Cheng, K. 2003: Growth and recovery in Mongolia during transition
IMF working paper no. 03/217, Washington.
Christoffersen, P and Doyle, P. 2000: From inflation to growth:
Eight years of the transition. Economics of Transition 8(2): 421-451.
De Melo, M, Denizer, C and Gelb, A. 1996: From plan to market:
Patterns of transition. World Bank Economic Review 10 (September):
397-424.
Easterly, W, Rodriguez, C and Schmidt-Hebbel, K. 1994: Public
sector deficits and macroeconomic performance. Oxford University Press:
New York.
Fischer, S and Sahay, R. 2000: The transition economies after ten
years IMF working paper no. 00/30, Washington.
Fischer, S and Sahay, R. 2004: Transition economies: The role of
institutions and initial conditions Paper presented at a conference in
honor of Guillermo A. Calvo, Washington.
Fischer, S, Sahay, R and Vegh, C. 1996: Economies in transition:
The beginnings of growth. American Economic Review 86(2): 229-253.
Fischer, S, Sahay, R and Vegh, C. 1997: from transition to market:
Evidence and growth prospects. In: Zecchini, S (ed) Lessons from the
Economic Transition: Central and Eastern Europe in the 1990s. Kluwer:
Dordrecht, 79-101.
Garibaldi, P, Mora, N, Sahay, R and Zettelmeyer, J. 2001: What
moves capital to transition economies? IMF Staff Papers 48(Special
Issue): 109-145.
Gelbard, E, McHugh, J, Iradian, G, Beddies, C and Redifer, L. 2005:
Growth and poverty reduction in Armenia: Achievements and challenges.
International Monetary Fund: Washington.
Gupta, S, Clements, B, Baldacci, E and Mulas-Granados, C. 2005a:
Fiscal policy, expenditure composition, and growth in low-income
countries. Journal of International Money and Finance 24(April):
441-463.
Gupta, S, Baldacci, E, Clements, B and Tiongson, E. 2005b: What
sustains fiscal consolidation in emerging market countries?
International Journal of Finance and Economics 10(4): 304-321.
Hausmann, R, Pritchett, L and Rodrik, D. 2005: Growth
accelerations. Journal of Economic Growth 10(4): 303-329.
Havrylyshyn, O. 2001: Recovery and growth in transition: A decade
of evidence. IMF Staff Papers 48(Special Issue): 53-87.
Havrylyshyn, O and van Rooden, R. 2003: Institutions matter in
transition, but so do policies. Comparative Economics Studies 45 (1):
2-24.
Havrylyshyn, O, Berengaut, J, de Castello Branco, M,
Mercer-Blackman, V, Van Rooden, R and Wolf, T. 1998: Growth experience
in transition countries 1990-1998 IMF Occasional paper no. 184,
Washington.
Helbling, T, Mody, A and Sahay, R. 2004: Debt accumulation in the
CIS-7 countries: Bad luck, bad policies, or bad advice? In: Shiells, C
and Sattar, S (eds). The Low-income Countries of the Commonwealth of
Independent States. International Monetary Fund and World Bank:
Washington, 15-49.
Hernandez-Cata, E. 1997: Liberalization and the behavior of output
during the transition from plan to market. IMF Staff Papers 44(4):
405-429.
Hsiao, C and Sun, B. 2000: To pool or not pool panel data. In:
Krishnamukar, J and Ronchetti, E (eds). Panel Data Econometrics: Future
Directions. Elsevier Science B.V.: Amsterdam.
Islam, N. 2000: Small sample performance of dynamic panel data
estimators in estimating the growth-convergence equation: A Monte Carlo
study. Advances in Econometrics 15: 317-339.
Kaufman, R and Segura-Ubiergo, A. 2001: Globalization, domestic
politics and social spending in Latin America: A time series cross
section analysis. World Politics 53: 553-587.
Krueger, G and Ciolko, M. 1998: A note on initial conditions and
liberalization during transition. Journal of Comparative Economics 26:
718-754.
Loukoianova, E and Unigovskaya, A. 2004: Analysis of recent growth
in low-income CIS countries IMF working paper no. 04/151, Washington.
McDermott, J and Wescott, R. 1996: An empirical analysis of fiscal
adjustments. IMF Staff Papers 43(4): 725-753.
Patillo, C, Gupta, S and Carey, K. 2005: Sustaining growth
accelerations and pro-poor growth in Africa IMF working paper no.
05/195, Washington.
Popov, V. 2000: Shock therapy versus gradualism: The end of the
debate (explaining the magnitude of transformational recession).
Comparative Economic Studies 42(1): 1-57.
Purfield, C. 2003: Fiscal adjustment in transition countries:
Evidence from the 1990s IMF working paper no. 03/36, Washington.
Rzonca, A and Cizkowicz, P. 2005: Non-Keynesian effects of fiscal
contraction in new member states ECB working paper no. 519, European
Central Bank: Frankfurt.
Segura-Ubiergo, A. 2007: The political economy of the welfare state
in Latin America: Globalization, democracy and development. Cambridge
University Press: New York.
Von Hagen, J and Strauch, R. 2001: Fiscal consolidations: Quality,
economic conditions, and success. Public Choice 109: 327-346.
Wolf, H.C. 1999: Transition strategies: choices and outcomes.
Princeton studies in international Finance No. 85.
World Bank. 2002: Transition--The first ten years. World Bank:
Washington.
(1) Transition economies are those that were initially organized on
the basis of government ownership of the factors of production and
central planning and changed their economic organization to market based
systems. The sample in this paper includes economies in Central and
Eastern Europe, the Baltics, Commonwealth of Independent States, and
Mongolia. See Table 2 for a detailed list of countries in each group.
(2) See, for example. Aslund et al. (1996); Berg et al. (2006);
Brunetti et al. (1998); Christoffersen and Doyle (2000); De Melo et al.
(1996); Fischer et al. (1996, 1997); Fischer and Sahay (2000);
Havrylyshyn et al. (1998);. Havrylyshyn and van Rooden (2003);
Hernandez-Cata (1997).
(3) For example, Fischer and Sahay (2004), who focus on the role of
institutional reforms in development, mention in passing that fiscal
adjustment is associated with higher growth. Their growth regressions
show a substantial and statistically significant positive coefficient
for the fiscal adjustment variable.
(4) See, for example, Easterly et al. (1994).
(5) These studies show that improving fiscal positions through the
rationalization of the government wage bill and public transfers rather
than increasing revenues and cutting public investment can foster higher
growth even in the short term. See, for example, McDermott and Wescott
(1996); Alesina and Perotti (1997); Alesina et al. (1998); Alesina and
Ardagna (1998); Buti and Sapir (1998); Alesina et al. (2002); and Von
Hagen and Strauch (2001).
(6) Hausmann et al. (2005) use the term 'growth
accelerations' to describe episodes where the per capita growth
rate increases by more than 2 percentage points a year and is sustained
for at least 8 years. By looking at jumps in country medium-term trends
they expect to gain insight into the sources of successful growth
transitions. While the paper does not follow this exact definition of
growth acceleration and methodology, the concept of acceleration being
used is consistent with a definition of this type.
(7) Other factors, such as the composition of spending, are more
likely to be the bottlenecks for growth in these countries. See Patillo
et al. (2005) for evidence on the importance of the composition of
fiscal spending on growth in low-income countries.
(8) For example, Krueger and Ciolko (1998) find that policy
measures are endogenous to output growth. Popov (2000) also finds policy
changes to be endogenous to initial conditions in former Soviet Onion
countries.
(9) Estimators with increased control on endogeneity include the
fixed-effect specifications that control for the related endogeneity,
the feasible general least square model (FGLS) that corrects for AR(I)
error autocorrelafion, and the Arellano-Bond GMM estimator that further
corrects for endogeneity by exploiting the moment conditions to obtain
consistent estimates in panels with short time series (Arellano and
Bond, 1991).
(10) On parameter heterogeneity, as Hsiao and Sun (2000) have
pointed out, the ability to exploit the information contained in panel
data depends critically on the plausibility of pooling. Hsiao and Sun
(2000, p. 181) note that the decision to pool or not to pool the data
depends on whether, [y.sub.it], the ith individual observation of the
dependent variable at time t, conditional on x, the independent
variable(s) of interest, can be viewed as a random draw from a common
population. This is the socalled exchangeability criterion. It implies
that the individual and time subscript, it, is simply a labeling device.
Observations on the dependent variable should be exchangeable so that, a
priori, E([y.sub.it]|x) E([y.sub.js]|x). In other words, the expected
probability of observing [y.sub.it] or [y.sub.js], conditional on x, the
independent variable(s) of interest, should be the same. If this
condition is satisfied, by pooling the data we can obtain more robust
and precise parameter estimates. However, if individual outcomes are
more appropriately viewed as stemming from a heterogeneous population,
then the subscript it contains important information that can be used to
determine the specific heterogeneous population from which the
particular observation is generated.
(11) Despite the merits of fixed effects, they also have the
limitation of masking some of the possible transmission channels.
Further research through more in-depth case studies can further
illuminate the relationship.
(12) Such as the Hausman test, the F-test, or the Chow test. The
test results support the inclusion of fixed effects in the model.
(13) The data are from the WEO database of the International
Monetary Fund.
(14) This section draws on Kaufman and Segura-Ubiergo (2001) and
Segura-Ubiergo (2007).
(15) In addition, the importance of the short-term effects
[DELTA][X.sub.i,t-1] depends on the size of [[beta].sub.k] and on how
long the effects of changes in [X.sub.i,t-1] persist through time. A
change in [X.sub.i,t-1] produces an immediate (contemporary) change in
[Y.sub.i,t] that is measured by [[beta].sub.k]. If at time t there is a
change in [X.sub.i,t] in the opposite direction to the change in
[X.sub.i,t-1], then there are no more effects. But if the change in
[X.sub.i,t-1] is sustained, then the impact will continue in subsequent
periods and can be measured by [DELTA][X.sub.i,t-1] [(1 + [phi]).sup.t],
where t is the number of periods after the initial change. Thus, for
example, 3 years after the initial change [DELTA][X.sub.i,t-1], the
effect will be [DELTA][X.sub.i,t-1] [(1 + [phi]).sup.3]. Since 0 <
[phi] < -1, the smaller the value of [phi], the longer the sustained
changes in X will persist through time.
(16) This index is developed by EBRD and data refer to the
aggregate reform index measuring progress along three fronts: (i) price
liberalization; (ii) trade and foreign exchange liberalization; and
(iii) privatization, restructuring and financial market reform. Given
the broad coverage of first and second stage reforms, the aggregate
index can be considered a proxy for institutions.
(17) Not all empirical studies include this variable. Its inclusion
or not in the model, as discussed in the next following section, does
not affect the main results.
(18) To be sure, countries began the transition in different years.
This paper follows the transition dating convention in Havrylyshyn et
al. (1998). Most countries are considered to have started transition in
1992. The exceptions are Bulgaria, the Czech Republic, Hungary, Poland,
Romania, and the Slovak Republic, which started the process in 1990, and
Albania, which started transitioning in 1991.
(19) See Table 2. This regional classification follows Havrylyshyn
et al. (1998).
(20) The Eastern European countries are Macedonia, Albania,
Bulgaria, and Romania. Romania had a low overall balance at the
beginning of the sample period.
(21) The main international commodity price alluded to is oil, as
it affects the cost of oil imports on which Belarus is highly dependent
and the price of oil and gas exports of Turkmenistan.
(22) The experience of two countries (Mongolia and Belarus) is
somewhat puzzling because at first sight they do not fit the strong
positive association between fiscal adjustment and growth presented in
Table 2 and discussed in this section. Belarus seems to have experienced
fast acceleration of economic growth even though structural reforms have
been limited and fiscal adjustment low. However, this result is driven
by other important exogenous factors that helped boost growth in
Belarus: the continuing sale by Russia of subsidized oil inputs, a
geographically advantageous position for energy transportation, and
favorable oil price developments. Mongolia, unlike Belarus, seems to
have undergone significant fiscal adjustment but no acceleration in
economic growth. Cheng (2003) suggests that favorable initial conditions
may have contributed to the milder initial recession. However, the
composition of fiscal adjustment in Mongolia may explain why fiscal
adjustment was not associated with higher growth: most of the fiscal
adjustment in Mongolia was revenue-based while the level of expenditures
was kept relatively high. This suggests that the public sector may not
have been sufficiently reformed and may have crowded out private sector
activity.
(23) Calculations of coefficients of variation and revenue and
expenditure elasticities are available from the authors upon request.
(24) It is not entirely obvious why the 'level' variable
of the fiscal deficit in the model without inflation is not
statistically significant. As the variable turns statistically
significant in the model with inflation, and inflation turns out to be
statistically significant in all models across the board, there is a
strong likelihood that Model 9 is misspecified.
(25) The discussion focuses on Models 5 and 6, which are robust to
the presence of outliers.
(26) Note that by long-term growth, we simply refer to growth over
the life of the model.
(27) The results are therefore consistent with a recent paper by
Rzonca and Cizkowicz (2005).
(28) Formal tests on the coefficient differences were conducted by
splitting the sample between CIS and non-CIS countries. Hausman tests
lead to rejection of the hypothesis of no difference in regression
coefficients at the 5% confidence level.
(29) While foreign direct investment averaged below 2% of GDP per
year at the beginning of transition, by 1999 these flows reached about
4% of GDP per year.
(30) Anecdotal evidence of unrecorded foreign remittances abound.
For Armenia, unrecorded private remittances are thought to be in the
form of 'pocket money' from Armenians working in Russia. See
Gelbard et al. (2005) for more details.
(31) The CIS-7 countries are Armenia, Azerbaijan, Georgia, the
Kyrgyz Republic, Moldova, Tajikistan, and Uzbekistan.
(32) Data limitations (consistency, validity, and reliability) did
not allow us to probe deeper in this area.
ALEX SEGURA-UBIERGO, ALEJANDRO SIMONE, SANJEEV GUPTA & QIANG
CUI
Fiscal Affairs Department, International Monetary Fund, 700 19th
Street, N. W., Washington D.C., 20431, USA. E-mail: sgupta@imf.org
Table 1: Description of main variables
Observations Mean Standard Source
deviation
Dependent variable
Real GDP growth 234 0.95 7.88 WEO
Independent variables
General gov. overall 260 -5.59 7.26 WEO
fiscal balance
(in percent of GDP)
Change in general gov. 234 0.85 5.21 WEO
overall fiscal balance
Current account balance 260 -5.54 11.68 WEO
(in percent of GDP)
Change in current 234 0.06 10.18 WEO
account balance
Average annual consumer 202 164.75 566.11 WDI
price index
Change in average annual 179 -93.33 533.73 WDI
consumer price index
EBRD reform index 240 3.12 0.80 EBRD
Change in reform index 216 0.13 0.24 EBRD
Sources: IMF, World Economic Outlook (WEO) database, EBRO, the World
Bank, and World Development Indicators (WDI) database
Table 2: Growth and fiscal adjustment experience by transition
regions, 1992-2001
Country Fiscal Fiscal Fiscal Growth
balance balance adjustment 1992-1996
1992-1996 1997-2001
Central and Eastern Europe
Poland -3.8 -3.9 -0.1 5.8
Romania -3.3 -4.1 -0.7 2.3
Croatia -1.5 -6.2 -4.8 3.5
Slovak Republic -4.5 -5.6 -1.1 3.8
Czech Republic -0.9 -2.8 -1.9 2.4
Slovenia 0.2 -1.0 -1.2 4.1
Hungary -6.6 -4.1 2.5 2.0
Macedonia -4.8 -1.4 3.4 -1.4
Albania -11.8 -10.7 1.0 5.4
Bulgaria -6.6 -0.5 6.2 -6.1
Average -4.3 -4.0 0.3 2.2
Baltics
Estonia -0.1 -1.3 -1.2 1.6
Latvia -1.6 -2.5 -0.9 0.4
Lithuania -3.4 -4.8 -1.3 -2.2
Average -1.7 -2.9 -1.1 -0.1
Commonwealth of independent states and Mongolia
Mongolia -12.5 -9.7 2.7 2.4
Uzbekistan -7.6 -2.6 5.0 -0.7
Armenia -22.3 -5.6 16.7 1.5
Georgia -20.7 -4.7 16.0 -3.2
Kyrgyz Republic -13.1 -9.7 3.4 -4.3
Belarus -1.9 -1.1 0.8 -3.3
Russia -10.2 -1.4 8.8 -5.6
Moldova -10.8 -2.4 8.4 -10.3
Kazakhstan -5.2 -2.7 2.5 -5.6
Ukraine -9.7 -2.0 7.7 -12.5
Tajikistan -12.8 -1.9 10.9 -9.5
Azerbaijan -5.7 -2.4 3.3 -9.5
Turkmenistan 2.5 -0.6 -3.1 -10.5
Average -10.0 -3.6 6.4 -5.5
Overall average -6.9 -3.7 3.2 -1.9
Country Growth Growth
1997-2001 acceleration
Central and Eastern Europe
Poland 3.5 -2.3
Romania 0.5 -1.8
Croatia 2.1 -1.4
Slovak Republic 2.9 -0.9
Czech Republic 1.4 -0.9
Slovenia 4.1 0.0
Hungary 4.5 2.6
Macedonia 1.9 3.4
Albania 9.4 4.1
Bulgaria 4.0 10.1
Average 3.4 1.3
Baltics
Estonia 4.4 2.9
Latvia 5.6 5.2
Lithuania 4.0 6.2
Average 4.7 4.7
Commonwealth of independent states and Mongolia
Mongolia 2.2 -0.2
Uzbekistan 3.2 3.9
Armenia 6.6 5.1
Georgia 3.1 6.3
Kyrgyz Republic 4.1 8.4
Belarus 5.6 8.9
Russia 4.0 9.6
Moldova -0.4 9.9
Kazakhstan 6.0 11.6
Ukraine 3.2 15.7
Tajikistan 6.9 16.4
Azerbaijan 9.6 19.1
Turkmenistan 15.5 26.0
Average 5.4 10.8
Overall average 4.5 6.4
Source: IMF, World Economic Outlook. These data are available
upon request from the authors
Table 3: Fiscal adjustment and growth acceleration in transition
countries
Fiscal adjustment Growth acceleration
Low (<2.5 percentage High (>2.5
points) percentage points)
Low (<2.5% GDP) Croatia [-4.8,0.5] (1) Albania [1.0,4.1]
Czech Rep.[-1.9, -1.0] Belarus [0.8, 8.8]
Poland [-0.1, -2.3] Estonia [-1.2, 2.9]
Romania [-0.7, -1.8] Latvia [-1.0, 5.21
Slovak f-1.1, -1.0] Lithuania [-1.3, 6.2]
Slovenia [-1.2,0.0] Turkmekistan [-3.1, 26.0]
High (>2.5% GDP) Mongolia [2.7, -0.2] Armenia [16.6, 5.1]
Azerbaijan [3.3, 19.1]
Bulgaria [6.2, 10.1]
Georgia [16.0, 6.3]
Hungary [2.5, 2.6]
Kazakhstan [2.5, 11.6]
Kyrgyz [3.4,8.4]
Macedonia [3.4,3.4]
Moldova [8.4,9.9]
Russia [8.8, 9.6]
Tajikistan [10.9, 16.4]
Ukraine [7.7, 15.7]
Uzbekistan [5.0, 3.9]
(1): If the big stump of 1992 (a large outlier) is included, Croatia
would move to the upper right cell in Table 3.
Note: The first figure in brackets corresponds to fiscal adjustment
and the second corresponds to growth acceleration. Bold
distinguishes the countries in the matrix diagonal.
Source: IMF, World Economic Outlook.
Table 4: Econometric results 1
OLS
[1] [2]
GDP growth (L) 0.5346 *** 0.3462 ***
[0.0673] [0.0966]
Reform index (L) 1.4998 *** 5.9487 ***
[0.7805] (2.0655]
Reform index (D) -0.9755 1.5448
[2.19261 [2.6141]
Fiscal balance (L) 0.4047 *** 0.6025 ***
[0.1475] [0.1799]
Fiscal balance (D) 0.742 *** 0.9161 ***
[0.2167] [0.1986]
Current account balance (L) -0.0940 -0.0915
[0.0723] [0.0920]
Current account balance (D) [0.0503] -0.0217
[0.0841] [0.0950]
Inflation (L)
Inflation (D)
Constant -1.9036 -11.436
[2.6739] [7.5059]
Observations 192 155
F 22.89 24.41
Prob>F 0.0000 0.0000
Wald chi-square
Prob>Wald
[R.sup.2] 0.5427 0.6303
Root MSE 5.1058 3.7401
Fixed effects No Yes
OLS
[3] [4]
GDP growth (L) 0.453 *** 0.2378 **
[0.0698] [0.0974]
Reform index (L) -0.3138 0.9441
[0.6322] [1.6103]
Reform index (D) -0.4137 -1.2453
[2.0046] [2.2435]
Fiscal balance (L) 0.2562 * 0.6349 ***
[0.1479] [0.1742]
Fiscal balance (D) 0.6812 *** 0.7665 ***
[0.2034] [0.1832]
Current account balance (L) -0.0536 0.0365
[0.0569] [0.0803]
Current account balance (D) -0.0685 0.0365
[O.0814] [0.0803]
Inflation (L) -0.0083 *** -0.0083 ***
[0.0017] [0.0013]
Inflation (D) -0.007 *** -0.0074 ***
[0.0016] [0.0013]
Constant 4.5312 ** 9.8426
[2.3633] [6.5827]
Observations 155 155
F 24.41 20.48
Prob>F 0.0000 0.0000
Wald chi-square
Prob>Wald
[R.sup.2] 0.6303 0.7277
Root MSE 3.7401 3.4567
Fixed effects No Yes
Robust regression
[5] [61
GDP growth (L) 0.4608 *** 0.1469
[0.0454] [0.0526]
Reform index (L) 1.9951 ** 0.7881
[0.9905] [1.24931
Reform index (D) -1.0749 -2.5095
[1.5377] [1.8842]
Fiscal balance (L) 0.4666 *** 0.8769 **
[0.1159] [0.1369]
Fiscal balance (D) 0.9220 *** 0.8967 ***
[0.1273] [0.12991
Current account balance (L) -0.1005 *** 0.0411
[0.0454] [0.0721]
Current account balance (D) -0.1532 *** -0.0007
[0.0471] [0.0654]
Inflation (L) -0.0083 ***
[0.0012]
Inflation (D) -0.0062 ***
[0.0012]
Constant 1.2436 14.3635
[4.3452] [5.1236]
Observations 192 155
F 15.68 17.64
Prob>F 0.0000 0.0000
Wald chi-square
Prob>Wald
[R.sup.2] n.a n.a
Root MSE n.a n.a
Fixed effects Yes Yes
FGLS
[7] [8]
GDP growth (L) 0.3162 *** 0.2804 ***
[0.0608] [0.0543]
Reform index (L) 6.3277 *** 0.9341
[1.3411] [1.2423]
Reform index (D) 1.6359 -0.6626
[2.0336] [1.9831]
Fiscal balance (L) 0.6034 *** 0.6002 ***
[0.1552] [0.1382]
Fiscal balance (D) 0.8979 *** 0.7555 ***
[0.1688] [0.1369]
Current account balance (L) -0.0850 0.0892
[0.0617] [0.0725]
Current account balance (D) -0.0201 0.0308
[0.0625] [0.0687]
Inflation (L) -0.0084 ***
[0.0012]
Inflation (D) -0.0076 ***
[0.0013]
Constant -12.4879
[5.8945]
Observations 192 155
F
Prob>F
Wald chi-square 299.7000 504.58
Prob>Wald 0.0000 0.0000
[R.sup.2] n.a n.a
Root MSE n.a n.a
Fixed effects Yes Yes
GMM
[9] [10]
GDP growth (L) 0.4318 *** 0.2234 **
[0.0760] [0.0838]
Reform index (L) 8.8624 *** -0.2043
[1.9282] [2.5225]
Reform index (D) 5.1984 *** -3.4134
[2.5361] [2.8779]
Fiscal balance (L) 0.1839 0.5918
[0.2051] [0.2289]
Fiscal balance (D) 0.4947 *** 0.7825 ***
[0.2065] 0.1222
Current account balance (L) 0.0587 [0.1086]
[0.0750] 0.028
Current account balance (D) -0.0047 [0.0916]
[0.0691]
Inflation (L) -0.0086 ***
[0.0020]
Inflation (D) -0.0068
[0.0019]
Constant -0.0751 -0.2059
[0.2215] [0.2039]
Observations 192 134
F
Prob>F
Wald chi-square 139.28 120.68
Prob>Wald 0.0000 0.0000
[R.sup.2] n.a n.a
Root MSE n.a n.a
Fixed effects n.a n.a
(1): Variables followed by 'L' were lagged, while variables followed
by V were first-differenced. Current GDP Growth is the dependent
variable. 'n.a' means not applicable.
Notes: The first four models are estimated using ordinary least
squares with White-corrected (heroscedasticity-consistent)
standard errors. Models 5 and 6 use robust regression in
order to correct for outliers. Models 7 and 8 are estimated
with generalized least squares assuming a common autoregressive
process and homoscedastic panels. Models 9 and 10 are
estimated using the Arellano-Bond generalized method of moments
estimator. Where possible, Lagrange Multiplier tests were used to
test whether there was any remaining serial correlation. In most
cases no serial correlation was found. Corrogram results on the
baseline fixed-effect panel OLS are also provided in Table 7 of
Annex I to show that there is no significant threat of serial
correlation because no serial correlation is not rejected in any case
at the 1% level and in only I out of 48 cases at the 5% level with
tests up to two lags. This means that any possible bias due to the
simultaneous inclusion of a lagged dependent variable and fixed
effects is likely to be very small.
* significant at 10%; ** significant at 5%; *** significant at 1%.
Table 5: Selected economic indicators at the beginning and at the end
of the sample period (average for years 1992-1993 and 2000-2001)
Country Total revenues (TR) Total expenditures (TE)
1992-1993 2000-2001 1992-1993 2000-2001
Central and Eastern Europe
Poland 45.2 38.1 50.5 42.6
Romania 35.6 30.7 38.1 34.3
Croatia 33.2 45.5 35.5 52.1
Slovak republic 42.9 37.7 52.0 44.7
Czech republic 44.5 39.2 44.3 42.4
Slovenia 41.7 41.2 40.7 42.5
Hungary 45.9 45.8 54.4 49.6
Macedonia 39.7 35.5 51.3 37.5
Albania 23.2 22.5 36.4 31.1
Bulgaria 37.8 38.1 45.9 39.1
Average 39.0 37.4 44.9 41.6
Baltics
Estonia 37.2 37.8 37.6 37.9
Latvia 32.2 36.2 32.3 38.9
Lithuania 31.1 30.4 33.5 32.8
Average 33.5 34.8 34.5 36.5
Commonwealth of independent states and Mongolia
Mongolia 30.0 36.5 43.6 42.7
Uzbekistan 33.8 27.0 48.6 29.1
Armenia 28.8 17.5 75.6 22.5
Georgia 14.8 15.7 53.3 18.7
Kyrgyz Republic 21.0 19.5 37.1 27.2
Belarus 48.8 45.4 51.1 46.3
Russia 37.9 37.1 50.9 34.2
Moldova 26.4 30.0 42.6 30.6
Kazakhstan 27.0 23.7 31.3 22.7
Ukraine 38.5 33.5 56.5 34.9
Tajikistan 24.2 14.4 49.5 14.8
Azerbaijan 45.8 20.7 52.1 21.2
Turkmenistan 47.3 1.7 36.1 1.5
Average 32.6 24.8 48.3 26.6
Overall Average 35.2 30.8 45.4 33.5
Country Overall balance Current account
(percent of GDP)
1992-1993 2000-2001 1992-1993 2000-2001
Central and Eastern Europe
Poland -5.2 -4.5 -3.0 -5.0
Romania -2.5 -3.6 -6.2 -4.7
Croatia -2.4 -6.7 4.5 -3.1
Slovak republic -9.1 -7 -2.6 -5.9
Czech republic 0.3 -3.2 1.7 -5.5
Slovenia 1.0 -1.3 3.6 -1.3
Hungary -8.5 -3.9 -4.0 -4.8
Macedonia -11.6 -1.9 -1.7 -4.5
Albania -13.2 -8.7 -40.2 -6.8
Bulgaria -8.1 -0.9 -7.2 -5.9
Average -5.9 -4.2 -5.5 -4.7
Baltics
Estonia -0.5 -0.1 2.6 -5.9
Latvia -0.1 -2.7 6.6 -8.2
Lithuania -2.4 -2.4 -1.1 -5.4
Average -1.0 -1.7 2.7 -6.5
Commonwealth of independent states and Mongolia
Mongolia -13.7 -6.2 1.2 -5.9
Uzbekistan -14.8 -2.1 -7.2 0.3
Armenia -46.9 -5.1 -34.6 -12.3
Georgia -38.5 -3.0 -21.4 -5.5
Kyrgyz Republic -16.1 -7.7 -10.7 -5.0
Belarus -2.4 -1.0 -3.3 -3.1
Russia -13.0 2.9 0.0 14.0
Moldova -16.2 -0.6 -8.9 -6.9
Kazakhstan -4.3 1.0 -30.1 0.4
Ukraine -18.0 -1.4 -4.6 4.2
Tajikistan -25.3 -0.4 -22.1 -6.7
Azerbaijan -6.3 -0.5 -14.4 -2.2
Turkmenistan 11.2 0.2 56.5 3.4
Average -15.7 -1.8 -7.7 -1.9
Overall Average -10.2 -2.7 -5.6 -3.6
Country Change in TR Change in TE
Central and Eastern Europe
Poland -7.1 -7.8
Romania -4.9 -3.8
Croatia 12.3 16.6
Slovak republic -5.2 -7.3
Czech republic -5.3 -1.8
Slovenia -0.5 1.80
Hungary -0.1 -4.8
Macedonia -4.2 -13.9
Albania -0.7 -5.2
Bulgaria 0.3 -6.8
Average -1.6 -3.3
Baltics
Estonia 0.6 0.2
Latvia 4.0 6.6
Lithuania -0.7 -0.8
Average 1.3 2.0
Commonwealth of independent states and Mongolia
Mongolia 6.5 -0.9
Uzbekistan -6.8 -19.6
Armenia -11.3 -53.1
Georgia 0.9 -34.6
Kyrgyz Republic -1.5 -9.9
Belarus -3.4 -4.8
Russia -0.8 -16.7
Moldova 3.6 -12.0
Kazakhstan -3.3 -8.6
Ukraine -5.0 -21.6
Tajikistan -9.8 -34.7
Azerbaijan -25.1 -30.9
Turkmenistan -45.6 -34.6
Average -7.8 -21.7
Overall Average -4.4 -11.9
Source: IMF, World Economic Outlook
