Is Joining the European Union a Good Thing? A Synthetic Control Testing the Effects on the Slovak Republic. Nicole Smith. Abstract

Is Joining the European Union a Good Thing? A Synthetic Control Testing the Effects on the Slovak Republic Nicole Smith Abstract After the fall of the Soviet Union in 1989, many Soviet states and satellites adopted more neoliberal and democratic policies to catch up with the rest of the world. This paper analyzes three effects associated with the Slovak Republic joining the European Union by using the synthetic control approach proposed by Abadie, Diamond, and Hainmueller. The three treatments are (1) adopting the Copenhagen Criteria, which alter the social, political, and institutional policies of the state and propose an application for membership, (2) joining the European Union in the 2004 Enlargement, and (3) adopting the euro as official currency and joining the monetary policies of the European Monetary Union. The synthetic control method shows that the Copenhagen Criteria had adverse effects on GDP per capita in the Slovak Republic, but the actual membership of the EU and the EMU increase economic growth of the Slovak Republic. 1

Economists have frequently debated the idea that poorer countries will have a higher growth rate once certain intervention measures are met in order to catch up with, or converge, with the rest of the developed world, specifically Western countries. 1 The former Soviet states of Eastern Europe are no exception to this theory, as observed through higher growth in the years following the abolishment of a centrally planned economy. Under communist rule, many of these countries missed out on a significant portion of economic growth compared to the rest of Europe, who conducted policies under neoliberalism and market freedoms, as well as receiving aid through the US Marshall Plan. Not only did these countries specifically the Slovak Republic in this study miss out on longer periods of economic liberalization and democratization that is often associated with growth, but they also had a disadvantage in terms of cultural convergence, an idea from world polity theory, as they abided by communist economics during the formation of neoliberalism and the European Union. During the last two decades, many economic and structural changes have occurred in former Soviet states and satellite states that lead to an increase in GDP per capita (in terms of purchasing power parity) as seen in Figure 1. These changes include (1) the agreement of the conditions on monetary and fiscal policy in order to achieve EU candidacy, (2) accession into the European Union, and (3) the adoption of the euro and monetary policies under the European Monetary Union. Aside from leaving central planning, the Slovak Republic adopted three policy treatments to increase economic growth; the purpose of this paper is to address which, if any, had a positive impact and the potential to lead to convergence. By conducting three separate synthetic control analyses, this paper will address issues of economic growth at each step of the way; synthetic control attempts to provide the counterfactual to what would have happened 1 Robert J.Barro, and Xavier Sala-i-Martin, Convergence, Journal of Political Economy 100.2 (1992): 223. 2

Figure 1 without treatment. In other words, I am trying to isolate the effects of each step on economic growth in the Slovak Republic. Three potential hypothesis exist: Hypothesis 1: The removal of barriers, due to agreement with the Copenhagen Criteria, found in a central planned economy provided the greatest increase in economic growth that still has impacts today. Hypothesis 2: Joining the European Union provides the best explanation for increase in economic growth in Slovak Republic as it provided a safety net and better access to markets. 3

Hypothesis 3: Adopting a common monetary policy under the European Monetary Union and the Maastricht Treaty, as well as better exchange rates under the euro provided a positive and significant impact on economic growth in the post-soviet states. In theory, all three hypotheses will have a significant effect on GDP per capita in the Slovak Republic. The outline for the rest of the paper will be an extensive literature review in Section 2. Section 3 will show the methodology of synthetic control. In Section 4, I will present the results. The last section will conclude. Literature Review There is a vast literature on economic growth in post-soviet states and EU accession, and it will be presented for each treatment in terms of before and after in order to better assess the predictions and theories of neoclassical economics. Slovak Republic was chosen for this analysis because of its proximity to the rest of the European Union and its previous relationship with the Soviet Union. Slovak Republic has the following timeline for treatments: agreement to candidacy and policy changes in 1995, as well as EU accession during the 2004 Eastern enlargement, and the adoption of the euro and monetary policy under the European Monetary Union in 2009. Copenhagen Criteria and Other Agreement Conditions In order to understand how market conditions changed with the fall of communism, we must analyze the existing Socialist structure in Slovak Republic. Under this economic system, physical assets and capital goods cannot be privately owned, investment is manipulated by the state through bank credit and rate of investment via market mechanisms, and the right to 4

manage the firm is in the hands of its employees. 2 Essentially, in a centrally planned economy, there is the absence of neoliberalism as all markets are controlled and regulated by the state. There is little debate in economic theory that centrally planned economies are harmful to economic growth, and GDP per capita in previous communist states, would have been at least three times higher on average by 1992, if they had continued the capitalist way of development after World War II. 3 Because of this, one could assume that economic growth would greatly increase with re-entry into a capitalist market, which occurred in the early 1990s for most former Soviet states and satellites. Before achieving actual accession into the European Union, candidate countries must sign an agreement that outlines criteria for monetary and fiscal policies. Specifically, the Copenhagen Criteria are concerned with political, economic, and institutional aspects of the candidate country. The political criteria require that a country have democratic stability and basic human rights for all peoples. The economic criteria outline a functioning market economy and ability to address EU competitiveness. The institutional criteria enforce the ability to maintain the existing institutional structure established by other EU members, including certain laws and administrative functions. 4 Many of these criteria challenge the existing norms of the post-soviet Slovak Republic, and they were adequately implemented before full membership was agreed upon. Under the stipulation that potential member countries will converge with the rest of the European Union, additional measures in relation to health, education, agriculture, infrastructure, 2 Svetozar Pejovich, The Firm, Monetary Policy, and Property Rights in a Planner Economy, Western Economic Journal, 7.3 (1969): 193. 3 Marek Dabrowski and Jacek Rostowski, ed. Eastern Enlargement of the EU, (New York: Springer Science+Business Media, LLC, 2001): 3. 4 "Economic Accession Criteria." European Commission. September 10, 2014. Accessed December 17, 2015. http://ec.europa.eu/economy_finance/international/enlargement/criteria/index_en.htm. 5

transportation, privatization, and welfare such as pension plans will also need to be met. 5 Even unspecified policy norms were implemented into candidate countries as those on the outskirts of the European Union adopt its rules based on a rational cost-benefit analysis. If the rule is beneficial for both domestic and EU relations or if the EU pressures the country to adopt a policy as a condition for EU membership then it will most likely be implemented. 6 As outlined above, one can infer that joining the European Union is an expensive investment and requires government spending to make up for lack of the private sector at first; however, we are testing this assumption by investigating how GDP per capita changes with or without EU membership. New European Union Countries In 2004, the European Union underwent the largest expansion of its members by adding ten countries, primarily those in Eastern Europe who were previously under Soviet rule. In theory, this would greatly contribute to high economic growth rates in member countries because liberalization occurred on industrial products only in Eastern European countries attempting to join the EU. Countries who are seeking to join the EU improve greatly in terms of industry building, employment rates, and openness to trade. 7 EU membership also provided an advantage as new EU members receive income transfers from payments into the recipient countries public budgets originate from the extension of EU structural policies to new members, and payments to suppliers in the agricultural and food sector. 8 5 Dabrowski and Rostowski, 24. 6 F. Schimmelfennig and U. Sedelmeier, Governance by conditionality: EU rule transfer to the candidate countries of Central and Eastern Europe, Journal of European Public Policy 11.4 (2004): 661 79. 7 Hubert Gabrisch. Eastern Enlargement of the European Union: Macroeconomic Effects in New Member States. Europe-Asia Studies 49 (1997): 567-590. 8 Ibid., 575. 6

Dabrowski and Rostowski claim, there is a widespread expectation that these economies will grow faster than the current EU members, gradually closing the development gap based on the analysis presented by Barro and Sala-I-Martin. 9 Their reasons for fast-growing economies in former communist countries are: they have ended central planning typically considered to be a negative economic policy, the use of learning by doing in market economy institutions, structural reforms, and other EU benefits. 10 This can be observed in the charts displaying GDP growth in previously communist countries, as the GDP greatly increases in the Slovak Republic after joining the European Union. This accelerated growth rate is partially due to being a candidate for EU accession as countries often receive pre-accession aid. Joining the EU is an expensive investment, as countries will have to meet the standards set in place by the Union, including environmental regulations, infrastructure investment, public schools, as well as relying on the private sector to provide other needs. 11 The economic growth also improves from high deficit countries that lessen their debt to GDP ratio in order to better their chances at accession. Countries often move away from agriculture and more towards industrial and financial markets; before accession, Slovak Republic had larger agricultural sectors than the EU average. 12 After accession into the EU, we can observe increased stability in the financial sectors of the Eastern European countries, especially as trust is built with the international community and a bond market emerges. 13 Broadman uses regressions to show that investment in most countries in the EU or in EU accession countries has a positive coefficient, whereas population growth has a negative 9 Dabrowski and Rostowski, 5. 10 Ibid., 5. 11 Ibid., 24. 12 Ibid., 177. 13 Corneliu Stirbu, Financial Market Integration in the Wider European Union, Hamburg Institute of International Economics Discussion Paper 297 (2004: 34 7

coefficient for non-eu countries and a positive coefficient for some EU countries. The authors summarize the effects of several factors on economic growth: investment, population growth, human capital, inflation, and government consumption. The paper argues, the accumulation of human capital is extremely important for enhancing economic growth. 14 Through this, the paper finds convergence in economic growth. They also argue that inflation appears to have a negative impact on growth, whereas government consumption and trade have an unclear result. European Monetary Union and the Maastricht Treaty Most EU members agree to join the European Monetary Union (EMU) and adopt the euro eventually. Slovak Republic adopted the euro in 2009. The Maastricht Treaty has several requirements for candidates countries that wish to adopt the euro: (1) low long-term interest rates, (2) government budget deficit, (3) low inflation cannot exceed +1.5 percentage points, and (4) nominal exchange rates in fluctuation with the euro. 15 These criteria would encourage higher economic growth rates for countries, which we observe in Figure 1. Because the euro adoption occurred either slightly before or after the financial crisis, the effects are difficult to observe. However, in theory, the adoption of a single currency in a monetary union would lead to drastic changes in both the private and public sector. 16 This can be observed in the Slovak Republic by looking at the existence of private firms and institutions 14 Harry G. Broadman, Tiiu Paas, and Paul J.J. Welfens, ed,. Economic Liberalization and Integration Policy: Options for Eastern Europe and Russia. (Berlin: Springer, 2006), 293. 15 Anna Lipinska, The Maastricht Convergence Criteria and Optimal Monetary Policy for the EMU Accession Countries, European Central Bank: Working Paper Series. No. 896 (2008): 7. 16 Carlo Monticelli and Oreste Tristani, What does the single monetary policy do?: A Svar benchmark for the European Central Bank, European Central Bank Working Paper Series (1999): 15. 8

before and after joining the European Union according to the World Development Indicators, Slovenia s domestic credit to the private sector was approximately 71% of GDP in 2013. 17 If we analyze economic growth in the Slovak Republic after the financial crisis in 2008, we observe that overall economic growth continued in the country but joblessness increase. This may be due to the continued help of foreign investment from the rest of the EU members, and the new challenge for the country is how to continue increased economic growth based solely on domestic sources rather than FDI. 18 Domestic and private investments were emphasized earlier as a means of observing economic growth in accession countries, and I argue its importance again here in a recovering EU economy. The literature that exists on new member EU countries is vast and not specifically oriented to one subject of the financial sector; however, by looking at data and regressions before and after European Union accession will provide an idea of both predictions and results. This will be crucial for understanding how economic growth changed due to the joining of one central bank, and hopefully a synthetic control or matching approach will yield results that show the difference in growth between a hypothetical Slovak Republic and the actual. Methodology: Synthetic Control This is a technique developed by Abadie, Diamond, and Hainmueller in 2010, a synthetic unit, a country in my example, is created using similarities in the entire dataset, where other units are weighted. 19 The actual country receives the treatment adoption of Copenhagen criteria, EU 17 World Development Indicators Domestic credit to private sector (% of GDP). http://data.worldbank.org/topic/private-sector?display=graph 18 Jarko Fidrmuc, Caroline Klein, Robert Price, and Andreas Worgotter, Slovakia: A Catching Up Euro Area Member In and Out of the Crisis, OECD Economics Department Working Papers, No. 1019, (2013). 19 Alberto Abadie, Alexis Diamond, and Jens Hainmueller, Comparative Politics and the Synthetic Control Method, American Journal of Political Science 00.0 (2014): 1-16. 9

accession, and EMU membership in this example whereas the synthetic country, modeled after the actual country pre-treatment and the addition of weighted countries, does not receive the treatment. This is essentially a counterfactual to what would have happened if the Slovak Republic had continued as normal. The donor units for the synthetic country are taken from the sample of countries in the world with data for GDP per capita, but exclude any possible countries that have the same treatment. Using the real gross domestic product divided by the population, I test the effect of each treatment on GDP per capita in terms of purchasing power parity (PPP), using 2005 USD from the Penn World Tables. 20 The predictor variables for each synthetic control vary by treatment; they are elaborated on in the appendix. For the Copenhagen criteria, the pre-treatment period begins in 1990 and the posttreatment period ends in 2003, which is one year before the next treatment is given. The data begins in 1990 due to data constraints from the Former Soviet Union; however, I think that it is sufficient to analyze the effects of applying for European Union membership. This treatment period is testing the effects of two policy changes that coincide with one another: the adoption of neoliberal economics and the Copenhagen criteria (which is based off of an open market economy) to replace the preexisting socialist state. There is a possibility that the results from this treatment are more due to the end of the socialist regime rather than the desire to join the EU; however, the ending of old market and the beginning of the new market are one in the same transformation. The 2004 enlargement of the European Union is captured by the time period 1996-2008 for Slovak Republic, due to the differing years in the other two treatments; this time period is forced in order to avoid the confounding events of Copenhagen criteria and the adoption of the 20 Robert C. Feenstra, Robert Inklaar and Marcel P. Timmer, "The Next Generation of the Penn World Table," forthcoming American Economic Review, (2015) available for download at www.gdpc.net/pwt 10

euro. It would be more revealing if the post-treatment period could be extended more than three years, but the third treatment interferes with this desire. Because I am testing the three different treatments, I can infer that no other manipulative events occurred during this treatment period. The existing treatment period (2005) and the final data collected by Penn World tables restrict the period of euro membership (2011). This is the shortest period, and continually updated data will be able to further elaborate the effects. The financial crisis of 2008 occurred during this time period, but it was a global effect. While it may not have equally affected each country, the treatment should not be altered by the simultaneity. Results Treatment 1: Application The first treatment tests the effect of applying for membership to the European Union and the conditionality of the Copenhagen criteria. All countries that were already members of the European Union or previously applied for membership are excluded from the synthetic. The predictor indicators are lagged values of GDP per capita for pre-treatment years (1990, 1992, and 1994), gross capital formation, percent of urban population, fertility rate, Polity Score to measure the democratic level of Slovak Republic 21 and ethnolinguistic fractionalization measure (elf15). 22 Simply, by looking at the graph presented in Figure 2, GDP per capita in the Slovak Republic has fallen behind the predictor found in the synthetic control country, which is 21 Systemic Peace Project, Polity IV. http://www.systemicpeace.org/inscrdata.html 22 William Easterly and Ross Levine. Africa s Growth Tragedy: Policies and Ethnic Divisions, The Quarterly Journal of Economics 112.4 (1997): 1203 50. 11

composed of Albania, Belarus, Canada, and Qatar. 23 The synthetic control country matches very well pretty well on the predictor indicators. Figure 2: While this gap between the treated country and the synthetic may seem large and significant, it is important to understand what this effect actually means through placebo tests of changing the year and country of the treatment, which will be elaborated on in future sections. Table 1: Match of Treated and Synthetic Countries, Treatment 1 Treated Synthetic rgdppc (1990) 14930.8 13660.73 rgdppc (1992) 10972.67 11671.09 rgdppc (1994) 11587.36 11614.5 grosscap 28.14407 20.71964 urban 56.6496 60.49301 fert 43.86028 28.56992 polity2 7 3.0176 elf15 0.3075 0.4016629 23 Qatar and Canada are geographically and compositionally different from the Slovak Republic. However, each country makes up less than one-fifth of the synthetic country. Qatar is probably a match due to the economic indicators, rather than the non-economic (polity and ethnolinguistic), as it is a donor country in several models. While Canada s high GDP per capita may skew the results, the placebo test will reveal the significance. 12

Robustness Checks In order to fully comprehend the composition of the synthetic control, we must alter the predictor variables to test for robustness of the effects. For this reason, I run the same synthetic control but simply drop the non-economic variables of world polity score and ethnolinguistic Figure 3: fractionalization. The results are consistent. The treated country has a lower GDP per capita than the synthetic country; in this case, the gap between the two appears to be greater. Slovak Republic may have been better off in the short run if they had refrained from abiding by the policy changes in the Copenhagen Criteria. The placebo test, along with the statistical match of the predictor variables, is presented in the Appendix, which reveals the significance of this gap and robustness check. Placebo tests In order to analyze the significance of the results presented in previous sections, it is essential to understand if the Slovak Republic is an outlier. For this, I run the treatment on every country in the donor pool. The results of this test are presented in Figure 4. Because the Slovak 13

Republic is not the most extreme outlier but is also not perfectly aligned with the mean effect of the treatment, I conclude that it is slightly significant, meaning that GDP per capita in the Slovak Figure 4: Republic would have been greater without the EU application process. The RMPSE for each country in the donor pool are listed in the Appendix, but the average is 330.91 and the Slovak Republic is 671.49. Treatment 2: European Union Membership In order to fully understand the effects of an open market and joining the European Union, it is essential to analyze the actual EU membership, which provides many benefits such as greater access to trade, Schengen borders that allow EU members to travel between borders, and more stable financial markets. In theory, we would expect EU membership to be significantly beneficial to the economy. We observe the effects in Figure 5. There seems to be the opposite effect from Treatment 1, where the treated country exceeds the GDP per capita of the synthetic country. This is good for the European Union, as it supports the foundation of cooperation 14

between independent nation-states. The donor countries selected for the synthetic are somewhat counterintuitive as most are in East Asia and therefore, they are not geographically similar. 24 However, Slovak Republic experienced high economic growth rates, based on GDP per capita, which is reflected by the synthetic composition of East Asian countries. Since the Slovak Figure 5: Republic is still transitioning from a communist to a democratic regime, the Polity IV index best matches with East Asian countries. Once again, we must analyze robustness and a placebo test in order to understand the significance of gap between treated and synthetic countries. An interesting improvement can be observed when comparing the changes in predictor variables between treatments, as there is a great increase in GDP per capita and fertility is nearly halved. Over time, both economic and non-economic conditions are improving in the Slovak Republic. 24 East Asian economies have had unprecedented high growth rates in recent history; this may skew the validity of the synthetic control. However, robustness checks should provide statistical evidence in favor of either situation. 15

Table 2: Match of Treated and Synthetic Countries, Treatment 2 Treated Synthetic rgdppc (1996) 12406.68 12398.43 rgdppc (2000) 12819.89 12815.76 rgdppc (2003) 14384.53 14379.47 grosscap 31.46 31.44 urban 56.23 56.71 fert 24.39 30.42 polity2 8.5 3.0176 elf15 0.307 0.326 Robustness Checks & Placebo Tests Presented below are the robustness and placebo tests for the second treatment, EU membership. The theory is the same as above, and these tests help to explain the importance and significance of Figure 5. The robustness check (Figure 6A), presents a conundrum as the noneconomic indicators added as predictor variables show the opposite result. The placebo test (Figure 6B) is Figure 6A and 6B: similar to Treatment 1, in that it the effect is not extremely significant. The average RMSPE is 478.56, and the Slovak Republic is 862.46. Essentially, a potential conclusion from the statistical analysis is that while Slovak Republic has had beneficial results from joining the European 16

Union, the effects on other countries in the area may differ and be contradictory to the Slovak Republic; in other words, there is not necessarily external validity. Treatment 3: European Monetary Union One of the stipulations of becoming a member of the European Union is to eventually abide by the monetary policies of the European Monetary Union (EMU) and adopt the euro as the official currency of the country. Slovak Republic fulfilled this requirement in 2009, immediately after the financial crisis of 2008. By joining the EMU during the crisis, Slovak Republic fared better than if they had monetary policies of their own and a currency subject to foreign exchange rates, which are sometimes volatile. Figure 7 illustrates how joining the euro assisted the Slovak Republic economy. The composition of the synthetic country further supports the effectiveness of joining the EMU, as a majority of the countries selected for the synthetic control are from Europe and half belong to the European Union (although not the EMU). Croatia, Estonia, and Romania all belong to the EU, and suffered more during the financial crisis than those countries that were subject to the euro and the EMU. Figure 7 17

Once again, we observe an improvement in GDP per capita, a reduction in the fertility rate (typically correlated with improvements in development), and a Polity IV score of that is close to 10 (the highest score of the index). Not only did joining the euro and the EMU help the Slovak Republic during the financial crisis, but also the combination of the treatments has yielded an overall improvement, both economically and non-economically. Table 3: Match of Treated and Synthetic Countries, Treatment 3 Treated Synthetic rgdppc (2005) 15720.95 15695.46 rgdppc (2007) 19477.67 19443.78 rgdppc (2008) 20996.15 20960.94 grosscap 28.99984 28.36325 urban 55.33675 59.61485 fert 20.704 29.37299 polity2 9.75 8.459 elf15 0.3075 0.2699542 Robustness Checks The robustness (Figure 8A) and placebo tests (Figure 8B) show that removing the noneconomic variables (Polity IV and ethnolinguistic fractionalization) show an even greater gap in GDP per capita between the treated and the synthetic country. This is mostly likely due to the fact that the composition of the synthetic country is made up of nearly every country in the dataset, which all performed differently during the financial crisis. The placebo test illustrates that the treated country is not statistically significant, as it is not an outlier. The average RMSPE is 526.14, and the Slovak Republic RMSPE is 467.89. The difference between the root mean squared percentage errors is small, and this effect is less significant than the other treatments. 18

Figure 8A and 8B: Conclusion Overall, there are mixed results from the synthetic control tests for the three treatments on the effects of joining the European Union. The first treatment, abiding by the policy recommendations of the Copenhagen Criteria, has a negative effect on the GDP per capita of the Slovak Republic. It would have been better for the Slovak Republic to develop without the stipulations of the Copenhagen Criteria, which are aligned with neoliberal ideologies. This negative impact is probably due to the downsides of a transitioning economy as different sectors understand their new role in investment and development of the national economy. However, after joining the European Union and the adopting the euro as official currency, the neoliberal policies and the politico-economic union policies of the European Union are helpful in improving gross domestic product for the Slovak Republic. Because of these conclusions, one might assume that the benefits of joining the EU are valid for countries elsewhere. However, the robustness checks and placebo tests question the external validity of the joining the EU. In order to fully understand the positives associated with 19

joining the European Union and adopting its official currency, synthetic control would need to be constructed for several other countries. Yet, we can conclude that it was beneficial to the Slovak Republic and its policies helped the country to develop after years of centrally planned economics under the Soviet Union. One might hypothesize that the effects of joining the European Union would be similar and positive for countries that have similar histories and composition to the Slovak Republic. Nonetheless, we know that the Slovak Republic has benefited from its membership in the European Union, and this support is crucial for current situations as there are rumors of the EU dissolving after the financial crisis and refugee crisis damage relationships internally. 20

Bibliography Abadie, Alberto, Alexis Diamond, and Jens Hainmueller. Comparative Politics and the Synthetic Control Method. American Journal of Political Science 59, no. 2 (2015): 495 510. doi:10.1111/ajps.12116. Barro, Robert J., and Xavier Sala-i-Martin. Convergence. Journal of Political Economy 100, no. 2 (1992): 223. doi:10.1086/261816. Broadman, Harry G., Tiuu Paas, and Paul J.J. Welfens. Economic Liberalization and Integration Policy: Options for Eastern Europe and Russia. Springer, 2006. doi:10.1007/s13398-014- 0173-7.2. Dabrowski, Marek, and Jacek Rostowski. The Eastern Enlargment of the EU. Springer Science+Business Media, LLC, 2001. doi:10.1007/s13398-014-0173-7.2. Easterly, William, and Ross Levine. Africa s Growth Tragedy: Policies and Ethnic Divisions. The Quarterly Journal of Economics 112, no. 4 (1997): 1203 50. "Economic Accession Criteria." European Commission. September 10, 2014. Accessed December 17, 2015. http://ec.europa.eu/economy_finance/international/enlargement/criteria/index_en.htm. Feenstra, Robert C., Robert Inklaar and Marcel P. Timmer, "The Next Generation of the Penn World Table" forthcoming American Economic Review, (2015) available for download at www.gdpc.net/pwt Fidrmuc, Jarko, Caroline Klein, R Price, and Andreas Wörgötter. Slovakia: A Catching Up Euro Area Member In and Out of the Crisis. OECD Economics Department Working Papers, No. 1019, 2013, 1 28. http://ftp.iza.org/pp55.pdf. Gabrisch, Hubert. Eastern Enlargement of the European Union : Macroeconomic Effects in New Member States. Europe-Asia Studies 49, no. 4 (2015): 567 90. Lipinska, Anna. The Maastricht Convergence Criteria and Optimal Monetary Policy for the EMU Accession Countries. ECB Working Paper Series No 896 (2008): 1 66. Monticelli, C, and O Tristani. What Does the Single Monetary Policy Do? A Svar Benchmark for the European Central Bank. European Central Bank, Working Paper Series, no. 2 (1999): 1 20. http://www.suomenpankki.fi/pdf/89902.pdf. Pejovich, Svetozar. The Firm, Monetary Policy, and Property Rights in a Planned Economy. Economic Inquiry 7, no. 3 (1969): 193 200. Schimmelfennig, F. and Sedelmeier, U. Governance by conditionality: EU rule transfer to the candidate countries of Central and Eastern Europe, Journal of European Public Policy, 21

no. 11 (2004): 661 79. Stirbu, Corneliu. Financial Market Integration in a Wider European Union. HWWA Disussion Paper, 2004, 1 40. Systemic Peace Project, Polity IV. http://www.systemicpeace.org/inscrdata.html The World Bank. 2015. World Development Indicators. 22

Appendix Table A1: Description of variables Variable Description Source rgdppc Real gross domestic product (rgdpce) divided by the population (pop) to account for purchasing power parity grosscap urban fert polity2 elf15 reserves unemp Gross capital formation (percent of GDP); measures fixed assets and inventories Urban population (percent of total population) Fertility rate, total (births per woman); number of children that would be born if a woman lived to the end of childbearing years Polity IV score; measures democratic and autocratic regimes on an index with the range: -10 to +10. Ethnolinguistic fractionalization, which is the probability that two randomly chosen people will belong to the same ethnic or linguistic family Total reserves holdings including gold, foreign exchange, and reserves of IMF members Unemployment, total (percent of total labor force); person who is without work but available and searching for employment Penn World Tables, version 8.1 World Development Indicators (2015) World Development Indicators (2015) World Development Indicators (2015) World Polity Score, Systemic Peace Project Easterly and Levine (1997) World Development Indicators (2015) World Development Indicators (2015) Table A2: Descriptive statistics for dependent and independent variables Variable Obs Mean Std. Dev. Min Max rgdppc 3,036 10837.76 13181.11 184.8175 134040.4 grosscap 2,941 23.14909 11.5964-2.42358 219.0694 urban 3,036 54.01074 23.16527 5.416 100 fert 3,036 66.44664 51.94273 1.857 228.642 polity2 3,016 3.577586 6.474695-10 10 elf15 3,036.4818783.2992333.003.9647 23

Table A3: Synthetic Composition for Treatment 1 Country Percent Albania 0.439 Belarus 0.195 Canada 0.191 Qatar 0.175 Table A4: Root Mean Square Prediction Error (RMSPE) for Treatment 1 Placebo Country RMSPE Country RMSPE Country RMSPE Albania 718.26 Gabon 348.59 Nigeria 80.86 Argentina 1124.45 Gambia 24.61 Oman 153.90 Armenia 490.69 Ghana 32.92 Pakistan 12.02 Australia 98.90 Guatemala 80.35 Panama 127.98 Azerbaijan 644.77 Guinea 15.84 Paraguay 19.90 Bahrain 807.92 Guinea-Bissau 132.58 Peru 47.80 Bangladesh 11.37 Honduras 14.37 Philippines 42.27 Belarus 317.75 India 13.99 Qatar 862.57 Benin 6.47 Indonesia 122.93 Russia 1577.85 Bhutan 54.90 Iran 281.56 Rwanda 138.44 Bolivia 10.05 Israel 440.07 Saudi Arabia 1041.01 Botswana 112.57 Jamaica 62.46 Senegal 4.70 Brazil 39.99 Japan 542.69 Sierra Leone 158.46 Burkina Faso 29.37 Jordan 160.40 Singapore 1973.95 Burundi 24.65 Kazakhstan 638.10 Slovak Rep. 671.49 Cambodia 52.43 Kenya 71.53 South Africa 85.44 Cameroon 35.07 Korea, Rep. 750.96 Sri Lanka 27.00 Canada 247.46 Kuwait 7063.35 Sudan 45.30 Central Afr. 27.17 Kyrgyz Republic 265.68 Suriname 41.75 Chad 94.24 Lesotho 33.84 Swaziland 479.38 Chile 60.71 Macedonia, FYR 178.80 Syria 37.07 China 141.53 Madagascar 11.58 Tanzania 7.85 Colombia 42.16 Malawi 35.49 Thailand 199.28 Comoros 50.92 Malaysia 19.91 Togo 51.57 Congo, Dem. 28.72 Mali 46.45 Trinidad & Tob. 154.24 Congo, Rep. 150.47 Mauritania 60.79 Tunisia 91.54 Costa Rica 36.70 Mauritius 36.06 Uganda 53.76 Cote d Ivoire 23.54 Mexico 388.98 Ukraine 221.68 Dominican Rep 72.74 Moldova 539.44 United States 5635.80 Ecuador 55.20 Mongolia 366.23 Uruguay 123.27 Egypt 60.10 Mozambique 36.72 Venezuela 313.36 El Salvador 30.94 Namibia 163.37 Vietnam 17.83 Eq. Guinea 57.62 Nepal 52.89 Yemen 14.16 Ethiopia 17.17 New Zealand 669.10 Zimbabwe 321.59 Fiji 80.40 Niger 19.58 24

Average RMPSE = 330.91 Figure A1: Placebo test for robustness check (change variables) This robustness test is significant because the actual treated company is mostly an outlier, when compared to other donor countries that receive the same treatment. In Table A4, we observe that some of the donor countries are similar throughout the models. Table A5: Synthetic Composition for Treatment 1, Robustness Change Variables Country Percent Albania 0.532 Canada 0.032 Qatar 0.436 Table A6: RMSPE for Treatment 1, Robustness Change Variables Country RMSPE Country RMSPE Country RMSPE Albania 726.76 Gabon 298.04 Nigeria 80.80 Argentina 814.50 Gambia 12.30 Oman 197.48 Armenia 410.78 Ghana 40.02 Pakistan 9.25 Australia 421.02 Guatemala 49.22 Panama 449.93 Azerbaijan 744.08 Guinea 16.61 Paraguay 41.56 25

Bahrain 233.23 Guinea-Bissau 76.39 Peru 38.61 Bangladesh 12.11 Honduras 26.11 Philippines 169.93 Belarus 319.26 India 42.87 Qatar 861.16 Benin 18.59 Indonesia 153.54 Russia 1574.80 Bhutan 104.06 Iran 153.54 Rwanda 139.88 Bolivia 18.29 Israel 987.34 Saudi Arabia 295.97 Botswana 213.95 Jamaica 72.40 Senegal 14.61 Brazil 23.89 Japan 558.53 Sierra Leone 194.60 Burkina Faso 28.72 Jordan 127.70 Singapore 2214.55 Burundi 7.67 Kazakhstan 627.42 Slovak Rep. 643.40 Cambodia 44.27 Kenya 66.13 South Africa 121.12 Cameroon 137.63 Korea, Rep. 1109.00 Sri Lanka 36.44 Canada 244.64 Kuwait 7050.28 Sudan 35.00 Central Afr. 20.30 Kyrgyz Republic 329.77 Suriname 337.09 Chad 83.62 Lesotho 83.72 Swaziland 283.90 Chile 54.97 Macedonia, FYR 132.85 Syria 21.30 China 129.85 Madagascar 17.94 Tanzania 12.27 Colombia 56.58 Malawi 35.06 Thailand 722.88 Comoros 27.55 Malaysia 4.80 Togo 79.98 Congo, Dem. 26.24 Mali 34.04 Trinidad & Tob. 221.95 Congo, Rep. 158.47 Mauritania 53.71 Tunisia 10.12 Costa Rica 236.49 Mauritius 534.43 Uganda 81.10 Cote d Ivoire 18.77 Mexico 37.15 Ukraine 347.21 Dominican Rep 67.85 Moldova 472.51 United States 5635.80 Ecuador 56.29 Mongolia 283.99 Uruguay 78.93 Egypt 9.56 Mozambique 33.15 Venezuela 175.25 El Salvador 39.11 Namibia 161.57 Vietnam 17.90 Eq. Guinea 56.81 Nepal 51.32 Yemen 12.91 Ethiopia 18.77 New Zealand 748.83 Zimbabwe 412.03 Fiji 30.40 Niger 24.80 Average RMSPE = 343.24 Table A7: Synthetic Composition for Treatment 2 Country Percent Equatorial Guinea 0.028 Jamaica 0.075 Japan 0.109 Korea, Rep. 0.326 Mongolia 0.092 Thailand 0.37 26

Table A8: RMSPE for Treatment 2 Placebo Country RMSPE Country RMSPE Country RMSPE Albania 228.43 Gabon 624.01 Nigeria 143.13 Angola 274.57 Gambia 52.19 Norway 1213.87 Argentina 288.06 Georgia 104.12 Oman 438.50 Armenia 72.57 Ghana 137.99 Pakistan 57.84 Australia 1851.31 Guatemala 109.46 Panama 578.77 Azerbaijan 107.6 Guinea 148.10 Paraguay 190.61 Bahrain 889.64 Guinea-Bisseau 102.12 Peru 135.53 Bangladesh 25.73 Honduras 83.63 Philippines 153.18 Belarus 208.25 India 74.13 Qatar 14117.68 Benin 30.12 Indonesia 153.27 Romania 275.45 Bhutan 312.11 Israel 487.90 Russia 504.51 Bolivia 106.29 Jamaica 146.01 Rwanda 23.96 Botswana 949.17 Japan 2470.02 Saudi Arabia 916.25 Brazil 137.02 Jordan 164.24 Senegal 28.92 Bulgaria 227.24 Kazakhstan 130.65 Sierra Leone 121.57 Burkina Faso 28.15 Kenya 57.27 Singapore 744.51 Burundi 26.23 Korea, Rep 492.20 Slovak Rep 862.46 Cambodia 49.08 Kuwait 3906.12 South Africa 268.83 Cameroon 136.57 Kyrgyz Rep 169.02 Sri Lanka 146.99 Canada 923.04 Lao PDR 51.99 Sudan 121.17 Central Afr. 11.16 Lesotho 202.25 Suriname 400.39 Chad 105.24 Liberia 139.32 Swaziland 915.82 Chile 342.09 Macedonia 114.60 Switzerland 787.70 China 94.26 Madagascar 33.35 Syria 70.07 Columbia 140.93 Malawi 34.26 Tanzania 16.71 Comoros 46.09 Malaysia 218.87 Thailand 276.02 Congo, Dem 205.88 Mali 44.59 Togo 41.50 Congo, Rep 415.22 Mauritania 59.80 Trinidad & Tob 776.14 Costa Rica 388.49 Mauritius 237.12 Tunisia 84.95 Cote d Ivoire 159.92 Mexico 400.33 Turkey 283.80 Croatia 604.45 Moldova 135.31 Uganda 35.31 Dominican Rep 451.56 Mongolia 43.61 Ukraine 33.36 Ecuador 218.53 Morocco 164.61 United States 2999.95 Egypt 47.53 Mozambique 40.42 Uruguary 159.16 El Salvador 33.15 Namibia 79.53 Venezuela 482.34 Eq. Guinea 1935.62 Nepal 25.72 Vietnam 100.89 Ethiopia 56.11 New Zealand 1460.53 Yemen 85.73 Fiji 168.93 Niger 36.25 Average RMSPE = 478.56 27

Figure A2: Placebo test for robustness check (change variables) This robustness test is also somewhat significant, as the Slovak Republic is somewhat of an outlier, but not extremely. Table A9: RMSPE for Treatment 2, Robustness Change Variables Country RMSPE Country RMSPE Country RMSPE Albania 228.43 Gabon 624.01 Nigeria 143.13 Angola 274.57 Gambia 52.19 Norway 1213.87 Argentina 288.06 Georgia 104.12 Oman 438.50 Armenia 72.57 Ghana 137.99 Pakistan 57.84 Australia 1851.31 Guatemala 109.46 Panama 578.77 Azerbaijan 107.60 Guinea 148.10 Paraguay 190.61 Bahrain 889.64 Guinea-Bissau 102.12 Peru 135.53 Bangladesh 25.73 Honduras 83.63 Philippines 153.18 Belarus 208.35 India 74.13 Qatar 14117.68 Benin 30.13 Indonesia 153.37 Romania 275.45 Bhutan 312.11 Israel 487.90 Russia 504.51 Bolivia 106.29 Jamaica 146.01 Rwanda 23.96 Botswana 949.17 Japan 2470.02 Saudi Arabia 916.25 Brazil 137.02 Jordan 164.24 Senegal 38.92 Bulgaria 277.24 Kazakhstan 130.65 Sierra Leone 121.58 Burkina Faso 28.15 Kenya 57.27 Singapore 744.51 Burundi 26.22 Korea, Rep 492.20 Slovak Rep 862.46 Cambodia 49.98 Kuwait 3906.12 South Africa 268.83 28

Cameroon 136.58 Kyrgyz Rep 169.02 Sri Lanka 146.99 Canada 923.04 Lao PDR 51.99 Sudan 121.17 Central Afr. 11.16 Lesotho 202.25 Suriname 400.39 Chad 105.24 Liberia 139.32 Swaziland 915.82 Chile 342.09 Macedonia 114.60 Switzerland 787.69 China 94.26 Madagascar 33.35 Syria 70.07 Columbia 140.93 Malawi 34.26 Tanzania 16.71 Comoros 46.07 Malaysia 318.87 Thailand 276.02 Congo, Dem 205.88 Mali 44.59 Togo 42.50 Congo, Rep 415.22 Mauritania 59.80 Trinidad & Tob 776.14 Costa Rica 388.49 Mauritius 237.12 Tunisia 84.96 Cote d Ivoire 159.92 Mexico 400.33 Turkey 283.80 Croatia 604.45 Moldova 135.31 Uganda 35.305 Dominican Rep 451.56 Mongolia 43.61 Ukraine 33.36 Ecuador 218.53 Morocco 164.61 United States 2999.95 Egypt 47.53 Mozambique 40.42 Uruguay 159.16 El Salvador 33.15 Namibia 79.53 Venezuela 483.34 Eq. Guinea 1935.62 Nepal 25.72 Vietnam 100.89 Ethiopia 56.11 New Zealand 1460.53 Yemen 85.73 Fiji 168.93 Niger 36.25 Average RMSPE = 479.54 Table A10: Synthetic Composition of Countries, Treatment 3 Country Percent Burundi 0.011 Croatia 0.065 Estonia 0.151 Kuwait 0.039 Romania 0.608 Switzerland 0.0125 Table A11: RMSPE for Treatment 3 Placebo Country RMSPE Country RMSPE Country RMSPE Albania 1.22 Gabon 171.23 Niger 48.16 Angola 35.73 Gambia 36.98 Nigeria 65.26 Argentina 16.72 Georgia 48.03 Norway 114.89 Armenia 97.66 Ghana 20.57 Oman 1470.10 Australia 230.81 Guatemala 73.79 Pakistan 7.91 Azerbaijan 883.92 Guinea 18.94 Panama 89.60 Bahrain 1584.61 Guinea-Bissau 26.59 Paraguay 45.39 Bangladesh 5.33 Honduras 22.85 Peru 21.12 Belarus 1.43 Hungary 91.37 Philippines 12.16 29

Benin 26.93 India 1.20 Poland 339.28 Bhutan 205.83 Indonesia 3.23 Qatar 42823.29 Bolivia 68.10 Israel 151.48 Romania 115.83 Botswana 212.01 Jamaica 79.01 Russia 312.13 Brazil 30.09 Japan 73.78 Rwanda 19.79 Bulgaria 199.52 Jordan 10.62 Saudi Arabia 185.15 Burkina Faso 4.05 Kazakhstan 33.24 Senegal 37.02 Burundi 61.31 Kenya 20.83 Sierra Leone 4.42 Cambodia 32.62 Korea, Rep. 7.32 Singapore 1857.57 Cameroon 8.95 Kuwait 1053.29 Slovak Rep. 467.89 Canada 23.32 Kyrgyz Rep. 34.85 South Africa 11.31 Central Afr. 32.18 Lao PRD 78.09 Sri Lanka 8.55 Chad 78.62 Latvia 173.98 Sudan 59.32 Chile 200.20 Lesotho 2.99 Swaziland 241.40 China 2.48 Liberia 6.85 Sweden 400.44 Colombia 41.17 Lithuania 250.44 Switzerland 1120.91 Comoros 23.54 Macedonia 85.01 Tanzania 9.62 Congo, Dem. 156.35 Madagascar 6.13 Thailand 11.94 Congo, Rep. 228.52 Malawi 27.64 Togo 23.08 Costa Rica 151.96 Malaysia 142.53 Tunisia 9.80 Cote d Ivoire 60.24 Mali 7.69 Turkey 65.47 Croatia 393.53 Mauritania 183.22 Uganda 7.05 Czech Rep. 269.14 Mauritius 141.88 Ukraine 117.82 Denmark 129.10 Mexico 20.57 United King. 94.19 Dominican R 104.87 Moldova 15.00 United States 888.84 Ecuador 24.83 Mongolia 63.81 Uruguay 95.66 Egypt 29.53 Morocco 5.61 Venezuela 70.09 El Salvador 20.80 Mozambique 5.24 Vietnam 8.78 Eq. Guinea 944.48 Namibia 71.44 Zimbabwe 232.54 Estonia 144.09 Nepal 8.07 Ethiopia 15.01 New Zealand 513.78 Average RMSPE = 536.13 Figure A3: Placebo test for robustness check (change variables) 30

This robustness check for the placebo test is the least significant of all the robustness checks because it is near the average of the effect of treatments. Table A12: Placebo Test for Treatment 3, Robustness Country RMSPE Country RMSPE Country RMSPE Albania 1.22 Gabon 171.23 Niger 48.16 Angola 35.73 Gambia 36.98 Nigeria 65.26 Argentina 16.72 Georgia 48.03 Norway 114.89 Armenia 97.66 Ghana 20.57 Oman 1470.10 Australia 230.81 Guatemala 73.79 Pakistan 7.91 Azerbaijan 883.92 Guinea 18.94 Panama 89.60 Bahrain 1584.61 Guinea-Bissau 26.59 Paraguay 45.39 Bangladesh 5.33 Honduras 22.85 Peru 21.12 Belarus 1.43 Hungary 91.37 Philippines 12.16 Benin 26.93 India 1.20 Poland 339.28 Bhutan 205.84 Indonesia 3.23 Qatar 42823.29 Bolivia 68.10 Israel 151.48 Romania 115.83 Botswana 212.01 Jamaica 79.01 Russia 312.13 Brazil 30.09 Japan 73.78 Rwanda 19.79 Bulgaria 199.52 Jordan 10.62 Saudi Arabia 185.15 Burkina Faso 4.05 Kazakhstan 33.24 Senegal 37.02 Burundi 61.31 Kenya 20.83 Sierra Leone 4.42 Cambodia 32.62 Korea, Rep. 7.32 Singapore 1857.57 Cameroon 8.96 Kuwait 1053.29 Slovak Rep. 467.89 Canada 23.32 Kyrgyz Rep. 34.85 South Africa 11.31 Central Afr. 32.18 Lao PRD 78.98 Sri Lanka 8.55 Chad 78.62 Latvia 173.98 Sudan 59.32 Chile 200.20 Lesotho 2.99 Swaziland 241.40 China 2.48 Liberia 6.85 Sweden 400.44 Colombia 41.17 Lithuania 250.44 Switzerland 1120.91 Comoros 23.54 Macedonia 85.01 Tanzania 9.62 Congo, Dem. 156.35 Madagascar 6.13 Thailand 11.94 Congo, Rep. 228.52 Malawi 27.64 Togo 23.08 Costa Rica 151.96 Malaysia 142.53 Tunisia 9.80 Cote d Ivoire 60.24 Mali 7.69 Turkey 65.47 Croatia 393.53 Mauritania 183.22 Uganda 7.05 Czech Rep. 269.10 Mauritius 141.88 Ukraine 117.82 Denmark 129.10 Mexico 20.57 United King. 94.19 Dominican R 104.87 Moldova 15.00 United States 888.84 Ecuador 24.83 Mongolia 63.81 Uruguay 95.66 Egypt 29.53 Morocco 5.61 Venezuela 70.09 El Salvador 20.80 Mozambique 5.34 Vietnam 8.78 Eq. Guinea 944.48 Namibia 71.44 Zimbabwe 232.54 31

Estonia 144.09 Nepal 8.07 Ethiopia 15.01 New Zealand 513.78 Average RMSPE = 526.14 32