The EU was founded in 1957 with aim of bringing peace and prosperity to territories that have experienced war for at least eleven centuries. In 2024, this political union represents 450 million people and one-sixth of the world’s GDP. In May 2004, 75 million people across ten countries (Cyprus, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovakia, and Slovenia) became citizens of the EU. Between 2004 and 2019, the GDP per capita of these countries almost doubled from $18,314 to $34,753. GDP per capita is the most common measure of standard of living. It gives the revenue that an average inhabitant receives each year, and when measured in real terms it corrects for inflation. According to the World Bank, based on this measure, eight of the ten countries that joined the EU in 2004 were in the middle-income group in 2004 (except for Cyprus and Malta) and are now in the high-income group.
Was part of this economic miracle the result of accession to the EU? What has been the effect of this enlargement on the 15 existing EU members in 2004? What are the main drivers of this miracle? In recent research (Grassi 2024), I try to answer these questions.
In the last few years, there have been discussions about further enlargement of the EU to the nine current candidate countries, which include Ukraine. The effects of past EU enlargement have been relatively understudied in the economics literature, with notable exceptions including Campos et al. (2014) and Caliendo et al. (2021).
I examine the effect of the EU enlargement on two aggregated economies: the EU-2004, which includes the new member states that joined the Union in 2004; and the EU-15, which consists of the member states prior to the 2004 enlargement. Figure 1 shows the GDP per capita for the EU-2004, the EU-15, and a few select countries in the left panel, and their GDP per capita relative to the EU-15 in the right panel. The figure shows that GDP per capita is growing for both the EU-15 and the EU-2004. Furthermore, GDP per capita before 2004 seems to be constant relative to the EU-15 while it is growing after 2004. The EU-2004 seems to catch up with the EU-15 in terms of standard of living after 2004.
Figure 1 Real GDP per capita
Note: Left panel: GDP per capita in current PPP express in 2017 US dollars. Right panel: Ratio of GDP per capita and GDP per capita of the EU-15 normalised to one in 2004. EU-2004: Aggregation of the ten countries that joined in 2004. EU-15: Aggregation of the 15 EU members before 2004. OECD: Aggregation of OECD countries.
Data source: Penn World Table 10.0.
Challenge and methodology
To evaluate the causal impact of a policy change, researchers usually compare the outcome of the treated group, i.e. those subject to the policy change, with a control group that was not subject to the policy change. Ideally, the treated and control groups would have identical characteristics. If the treatment group is doing better than the control group, this indicates that there is a causal effect of the policy change on the outcome. In the case of the 2004 EU enlargement, there are no countries similar to the EU-2004 that did not join the EU that could be used as control. Similarly, there are no unions similar to the EU-15 that did not experience an enlargement.
In my paper, I use a methodology that can address this identification challenge: the synthetic control method introduced by Abadie and Gardeazabal (2003). The idea is to construct a ‘synthetic control’ as the weighted average of some countries not affected by the EU enlargement taken from a ‘donor pool’. The weighted are chosen such that the dynamics of the treated countries – the EU-2004 or the EU-15 – and the synthetic controls are the same before 2004. If there was a causal effect of the EU enlargement in 2004, then the GDP per capita of the treated country will differ from that of the synthetic control.
Main results
Figure 2 plots GDP per capita of the EU-2004 (left) and the EU-15 (right), and their respective synthetic controls. According to these calculations, the difference between the EU-2004 and its synthetic control is $8,433 in 2019. The accession of these countries to the EU in 2004 led to an increase in their GDP per capita of 32%. Almost a third of their current level of standard of living can be attributed to their accession to the EU, which is about half of the increase in GDP per capita between 2004 and 2019. This is a very large positive effect due to only changes in policies, regulations, trade barriers, and so on.
A similar calculation for the EU-15 does not point to a large positive or negative effect. As the right panel of Figure 2 shows, the dynamics of the GDP per capita of the EU-15 closely follow the dynamics of the synthetic control.
Figure 2 Dynamics of GDP per capita EU-2004/EU-15 and their synthetic controls
Note: Synthetic control was estimated by matching the real GDP per capita from 1991 to 2003. Left panel: EU-2004: Aggregation of the ten countries that joined in 2004. Right panel: EU-15: Aggregation of the 15 EU members before 2004. The red vertical line indicates the year 2004.
Data source: Penn World Table 10.0.
As I show in my paper, these results are robust to a variety of tests, such as a ‘leave one out’ test, in-time placebo, in-country placebo, alternative donor pools, and alternative specifications. Across these tests, the large and positive effect of EU accession on GDP per capita for the new member states is always present. For the EU-15, these robustness checks are consistent with an absence of an effect of the 2004 enlargement.
The EU enlargement of 2004 was remarkably beneficial to the new member states while having no negative consequences for the standard of living of existing members.
Figure 3 TFP gap with the EU-15
What are the drivers of this large and positive effect?
To explore the drivers of the large and positive effect attributed to EU accession in 2004, I perform a simple growth accounting exercise, in the spirit of Solow (1957) and following Baqaee and Farhi (2019). This exercise consists of decomposing growth of GDP into the contribution of factors of production such as capital and labour and into the contribution of the ‘Solow residual’. The latter is often considered a measure of productivity and captures technological progress, better factor allocation, and any change in frictions on the capital and labour market. Such an exercise can tell us how much of the gain in growth is due to more people working (the contribution of labour), to more machines being used (the contribution of capital), or to better productivity.
This growth accounting exercise shows that the contribution to GDP growth of capital and labour is around 60% higher in the EU-2004 than for its synthetic control. Importantly, the contribution of the ‘Solow residual’ is almost three times larger for the EU-2004 than for its synthetic control. The accession to the EU seems to generate a much higher and sustained increase in productivity growth.
Further inspection of data on the components of demand (consumption, investment, government spending, exports/imports) and other macroeconomics aggregates such as the employment rate or foreign direct investment show convergence to a stable level either before or around 2004. They converge to either higher, lower, or similar levels than the EU-15.
Indices on product market regulation produced by the OECD measure state control, barriers to trade and investment, and barriers to entrepreneurship. The last two indicators for the EU-2004 have converged to similar or even lower levels than the EU-15. However, the data are only available for a few years and for a subset of the new member countries (mostly Poland and Hungary before 2008).
Another measure of productivity, total factor productivity (TFP), as estimated by the Penn World Tables (Feenstra et al. 2015) shows signs of convergence to the EU-15 level. As shown in Figure 3, TFP relative to the EU-15 seems to have a higher growth rate after than before 2004. Consistent with the above growth accounting exercise, productivity seems to keep growing and closing the gap with the EU-15.
EU accession resulted in a rapid convergence of the main macroeconomic variables except for TFP, which keep closing the gap with the EU-15. Membership of the EU has a large effect on the GDP per capita of its new members that materialises in sustained productivity gains.
Conclusion
There is a large positive effect of becoming a member of the EU without cost to previous members; EU enlargement seems to be a positive sum game. An analysis of the data points to a large role of productivity measured either by the Solow residual or TFP.
These results prompt more questions than they give answers. Further research on the accession process is needed to understand the mechanism through which a change in policies, regulations, and institutions can have such large positive effects on GDP per capita and productivity.
Several mechanisms are worth exploring, including technological transfer, competition, trade, migration, fiscal transfers, and monetary policy, to name a few. Micro-level data exist that could be exploited. Equipped with a good understanding of these mechanisms, we could evaluate the qualitative and quantitative impact of future accession, which would potentially include Ukraine.
References
Abadie, A and J Gardeazabal (2003), “The Economic Costs of Conflict: A Case Study of the Basque Country,” American Economic Review 93(1): 113–132.
Baqaee, D R and E Farhi (2019), “The Macroeconomic Impact of Microeconomic Shocks: Beyond Hulten’s Theorem,” Econometrica 87(4): 1155–1203.
Caliendo, L, L D Opromolla, F Parro, and A Sforza (2021), “Goods and Factor Market Integration: A Quantitative Assessment of the EU Enlargement,” Journal of Political Economy 129(12): 3491–3545.
Coricelli, F, N Campos, and L Moretti (2014), “How much do countries benefit from membership in the European Union?”, VoxEU.org, April.
Feenstra, R C, R Inklaar, and M P Timmer (2015), “The Next Generation of the Penn World Table,” American Economic Review 105(10): 3150–82.
Grassi, B (2024), “The EU Miracle: When 75 Million Reach High Income”, IGIER Working paper, May.
Solow, R M (1957), “Technical Change and the Aggregate Production Function,” The Review of Economics and Statistics 39(3): 312–320.
