Information about Europe Europa
Chapter

IV The Experience of Earlier Accession Countries

Author(s):
International Monetary Fund
Published Date:
April 2002
Share
  • ShareShare
Information about Europe Europa
Show Summary Details

In their drive to join the EU, the Baltics face fiscal challenges similar to those of the countries that joined in the past three decades and had incomes below the EU average. Such countries were Greece, Ireland, Portugal, and Spain—henceforth referred to as EACs (earlier accession countries). The harmonization of indirect taxation, the Common Agricultural Policy (CAP), the implementation of EU-related transportation projects and environmental standards, and, more recently, the adherence to Maastricht Treaty criteria and SGP provisions have all required a substantial reorientation in the EACs’ fiscal policies. The way in which these countries managed these challenges, often requiring a considerable change in the size and composition of revenues and expenditures, can provide a lesson for the Baltic countries and other EU accession candidates.

Similarities and Differences Between the Baltics and the EACs at the Point of Accession

The Baltic countries share a number of economic similarities with the EACs at their point of accession but also exhibit some striking differences (Table 3). The Baltic region represents only a small proportion of the EACs’ geographical size and population. The Baltics’ per capita GDP in 2000 was just over one-third that of the EACs’ average upon accession, if measured at 2000 prices. This remains true, albeit to a lesser extent, even when GDP is measured on a purchasing power parity (PPP) basis. In turn, this is reflected in a larger income gap with the EU, with the Baltic countries accounting for less than 30 percent of the average income in the EU, compared with 70 percent for EACs (Figure 3). This suggests, among other things, that the Baltics will require either larger per capita capital inflows (which appear to be forthcoming) or a longer time horizon to complete their convergence process. The better inflation performance of the Baltic countries relative to the EACs mostly reflects the effective discipline of their hard pegs.

Figure 3.Earlier Accession Countries (EACs)1 and the Baltic Countries: Real Convergence

(Ratio of per capita income to EU average, on PPP basis)2

Source: IMF, World Economic Outlook

1 The EACs are Greece, Ireland, Portugal, and Spain.

2 PPP, purchasing power parity.

Table 3.Earlier Accession Countries (EACs) and the Baltic Countries: Key Macro economic Indicators
EACs (at time of accession)Baltic Countries
Greece (1981)Ireland (1973)Portugal (1986)Spain (1986)Estonia (2000)Latvia (2000)Lithuania (2000)
Population (millions)9.53.010.038.41.52.63.8
GDP (millions of U.S. dollars)184,31827,92552,903375,3174,9737,11811,202
Per capita GDP (U.S. dollars)18,9239,3775,2809,7833,2152,7522,980
On purchasing power parity basis (U.S. dollars)1,211,25410,28210,54313,0995,9574,5785,319
In percent of EU average (ratio)75.571.659.073.330.523.527.3
Average CPI (change in percent)7.611.411.68.94.01.81.0
Economic sectors (percent of GDP)
Agriculture14.09.18.95.65.74.06.9
Industry31.625.739.535.725.122.329.4
Services354.465.251.658.769.273.763.7
General government (percent of GDP)
Revenue26.723.435.834.638.937.430.3
Expenditure and net lending35.027.742.340.339.240.733.0
Fiscal balance-8.4-4.4-6.4-5.7-0.3-3.3-2.7
Total public and publicly guaranteed debt29.6n.a.68.444.26.215.028.8
External sector (percent of GDP)
Exports of goods and nonfactor services24.534.628.819.095.446.047.1
Imports of goods and nonfactor services40.540.731.117.1100.454.552.0
Current account balance-5.4-2.83.51.5-6.4-6.9-6.0
FDI inflows and net equity41.41.01.92.07.35.04.4
Sources: National authorities and IMF staff estimates

Valued at 2000 prices for the earlier EU accession countries.

Estimated based on nominal and PPP-based GDP at actual prices for the earlier EU accession countries.

Includes financial intermediation and net taxes.

For Greece, only FDI inflows are available for 1981.

Sources: National authorities and IMF staff estimates

Valued at 2000 prices for the earlier EU accession countries.

Estimated based on nominal and PPP-based GDP at actual prices for the earlier EU accession countries.

Includes financial intermediation and net taxes.

For Greece, only FDI inflows are available for 1981.

The Baltics show a more advanced structure of the economy than the EACs. Agriculture accounted for only 5 percent of output in 2000, compared with 9 percent in EACs, partly reflecting the rapid restructuring of the sector after independence. At the same time, industrial output was somewhat lower than in the EACs, since the Baltics seem to have bypassed the years of heavy industrialization typical of western economies in the 1970s. The share of services in output was 11 percentage points higher in the Baltics because tourism, banking, and other services play a substantial role in output.

The Baltic countries have pursued a more prudent fiscal policy than the EACs, Their fiscal deficits in 2000 amounted to 2 percent of GDP on average, compared with an average of about 6 percent of GDP in the EACs. The Baltics’ revenue-to-GDP ratio exceeds the ratio of those EACs that joined the EU first, namely Ireland and Greece. The size of the public sector varied within each group, but the average size was somewhat higher in the Baltic region, at 38 percent of GDP. The Baltic countries also enjoyed a substantially lower public debt burden than the EACs.

The external sector dominates economic activity in the Baltics. Trade and service flows averaged 130 percent of GDP in 2000, more than double the EACs’ flows. This difference can be partly explained by the liberal trade regime and the small size of the Baltic countries. The Balics’ somewhat larger current account deficits mainly reflect large FDI-related imports.

It is generally expected that the main economic benefit of EU integration will be an accelerated convergence to the EU’s average income level, as increased FDI sustains a higher real growth rate and reduces the income differential. Real convergence has already taken place in the EACs, albeit with somewhat differing results. Although Ireland has achieved full convergence to the EU’s average income level in the past ten years, Greece and, to a lesser extent, Portugal and Spain have lagged behind. The evidence presented in this section suggests that macroeconomic stability and fiscal discipline may have played a significant role in determining the speed of convergence in EACs.

Fiscal Challenges of EU Integration

EU accession generally requires significant changes in the composition of revenue and expenditure. These changes played themselves out over many years in the EACs. The harmonization of indirect taxes was established in these countries’ respective accession agreements, but it only became fully effective prior to the creation of the European Single Market. The EACs were also in a position to influence the contents of various EU directives and initiatives, by requesting transitional periods and additional financing for implementation as these directives were adopted. This gave EACs’ governments more time to adjust the composition of spending.

The Rising Tax Burden in the EACs

The tax burden in the EACs has risen substantially over the past decades. The tax-to-GDP ratio rose from an average of 21 percent of GDP in 1975 to 36 percent of GDP in 2000. This increase has reduced the differential with the original founders of the EU (the EU-6 countries: Belgium, France, Germany, Italy, Luxembourg, and the Netherlands), especially after 1980 (Figure 4).

Figure 4.EACs and the EU-6: Fiscal Developments1

(Percent of GDP)

Source: OECD Analytical Database; IMF, International Financial Statistics, and IMF staff estimates.

1 The EU-6 countries are the original founders of the EU: Belgium, France, Germany, Italy, Luxembourg, and the Netherlands.

This decline in the tax differential could be partially attributed to the European integration process. This might have eroded the geographical advantage (“agglomeration economies”) of the EU-6 (Baldwin and Krugman, 2000). Baldwin and Krugman demonstrate in a model with geographical externalities that higher investment inflows would result in an increase in the optimal tax rate of the recipient country. They conclude that the convergence process in tax ratios is the result of increased capital mobility within the EU, especially after the establishment of the EMU, which has led to higher FDI in the EACs and the harmonization of tax policies.

Although economic integration played a role, institutional factors also contributed to the convergence of the tax burden. The harmonization of indirect taxes was part of the EACs’ accession agreements and was gradually introduced before the establishment of the European Single Market. Moreover, the increased spending pressures related to the EU, as well as discretionary domestic spending decisions, forced the EACs to raise additional domestic resources, mainly in Greece and Portugal. This was done through an increase in taxation, ultimately leading to little divergence from the tax burden in the EU-6.

Both the EACs’ integration experience and the institutional factors associated with EU accession suggest that reducing the tax burden over the next decade in the Baltic countries may present a major challenge. The scope for tax reductions would be limited by EU directives to changes in direct taxes that do not infringe on the “EU Code of Conduct on Business Taxation.” But such efforts would be constrained by the Baltics’ balanced-budget objective, as well as by the requirements to meet the criteria under the Maastricht Treaty and the SGP.

Indirect Tax Harmonization

The harmonization of indirect taxation required a substantial restructuring of the tax systems in the EACs.22 The introduction of the value-added tax (VAT) represented a particularly difficult challenge, since all of the EACs applied a cascading sales tax prior to accession. It required a large investment in the institutional capacity of the EACs’ tax administration. Harmonization of the VAT tax bands and the tax base was fully implemented only at the start of the European Single Market in 1993.

The harmonization of customs tariffs with the EU Common External Tariff (CET) required a lengthy transitional period to avoid major disruptions in domestic production. Tariffs before accession were noticeably higher than EU levels, except for Ireland. Most noteworthy, excise taxation was harmonized only at the inception of the European Single Market.

The Baltics face less of a challenge in the harmonization of indirect taxation, as stated above. Unlike the EACs, they all apply VAT and excise taxation that are already broadly in line with EU directives, although raising some excise rates may be politically difficult to implement. Adjustments may be required in the tax base, but the institutional capacity to administer tax refunds is already in place. Trade barriers with the EU were eliminated in the context of the 1998 Europe Agreement between the EU and the Baltic countries.

The main challenge will be the adoption of the CET. The Baltics have lower tariff protection than the EU and signed several free-trade agreements with non-EU and non-EFTA (European Free Trade Association) countries. The adoption of the CET will simply an increase in trade protection in the Baltics, where trade and service flows account on average for 130 percent of GDP. While the share of trade with the EU is close to 70 percent in Estonia and Latvia and 50 percent in Lithuania, an increase in trade protection to other countries may still have a sizable effect on prices of production inputs and consumer goods (with clear welfare losses) and could impinge on the Baltic countries’ geographical advantage as a link between Russia and the EU.

Size of Government in the EACs

The size of government in the EACs rose dramatically during the 1970s and 1980s, coming close to government size in the EU-6 (see Figure 4). Since then, it has fallen somewhat, reflecting the EU-wide Fiscal consolidation since the mid-1990s, which was partly motivated by the need to meet the Maastricht criteria in 1997 to join the euro area and adhere to the SGP thereafter. The differential with the EU-6 fell until the mid-1980s and has since remained at about 10 percentage points of GDP.

This convergence in the size of the government in the EU partially reflected EU-related spending, but other factors also were at play. In Greece, for example, spending financed by EU grants rose from I percent of GDP in 1981 to 5 ½ percent of GDP in 1992, before declining to above 4 percent of GDP by 1999. Large transfers from the EU budget also led to a larger government sector in the other EACs. However, other factors contributed to the expansion of the public sector as well, such as a discretionary easing of fiscal policy, political business cycles, and an overregulated economy, especially in Greece.

The size of government in the Baltics is likely to be less affected by EU transfers. The EU agreed in 1999 that annual transfers to new members would be capped at 4 percent of the recipient’s GDP. The cap was intended to ensure that the implications of enlargement on the EU budget would be contained. Although a portion of EU transfers is likely to finance existing domestic spending (see Section V), the increase in the size of the public sector in the Baltics is likely to be contained in view of the authorities’ twin objectives of moving toward a balanced budget while lowering the tax burden.

Common Agricultural Policy

The CAP accounts for the largest share of EU transfers to EACs. In the case of Ireland, explicit budgetary transfers under the CAP amounted to about 3 ½ percent of GNP annually between 1979 and 1986, more than half of total transfers. The other EACs received somewhat less as a share of GNP; however, the CAP for all countries still represents the largest net transfer from the EU. The reform of the CAP in 1988—stipulating that agricultural expenditure was to grow each year by less than 75 percent of the annual real GNP growth rate of the EU—limited the share of agricultural subsidies to EACs. However, subsidies under the CAP still accounted for 47 percent of the EU budget in 1999.

The introduction of the CAP in the EACs led to a real increase in both explicit and implicit subsidies to the agricultural sector and the eventual elimination of intra-EU trade barriers. The main implicit subsidy resulted from effective trade barriers against non-EU producers under the CET. Both explicit and implicit subsidies were, however, limited by the introduction of production quotas in 1984. To compensate for this, the EU increasingly shifted subsidies to support agricultural exports to non-EU countries. The elimination of intra-EU trade barriers led to a convergence of agricultural prices across the EU and increased specialization. It is not clear whether the CAP resulted in productivity gains, especially in small farms, since most of the agricultural investment guaranteed by EU funds was concentrated in large entities (see Keane and Lucey, 1997, and Rosenblatt and others, 1988).

The size of the agricultural sector in the Baltics has been steadily declining since independence. It averaged about 5 percent of GDP in 2000,23 only half the size of that in the EACs before their accession. The impact of the CAP on the Baltics is thus likely to be smaller relative to the EACs’ historical experience. Barring reform to the CAP before accession, explicit and implicit subsidies to agriculture in the Baltic countries will increase after accession along with domestic prices for agricultural products. The extent of the subsidies will depend on the allocation of production quotas to each country. The introduction of the CET and export subsidies is likely to affect agricultural trade with non-EU members.

Regional, Transportation, and Environmental Projects

About one-third of all EU transfers goes toward aid to poorer regions, under the auspices of the structural and cohesion funds,24 The goal of both funds is to reduce the income gap between different regions of member countries and “thereby promote economic and social cohesion.” The allocation for 2000–06 has increased somewhat in nominal terms, in part to take account of allocations for future member countries.

The EU’s regional policy has traditionally focused on large infrastructure projects with the aim to stimulate private investment in poorer regions and generate employment. In this context, large transportation projects have featured prominently, particularly highway construction to develop a European network of freight transport under the Common Transport Policy (CTP). These large investments have usually required a substantial portion of domestic finances. More recently, the CTP has strived, in the approval of new infrastructure projects, to take more carefully into account the absorptive capacity of the recipient country. In addition, environmental projects have gained in importance lately. This reflects the adoption of stringent environmental standards that became part of EU legislation in 1986, although their enforcement has been somewhat flexible.

The Baltic countries will face substantial pressures in the coming years to catch up with the regional projects and environmental standards of the EU. Even if a large share of these projects were to be financed by the EU, experience (especially that of Greece) shows that there is a limit to the absorptive capacity of recipient countries. To what extent the catching up will need to be done before accession is currently a matter of negotiation. However, the EACs’ experience suggests that the European Commission has, in the past, shown considerable flexibility in the implementation timetable. Against this background, the optimal rate of implementation must be one that does not endanger macroeconomic stability.

Real Convergence

The experience of the EACs suggests that real convergence is not an automatic process, as standard neoclassical growth theory appears to suggest (Box 5). The extent and speed of convergence has in fact varied among the four EACs during the past 30 years (see Figure 3). In the case of Greece, despite EU regional aid, the ratio of per capita GDP (measured on a PPP basis) to that of the EU average has remained roughly unchanged. Spain and Portugal have gradually closed their income gaps since joining the EU by about 10 and 20 percentage points, respectively, but still remain below the EU average.

The Irish convergence experience in the second half of the 1990s is strikingly different. Although the Irish income gap had narrowed by 11 percentage points in the 24 years before 1994, it was eliminated by 1998, and income stood at 16 percent above the EU average by 2000. This reflects a spectacular acceleration of economic growth in the past six years, fueled by very large FDI inflows of more than 20 percent of GDP in 2000 alone.25 It is worth noting that the successful convergence of the 1990s followed on a successful macroeconomic stabilization in the 1980s that was based on a substantial fiscal adjustment. The fiscal deficit was reduced from 14 percent of GDP in 1982 to less than 2 percent in 1989. At the same time, inflation fell from 17 percent to 4 percent.26

Empirical Evidence

Following the first oil price shock in 1973, the EACs experienced a protracted period of macroeconomic instability that in the end was tamed only by a sizable fiscal adjustment (Figures 4Figures 5). Ireland set the stage for fiscal adjustment in the 1980s, after a dramatic widening of the fiscal balance in the 1970s. This in turn led to a rapid decline of inflation from its peak in 1981. The Iberian countries followed a similar fiscal adjustment path in the mid-1980s, albeit from a lower initial imbalance, Greece delayed its adjustment until the 1990s, with a corresponding delay in the achievement of macroeconomic stability.

Figure 5.EACs: Macroeconomic Stability

(CPI inflation; percent a year)

Sources: IMF, International Financial Statistics; and IMF staff estimates

The conditions of macroeconomic instability in the 1970s and 1980s undercut the ability of the EACs to attract substantial amounts of FDI (Figure 6). Up to 1995. FDI inflows to all four EACs averaged about 2 percent of GDP annually, which even the advent of the European Single Market did not seem to alter immediately. However, after 1995, Ireland saw a strong rise in FDI inflows, with Spain and Portugal also registering a moderate increase.

Figure 6.EACs: Annual FDI Inflows1

(Percent of GDP)

Sources: IMF, International Financial Statistics; and IMF staff estimates.

1 FDI, foreign direct investment.

Box 5.Macroeconomic Convergence

Policymakers from accession countries widely believe that the main benefit of EU membership will be a faster convergence of living standards to EU averages (see, for example. Bank of Estonia, 2000). This belief may be based partly on the prediction of neoclassical growth models that, with the elimination of barriers to goods and factor movements, capital will flow to countries with a lower capital stock to take advantage of higher returns. In turn, this inflow of foreign investment will sustain a higher growth rate in the new member states and thus reduce the income gap with the EU average (based on Solow, 1956). This convergence process comprises both an alignment of the nominal price level (nominal convergence) and of per capita real GDP (real convergence).

The automatically of convergence predicated by neoclassical growth models has been challenged in recent years. (For a summary of the theoretical discussions and the implications for EU accession countries, see World Bank, 2002.) In the new growth literature, international factor mobility, the engine of real convergence in the neoclassical models, may not work in favor of convergence. If the assumption of diminishing returns to capital is relaxed, as in Romer (1986). foreign investment may not flow to less advanced economies where capital is scarce. In the same vein, if human capital is allowed to move freely, as in Lucas (1988), skilled labor may move from poorer to richer countries to take advantage of higher real wages, thus increasing cross-country divergence.

Real convergence may also not materialize as a result of spatial externalities. According to the new geography literature pioneered by Krugman (1991), returns on capital may be higher in areas where other firms are already producing similar products, since geographical proximity increases the probability of technology and human capital spillovers. These “agglomeration economies,” or “Silicon Valley” effects, would in turn reduce the incentives for capital inflows to less developed countries and slow, or even halt, the process of real convergence.

Two theoretical arguments can be made for a positive correlation between macroeconomic stability and the speed of convergence. The first one is the standard argument that macroeconomic stability is good for investment, since it reduces the uncertainty associated with country risk. This has also been interpreted in an intertemporal context, where budgetary rules can solve the time-inconsistency problem associated with discretionary fiscal policies (see Kopits, 2000, and IMF, 2001). The second argument is that a more prudent macroeconomic policy in one country, compared with a set of similar countries such as the EACs, may induce FDI to flow to that country first. As agglomeration economies occur, FDI flows to that country will be reinforced, thus resulting in a faster process of convergence. Anecdotal evidence on recent investment inflows in Ireland suggests that these “bandwagon” or “cascade” effects have strong empirical support (Barry, Bradley, and O’Malley, 1999).

Finally, although the theoretical and empirical literature on the positive effects of macroeconomic stability and fiscal prudence for growth is extensive, the specific role of fiscal policy and macroeconomic stability has received only marginal attention in the literature on convergence (for a notable exception, see World Bank, 2002).

Prudent fiscal policies seem to condition not only macroeconomic stability but also FDI inflows and, consequently, real convergence. A qualitative comparison between the fiscal balance (Figure 4, bottom panel) and the speed of real convergence (Figure 7) suggests a lagged positive relationship between the two. This is particularly evident in the 1990s, when Ireland showed a consistently high speed of real convergence following the fiscal adjustment of the 1980s, while Greece, and to a lesser extent Portugal and Spain, lagged behind. Since 1995, fiscal adjustment and the achievement of macroeconomic stability seem to have also started to pay off for the southern EU members, with a positive speed of real convergence, albeit much less than that of Ireland.

Figure 7.EACs: Speed of Real Convergence

(First difference of ratio of per capita income to EU average, on PPP basis)

Sources: IMF, International Financial Statistics; and IMF staff estimates.

Although many elements come into play in the process of real convergence, one can attempt to quantify the role of fiscal policy and macroeconomic stability. The following structural equation was estimated for each EAC:

where SICt is the speed of income convergence, 27 ΔFDIt is the change in inward FDI as a share of GDP, ΔGGBt, is the change in the general government balance in percent of GDP, ΔInflt is the change in annual CP1 inflation, and ESM is a dummy variable controlling for the introduction of the European Single Market in 1993.28 The L subscript indicates the significant lag length, which was determined empirically on the basis of the Aikaike information criterion.

The motivation for equation (3) rests on the recent literature about the role of prudent macroeconomic policies in economic growth. Both the achievement of macroeconomic stability and fiscal adjustment have been shown to have a direct, albeit lagged, impact on growth, and thus on the speed of convergence. Equation (3), however, should not be interpreted as a “full” convergence equation because other factors beyond the scope of this paper have a significant bearing on convergence as well (such as human capital, institutional settings, political stability, etc.). The sample data comprised 31 annual observations per country.

Formulation of equation (3) in first differences is dictated by the nonstationarity of the regressors in levels. The unit-root hypothesis cannot be rejected for the regressors in levels for every country. However, tests on the first difference of the regressors and the dependent variables indicate that the unit-root hypothesis can be rejected at least at the 5 percent level of significance, and for many at the 1 percent level of significance (Table 4), Based on these results, equation (3) was estimated using a standard ordinary least-squares (OLS) procedure with heteroscedasticity-consistent standard errors (Table 5).

Table 4.EACs: Phillips-Perron Unit-Root Tests(Sample: 1970—2000)
GreeceIrelandPortugalSpain
Variables1
Speed of income convergence-5.406**-3.529*-3.842**-2.989*
Change in FDI-4.733**-3.340*-5.031**-7.219**
Change in general government balance-5.653**-7.980**-6.937**-5.090**
Change in inflation-4.907**-3.249*-6.023**-4.770**

An asterisk (*) next to the statistics indicates significance at the 5 percent level. A double asterisk (**) indicates significance at the 1 percent level. The test for the speed of income convergence for Ireland includes a trend term.

An asterisk (*) next to the statistics indicates significance at the 5 percent level. A double asterisk (**) indicates significance at the 1 percent level. The test for the speed of income convergence for Ireland includes a trend term.

Table 5.EACs: Ordinary Least-Squares (OLS) Regression Results(White heteroscedasticity–consistent standard errors and covariance)
Dependent VariableGreeceIrelandPortugal 1Spain1
Speed of income convergence
Regressors2
Change in FDI inflows1.875 (-1)0.197 (-1)0.671 (-2)0.599 (-1)
(T-ratio)(2.162)*(2.029)*(1.781)+(1.394)
Change in general government balance0.189 (-2)0.187 (-1)0.345 (-1)0.144 (-1)
(T-ratio)(1.722)+(3.556)*(2.416)*(1.506)
Change in inflation-0.164-0.074-0.152 (-1)-0.209 (-3)
(T-ratio)(-2.213)*(-0.738)(-3.072)**(-2.255)*
Single European Market dummy (1993)-0.3274.2010.5990.352
(T-ratio)(-0.892)(6.178)**(l.698)(1.459)
Diagnostics
R-square0.3870.6770.4090.597
Adjusted R-square0.3130.6390.2800.505
S.E, of regression1.2931.5931.4140.912
F-statistics4.47486.7295.7915.840
Mean of dependent variable-0.0671.4910.5860.220
Residual sum of squares41.82565.97145.99318.234
DW-statistic1.7091.7661.5362.027
Serial correlation LM test (F-statistic)30.0670.4560.5291.810
Aikaike information criterion3.4803.8933.7132.840
Equation log-likelihood-46.459-54.388-47.837-33.764

For the Iberian countries, dummy variables were included for the years 1974 and 1979 to control for outliers associated with the first and second oil price shock. The coefficients on these dummy variables were all statistically significant. These outliers were not present in the other regressions.

Lagged values are indicated by negative numbers in parentheses next to the coefficient of the regressor. The coefficient on a constant term was insignificant in all regressions. A plus (+) next to the T-ratio indicates that the corresponding coefficient is significant at the 10 percent level; an asterisk (*) stands for significance at the 5 percent level; and a double asterisk (**) indicates that the coefficient is significant at the 1 percent level.

The Durbin-Watson (DW) statistics indicated no serial correlation in the residuals of each regression. To confirm these results, Breusch-Godfrey LM tests for serial correlation were run as well. These tests are significant at the 5 percent level in all regressions.

For the Iberian countries, dummy variables were included for the years 1974 and 1979 to control for outliers associated with the first and second oil price shock. The coefficients on these dummy variables were all statistically significant. These outliers were not present in the other regressions.

Lagged values are indicated by negative numbers in parentheses next to the coefficient of the regressor. The coefficient on a constant term was insignificant in all regressions. A plus (+) next to the T-ratio indicates that the corresponding coefficient is significant at the 10 percent level; an asterisk (*) stands for significance at the 5 percent level; and a double asterisk (**) indicates that the coefficient is significant at the 1 percent level.

The Durbin-Watson (DW) statistics indicated no serial correlation in the residuals of each regression. To confirm these results, Breusch-Godfrey LM tests for serial correlation were run as well. These tests are significant at the 5 percent level in all regressions.

The regression results seem to support the notion that prudent fiscal policy and macroeconomic stability contribute to convergence. The coefficients on the change in the fiscal balance are all positive and highly significant for Ireland and Portugal and, at the 10 percent level of significance, also for Greece. The magnitude of the coefficients indicates that a fiscal adjustment of 1 percent of GDP leads to an increase in the speed of income convergence of 0.19 (Ireland and Greece) and 0.35 (Portugal) percentage points with a lag of one or two years. The coefficients also confirm that an increase in inflation has a negative and, except for Ireland, significant effect on convergence, A 1 percent increase in inflation leads to a decline in the speed of income convergence of 0.15 to 0.2 percentage points with a lag of zero to three years.

As expected, FDI inflows have a positive effect on convergence. The variable is highly significant in Greece and Ireland and, at the 10 percent level of significance, also in Portugal. The coefficients indicate that an increase in FDI inflows of 1 percent of GDP would increase the speed of income convergence by 0.2 percentage points in Ireland, 0.7 percentage points in Portugal, and 1.8 percentage points in Greece with a lag of one to two years. These coefficients are negatively related to the amount of FDI inflows to each country in the past 30 years, suggesting that the law of diminishing returns may be at work here. Finally, the coefficient on the dummy variable controlling for the introduction of the European Single Market is statistically significant only in Ireland and seems to have had a very large impact on convergence there. This could be interpreted as supporting the idea of a “bandwagon” or “cascade” effect of FDI following the introduction of the European Single Market, which may have accelerated Ireland’s convergence.

The importance of prudent macroeconomic policies for convergence goes beyond what the estimated coefficients imply.29 The coefficients on the change in the fiscal balance and inflation may seem small compared with the ones on FDI inflows. However, the theoretical discussion above suggested that prudent policies were a precondition for large FDI inflows, which act as the engine for convergence. The indirect effects of policies on FDI might therefore be just as important as the direct ones captured by the regressions. As shown by the Irish example, the benefits from fiscal prudence and macroeconomic stability can be large and can materialize in a rather short time span, although other factors—such as flexible labor markets with highly qualified technical workers, the English language advantage, and close links to the rapidly growing U.S. economy—may have also played an important role.

In sum, the Baltic countries are well placed to reap the benefits of macroeconomic stability. The strict adherence to the discipline of currency board arrangements, together with a prudent fiscal policy, has already resulted in some of the highest per capita FDI inflows among accession countries. Continued prudence in macroeconomic management is likely to accelerate the process of convergence further before and after accession to the EU.

    Other Resources Citing This Publication