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

Transcription

1 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

2 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 (1992):

3 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

4 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 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

5 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): 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, Accessed December 17,

6 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, 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): Hubert Gabrisch. Eastern Enlargement of the European Union: Macroeconomic Effects in New Member States. Europe-Asia Studies 49 (1997): Ibid.,

7 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, Ibid., Ibid., Ibid., Corneliu Stirbu, Financial Market Integration in the Wider European Union, Hamburg Institute of International Economics Discussion Paper 297 (2004: 34 7

8 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 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), Anna Lipinska, The Maastricht Convergence Criteria and Optimal Monetary Policy for the EMU Accession Countries, European Central Bank: Working Paper Series. No. 896 (2008): 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

9 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 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) 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):

10 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 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 10

11 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 William Easterly and Ross Levine. Africa s Growth Tragedy: Policies and Ethnic Divisions, The Quarterly Journal of Economics (1997):

12 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) rgdppc (1992) rgdppc (1994) grosscap urban fert polity elf 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

13 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

14 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 and the Slovak Republic is 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

15 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

16 Table 2: Match of Treated and Synthetic Countries, Treatment 2 Treated Synthetic rgdppc (1996) rgdppc (2000) rgdppc (2003) grosscap urban fert polity elf 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 , and the Slovak Republic is Essentially, a potential conclusion from the statistical analysis is that while Slovak Republic has had beneficial results from joining the European 16

17 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 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

18 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) rgdppc (2007) rgdppc (2008) grosscap urban fert polity elf 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 , and the Slovak Republic RMSPE is The difference between the root mean squared percentage errors is small, and this effect is less significant than the other treatments. 18

19 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

20 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

21 Bibliography Abadie, Alberto, Alexis Diamond, and Jens Hainmueller. Comparative Politics and the Synthetic Control Method. American Journal of Political Science 59, no. 2 (2015): doi: /ajps Barro, Robert J., and Xavier Sala-i-Martin. Convergence. Journal of Political Economy 100, no. 2 (1992): 223. doi: / Broadman, Harry G., Tiuu Paas, and Paul J.J. Welfens. Economic Liberalization and Integration Policy: Options for Eastern Europe and Russia. Springer, doi: /s Dabrowski, Marek, and Jacek Rostowski. The Eastern Enlargment of the EU. Springer Science+Business Media, LLC, doi: /s Easterly, William, and Ross Levine. Africa s Growth Tragedy: Policies and Ethnic Divisions. The Quarterly Journal of Economics 112, no. 4 (1997): "Economic Accession Criteria." European Commission. September 10, Accessed December 17, 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 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, Gabrisch, Hubert. Eastern Enlargement of the European Union : Macroeconomic Effects in New Member States. Europe-Asia Studies 49, no. 4 (2015): Lipinska, Anna. The Maastricht Convergence Criteria and Optimal Monetary Policy for the EMU Accession Countries. ECB Working Paper Series No 896 (2008): 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): Pejovich, Svetozar. The Firm, Monetary Policy, and Property Rights in a Planned Economy. Economic Inquiry 7, no. 3 (1969): 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

22 no. 11 (2004): Stirbu, Corneliu. Financial Market Integration in a Wider European Union. HWWA Disussion Paper, 2004, Systemic Peace Project, Polity IV. The World Bank World Development Indicators. 22

23 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, grosscap 2, urban 3, fert 3, polity2 3, elf15 3,

24 Table A3: Synthetic Composition for Treatment 1 Country Percent Albania Belarus Canada Qatar Table A4: Root Mean Square Prediction Error (RMSPE) for Treatment 1 Placebo Country RMSPE Country RMSPE Country RMSPE Albania Gabon Nigeria Argentina Gambia Oman Armenia Ghana Pakistan Australia Guatemala Panama Azerbaijan Guinea Paraguay Bahrain Guinea-Bissau Peru Bangladesh Honduras Philippines Belarus India Qatar Benin 6.47 Indonesia Russia Bhutan Iran Rwanda Bolivia Israel Saudi Arabia Botswana Jamaica Senegal 4.70 Brazil Japan Sierra Leone Burkina Faso Jordan Singapore Burundi Kazakhstan Slovak Rep Cambodia Kenya South Africa Cameroon Korea, Rep Sri Lanka Canada Kuwait Sudan Central Afr Kyrgyz Republic Suriname Chad Lesotho Swaziland Chile Macedonia, FYR Syria China Madagascar Tanzania 7.85 Colombia Malawi Thailand Comoros Malaysia Togo Congo, Dem Mali Trinidad & Tob Congo, Rep Mauritania Tunisia Costa Rica Mauritius Uganda Cote d Ivoire Mexico Ukraine Dominican Rep Moldova United States Ecuador Mongolia Uruguay Egypt Mozambique Venezuela El Salvador Namibia Vietnam Eq. Guinea Nepal Yemen Ethiopia New Zealand Zimbabwe Fiji Niger

25 Average RMPSE = 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 Canada Qatar Table A6: RMSPE for Treatment 1, Robustness Change Variables Country RMSPE Country RMSPE Country RMSPE Albania Gabon Nigeria Argentina Gambia Oman Armenia Ghana Pakistan 9.25 Australia Guatemala Panama Azerbaijan Guinea Paraguay

26 Bahrain Guinea-Bissau Peru Bangladesh Honduras Philippines Belarus India Qatar Benin Indonesia Russia Bhutan Iran Rwanda Bolivia Israel Saudi Arabia Botswana Jamaica Senegal Brazil Japan Sierra Leone Burkina Faso Jordan Singapore Burundi 7.67 Kazakhstan Slovak Rep Cambodia Kenya South Africa Cameroon Korea, Rep Sri Lanka Canada Kuwait Sudan Central Afr Kyrgyz Republic Suriname Chad Lesotho Swaziland Chile Macedonia, FYR Syria China Madagascar Tanzania Colombia Malawi Thailand Comoros Malaysia 4.80 Togo Congo, Dem Mali Trinidad & Tob Congo, Rep Mauritania Tunisia Costa Rica Mauritius Uganda Cote d Ivoire Mexico Ukraine Dominican Rep Moldova United States Ecuador Mongolia Uruguay Egypt 9.56 Mozambique Venezuela El Salvador Namibia Vietnam Eq. Guinea Nepal Yemen Ethiopia New Zealand Zimbabwe Fiji Niger Average RMSPE = Table A7: Synthetic Composition for Treatment 2 Country Percent Equatorial Guinea Jamaica Japan Korea, Rep Mongolia Thailand

27 Table A8: RMSPE for Treatment 2 Placebo Country RMSPE Country RMSPE Country RMSPE Albania Gabon Nigeria Angola Gambia Norway Argentina Georgia Oman Armenia Ghana Pakistan Australia Guatemala Panama Azerbaijan Guinea Paraguay Bahrain Guinea-Bisseau Peru Bangladesh Honduras Philippines Belarus India Qatar Benin Indonesia Romania Bhutan Israel Russia Bolivia Jamaica Rwanda Botswana Japan Saudi Arabia Brazil Jordan Senegal Bulgaria Kazakhstan Sierra Leone Burkina Faso Kenya Singapore Burundi Korea, Rep Slovak Rep Cambodia Kuwait South Africa Cameroon Kyrgyz Rep Sri Lanka Canada Lao PDR Sudan Central Afr Lesotho Suriname Chad Liberia Swaziland Chile Macedonia Switzerland China Madagascar Syria Columbia Malawi Tanzania Comoros Malaysia Thailand Congo, Dem Mali Togo Congo, Rep Mauritania Trinidad & Tob Costa Rica Mauritius Tunisia Cote d Ivoire Mexico Turkey Croatia Moldova Uganda Dominican Rep Mongolia Ukraine Ecuador Morocco United States Egypt Mozambique Uruguary El Salvador Namibia Venezuela Eq. Guinea Nepal Vietnam Ethiopia New Zealand Yemen Fiji Niger Average RMSPE =

28 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 Gabon Nigeria Angola Gambia Norway Argentina Georgia Oman Armenia Ghana Pakistan Australia Guatemala Panama Azerbaijan Guinea Paraguay Bahrain Guinea-Bissau Peru Bangladesh Honduras Philippines Belarus India Qatar Benin Indonesia Romania Bhutan Israel Russia Bolivia Jamaica Rwanda Botswana Japan Saudi Arabia Brazil Jordan Senegal Bulgaria Kazakhstan Sierra Leone Burkina Faso Kenya Singapore Burundi Korea, Rep Slovak Rep Cambodia Kuwait South Africa

29 Cameroon Kyrgyz Rep Sri Lanka Canada Lao PDR Sudan Central Afr Lesotho Suriname Chad Liberia Swaziland Chile Macedonia Switzerland China Madagascar Syria Columbia Malawi Tanzania Comoros Malaysia Thailand Congo, Dem Mali Togo Congo, Rep Mauritania Trinidad & Tob Costa Rica Mauritius Tunisia Cote d Ivoire Mexico Turkey Croatia Moldova Uganda Dominican Rep Mongolia Ukraine Ecuador Morocco United States Egypt Mozambique Uruguay El Salvador Namibia Venezuela Eq. Guinea Nepal Vietnam Ethiopia New Zealand Yemen Fiji Niger Average RMSPE = Table A10: Synthetic Composition of Countries, Treatment 3 Country Percent Burundi Croatia Estonia Kuwait Romania Switzerland Table A11: RMSPE for Treatment 3 Placebo Country RMSPE Country RMSPE Country RMSPE Albania 1.22 Gabon Niger Angola Gambia Nigeria Argentina Georgia Norway Armenia Ghana Oman Australia Guatemala Pakistan 7.91 Azerbaijan Guinea Panama Bahrain Guinea-Bissau Paraguay Bangladesh 5.33 Honduras Peru Belarus 1.43 Hungary Philippines

30 Benin India 1.20 Poland Bhutan Indonesia 3.23 Qatar Bolivia Israel Romania Botswana Jamaica Russia Brazil Japan Rwanda Bulgaria Jordan Saudi Arabia Burkina Faso 4.05 Kazakhstan Senegal Burundi Kenya Sierra Leone 4.42 Cambodia Korea, Rep Singapore Cameroon 8.95 Kuwait Slovak Rep Canada Kyrgyz Rep South Africa Central Afr Lao PRD Sri Lanka 8.55 Chad Latvia Sudan Chile Lesotho 2.99 Swaziland China 2.48 Liberia 6.85 Sweden Colombia Lithuania Switzerland Comoros Macedonia Tanzania 9.62 Congo, Dem Madagascar 6.13 Thailand Congo, Rep Malawi Togo Costa Rica Malaysia Tunisia 9.80 Cote d Ivoire Mali 7.69 Turkey Croatia Mauritania Uganda 7.05 Czech Rep Mauritius Ukraine Denmark Mexico United King Dominican R Moldova United States Ecuador Mongolia Uruguay Egypt Morocco 5.61 Venezuela El Salvador Mozambique 5.24 Vietnam 8.78 Eq. Guinea Namibia Zimbabwe Estonia Nepal 8.07 Ethiopia New Zealand Average RMSPE = Figure A3: Placebo test for robustness check (change variables) 30

31 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 Niger Angola Gambia Nigeria Argentina Georgia Norway Armenia Ghana Oman Australia Guatemala Pakistan 7.91 Azerbaijan Guinea Panama Bahrain Guinea-Bissau Paraguay Bangladesh 5.33 Honduras Peru Belarus 1.43 Hungary Philippines Benin India 1.20 Poland Bhutan Indonesia 3.23 Qatar Bolivia Israel Romania Botswana Jamaica Russia Brazil Japan Rwanda Bulgaria Jordan Saudi Arabia Burkina Faso 4.05 Kazakhstan Senegal Burundi Kenya Sierra Leone 4.42 Cambodia Korea, Rep Singapore Cameroon 8.96 Kuwait Slovak Rep Canada Kyrgyz Rep South Africa Central Afr Lao PRD Sri Lanka 8.55 Chad Latvia Sudan Chile Lesotho 2.99 Swaziland China 2.48 Liberia 6.85 Sweden Colombia Lithuania Switzerland Comoros Macedonia Tanzania 9.62 Congo, Dem Madagascar 6.13 Thailand Congo, Rep Malawi Togo Costa Rica Malaysia Tunisia 9.80 Cote d Ivoire Mali 7.69 Turkey Croatia Mauritania Uganda 7.05 Czech Rep Mauritius Ukraine Denmark Mexico United King Dominican R Moldova United States Ecuador Mongolia Uruguay Egypt Morocco 5.61 Venezuela El Salvador Mozambique 5.34 Vietnam 8.78 Eq. Guinea Namibia Zimbabwe

32 Estonia Nepal 8.07 Ethiopia New Zealand Average RMSPE =