Hlavní práce73236_novt14.pdf, 0.9 MB Stáhnout

(1)

University of Economics, Prague Faculty of Economics

Field of Study: Economics

I ^S B ^ULGARIA B ^{ETTER OFF} A ^FTER J OINING THE E ^UROPEAN U ^NION ?

Bachelor thesis

Author: Tomáš Novák

Thesis supervisor: Ing. Martin Slaný, Ph.D.

(2)

Declaration of Authorship

This thesis is my original work and has not been previously submitted for an examination that has led to the award of a degree.

To the best of my knowledge and belief, this thesis contains no material previously published or written by another person except where due reference is made.

Signed:

Date:

(3)

Acknowledgements

Thank you to my thesis director, Ing. Martin Slaný, Ph.D., for his valuable comments and suggestions.

Thank you to University of Economics for creating such an inspiring environment. I would like to express my gratitude to Ing. Pavel Potužák, Ph.D., who has had a particular influence on my appreciation of macroeconomics and mathematics. I also thank my classmates for creating a friendly and inspiring environment.

Thank you to Ing. František Mašek for his valuable comments.

Thank you to my parents, Jitka and Pavel, to my brother Michal all of whom provided

(4)

Abstract

This thesis evaluates the quantitative effects of Bulgaria’s accession to the European Union in 2007. The dependent variables are GDP per capita and growth rate of GDP.

Using the Synthetic Control Method was created a control unit, which is used for estimating the development of the two analysed variables in the absence of the shock, which is represented by the entrance to the EU in the year 2007. The results show, that the accession had a positive effect on GDP per capita and a negative effect on the growth rate of GDP. Particularly, the accumulated gain is estimated at 11319,4 USD for GDP per capita and the loss is estimated at 7,9 pp. for the growth rate of GDP. Both accumulated gaps are within the observed period 2007-2018. Placebo tests showed, that the obtained results were not statistically significant.

Keywords

Bulgaria, European Union, GDP per capita, real GDP growth, Synthetic Control Method

Abstrakt

Tato práce kvantitativně vyhodnocuje efekt vstupu Bulharska do Evropské unie v roce 2007. Zkoumané proměnné jsou HDP na obyvatele a tempo růstu HDP. Za pomocí metody Synthetic control method byla vytvořena kontrolní jednotka, pomocí níž byl sledován vývoj sledovaných proměnných v případě absence šoku, který v tomto případě představuje vstup do EU v roce 2007. Výsledky ukazují, že vstup měl pozitivní vliv na HDP na obyvatele a negativní efekt na tempo růstu HDP. Akumulovaný zisk je 11319,4 USD pro proměnnou HDP na hlavu a akumulovaná ztráta je 7,9 pp. pro proměnnou tempo růstu HDP. Oba akumulované rozdíly byly sledovány pro období 2007-2018. Placebo testy ukázaly, že získané výsledky nejsou statisticky signifikantní.

Klíčová slova

Bulharsko, Evropská unie, HDP na osobu, tempo růstu HDP, Synthetic Control Method

(5)

L IST OF T ABLES

Table 1 - Predictor Means Before the Intervention ... 44

Table 2 - In Model Included Predictor Means Before the Intervention ... 45

Table 3 - Country’s Weights ... 45

Table 4 - Gap in the GDP per Capita Before Entering the EU ... 46

Table 5 - Gap in the GDP per Capita After Entering the EU ... 46

Table 6 - Predictor Means Before the Intervention ... 49

Table 7 - In Model Included Predictor Means Before the Intervention ... 50

Table 9 - Gap in the Real GDP Growth Before Entering the EU ... 51

Table 10 - Gap in the Real GDP Growth After Entering the EU ... 51

Table 11 - Predictor Means Before the Intervention (Comparison of Both Models) .... 55

Table 16 - Country’s Weights (Comparison) ... 61

Table 18 - Country’s Weights (Comparison) ... 64

(8)

L IST OF F IGURES

Figure 1 - Estimation of the effect of the accession to the EU on GDP per capita 1991-

2018 ... 43

Figure 2 - Gap in the GDP per Capita 1991-2018 ... 47

Figure 3 - Estimation of the effect of the accession to the EU on the real GDP growth 1991-2018 ... 48

Figure 4 - Gap in the Real GDP Growth 1991-2018 ... 52

Figure 5 - Estimation of the Effect on GDP per Capita After Changing the Time Optimizationperiod to 2004 ... 54

Figure 6 - Estimation of the Effect on the Growth Rate of GDP After Changing the Time Optimizationperiod to 2004 ... 57

Figure 7 - Placebo Test (North Macedonia) 1991-2018 ... 59

Figure 8 - Placebo Test (North Macedonia) 1991-2018 ... 62

Figure 9 - Gap in the Real GDP Growth for Macedonia 1991-2018 ... 65

Figure 10 - Robustness Test - Simulation for All Units for GDP per Capita, Ppp (Constant 2011 International $) ... 66

Figure 11 - Robustness Test – Rmspe Ratios for GDP per Capita, Ppp (Con- Stant 2011 International $) ... 67

Figure 12 - Robustness test - simulation for all units for the rate of GDP growth (annual %) ... 68

Figure 13 - Robustness Test – Rmspe Ratios for the Rate of GDP Growth (Annual %) ... 69

Figure 14 - Control countries ... 77

Figure 15 - Description of variables (A) ... 78

Figure 16 - Description of variables (B) ... 79

(9)

A CRONYMS

EU European Union

WB The World Bank

GDP Gross Domestic Product SC Synthetic Control

SCM The Synthetic Control Method RMSPE Root Mean Square Predicted Error

PHARE Poland and Hungary: Assistance for Restructuring Their Economies N. Macedonia North Macedonia

(10)

I NTRODUCTION

On the 1st January 2007, the sixth enlargement of the European Union took place, with the accession of Bulgaria and Romania. Both countries have applied a number of democratic processes and undergone economic transformation. Bulgaria and Romania began to change the structure of their economies after the collapse of the communist regime in the early 1990s. Both states are among the least developed European countries and Bulgaria is the weakest member of the European Union when its GDP per capita (in constant 2017 international $) accounted to 22 181.5 USD in 2018.

The goal of this thesis is to evaluate the effect of the intervention on the Bulgaria GDP per capita and growth rate of GDP. The effect will be estimated using the Synthetic Control Method, which was originally presented by Abadie & Gardeazabal (2003), and which creates a contra-factual Bulgaria in the absence of the treatment. Synthetic Bulgaria will be created by finding the weighted average of countries with similar characteristics to the Bulgaria indicators before the accession. The data-driven approach prevents us from choosing the comparative unit based on unspecified characteristics, while the counter-factual is constructed by building a synthetic control group and the method guards against extrapolation outside the convex hull of the data because the sum of the chosen control units has to sum to one. The results showed, that the accession had a positive effect on GDP per capita and negative effect on the growth rate of GDP when the accumulated positive effect within the observed period 2007-2018 was 11319,4 USD for GDP per capita and the negative effect accumulated effect was -7,9 pp. for the growth rate of GDP.

This thesis is structured as follows. In chapter 1 will presented economic growth models and the effect of institutions on the economy. In chapter 2 will be analysed Bulgaria's’ economic transformation, followed by its European Union accession. In chapter 4 will be described the Synthetic Control Method, including its application in economic empirical studies and the tests of robustness. the next chapter is about the data, including the arguments based on which the countries were selected for the donor pool.

The results are presented in the following chapter followed by a chapter showing the results of the placebo tests. There were made three types of placebo tests. At first, was changed the time of intervention, then was changed the treated unit and in the last placebo

(11)

test was analysed the gap between the real and its synthetic unit for each state of the donor pool. The last chapter summarizes and discusses the obtained results.

(12)

1 E CONOMIC G ROWTH

In the first chapter called Economic Growth will be presented two growth models followed by a look at institutions concerning economic growth. Firstly, will be presented the Solow model and its structure followed by an Endogenous Growth model. Those two models will be followed by the role of institutions in the economy. At first, will be presented North view on institutions and how they interact with the economic environment and then will be presented empirical studies examining the economic growth concerning institutions..

1.1 Solow Model

The Solow model was developed by Robert Solow, who was born on the 23rd of August 1942 in Brooklyn. Solow received a scholarship on Harvard University in 1940 and began his studies of sociology, anthropology and economics. In 1942, he left the university and entered the army. After the second world war, he returned to Harvard and served as a research assistant of Wassily Leontief. In the years 1949-1950, he spent one year at Columbia University, where he studies statistics. Since 1950, Robert Solow worked as an assistant professor at M.I.T. During his cooperation with Paul Samuelson was created many theories. In 1997, Solow received the Nobel prize for his contributions to the theory of economic growth (Encyclopædia Britannica, inc., 2019).

The Solow model is a type of dynamic growth model, which allows us to analyse economy growth in time. The sources of the total income of the economy in the Solow model are the factors of production, such as capital and labour, and the production technology (Mankiw et al., 1992).

Assumptions

The supply of goods in the Solow model is based on the neoclassical production function with diminishing marginal product of capital. The production function consists of Y as the output, K as the physical capital, A as the level of technology and L as the labour (Mankiw et al., 1992). Product of A and L determines the effective labour. The function is described as follows:

(13)

𝑌 = 𝐹(𝐾, 𝐴𝐿), (1.1) The model assumes exogenous savings and population growth. Technological progress and capital accumulation are exogenous as well. Steady-state capital per capita per worker is also determined exogenously by the growth rate of population and savings (Mankiw et al., 1992). The production function is described by Cobb-Douglas production function at time t as:

𝑌(𝑡) = 𝐾(𝑡)⁺(𝐴(𝑡)𝐿(𝑡))^,-+, (1.2) The model assumes, that there is a certain amount of working population denoted by L, which grows at exogenous growth rate n and technological progress is denoted by A, which grows at rate g.

One of the characteristics of the production function is, that the function has constant returns to scale. Which means, that increase in capital and labour lead to an equivalent increase in the output. We might denote the constant returns to scale in Cobb- Douglas function by:

𝑌 = 𝐾⁺(𝐴𝐿)^,-+, (1.3)

As:

𝑧𝑌 = 𝑧𝐾⁺(𝐴𝐿)^,-+,¹ (1.4)

where z is any positive number.

To obtain the production in the intensive form, we set 𝑧 =_/0^, and obtain:

𝐹 1_/0² , 14 =_/0^, 𝐹(𝐾, 𝐴𝐿), (1.5) By setting 𝑧 =_/0^,, we obtained the production function in its intensive form. The intensive form now shows that the output per worker is a function of the amount of capital per worker:

𝑦 = 𝑓(𝑘), (1.6)

(14)

when 𝑘 =_/0² is capital per effective labour and 𝑦 =_/0⁹.

The next assumptions about the production function in the intensive form are:

𝑓(0) = 0 (1.7)

𝑓′(𝑘) > 0 (1.8)

𝑓^==(>)< 0 (1.9)

when 𝑓′(𝑘) > 0 is the marginal product of capital per worker.

Inada conditions are assumptions about the shape of a production function. When the amount of capital is closer to zero, the highest is its marginal product and the higher the amount of capital we have, the lower the marginal product is. The conditions are described as:

>→Dlim𝑓⁼(𝑘) = ∞ (1.10)

>→Flim𝑓⁼(𝑘) = 0 (1.11)

Due to the diminishing marginal product of capital, this relation holds:

1^GH_G2> 0 ∧^GH_G0 > 04 ∧ 1_G2^G^J^H_J< 0 ∧^G_G0^J^H_J < 04 ^(1.12) when the limits of production are given by:

>→Flim

GH

G2 = 0 ∧ lim

>→D GH

G2 = ∞ (1.13)

>→Flim

𝜕𝑓

𝜕𝐿= 0 ∧ lim

>→D

𝜕𝑓

𝜕𝐿= ∞ (1.14)

As mentioned above, labour growths at a constant rate n and capital at constant rate g, as:

𝐿(𝑡) = 𝑛𝐿(𝑡) (1.15)

𝐴(𝑡) = 𝑔𝐴(𝑡) (1.16)

And the growth in time is described as:

𝐿(𝑡) = 𝐿(0)𝑒^OP (1.17)

(15)

𝐴(𝑡) = 𝐴(0)𝑒^QP (1.18) The amount of effective labour units grows at rate n+g.

The Solow model assumes, that a constant fraction of output, c, is consumed, and a constant fraction, s, is invested (Mankiw et al., 1992). Thus, the change in the stock of capital might be described as:

𝑘(𝑡) = 𝑠𝑦(𝑡) − (𝑛 + 𝑔 + δ)k(t) (1.19)

= 𝑠𝑘(𝑡)⁺ − (𝑛 + 𝑔 + δ)k(t) (1.20) When δ is the rate of depreciation, equation (1.20) converges to a steady-state of value k*:

𝑠𝑘^∗+ = (𝑛 + 𝑔 + δ)k^∗ (1.21) 𝑘^∗ = Y 𝑠

𝑛 + 𝑔 + δZ

,-+,

(1.22)

Equation (1.22) illustrates, that the steady-state value of k* is determined positively by savings and negatively by the rate of population growth and depreciation.

When substituting equation (1.22) into the production function, we obtain the steady-state income per capita:

𝑙𝑛 \𝑌(𝑡)

𝐿(𝑡)] = ln 𝐴(0) + 𝑔𝑡 + 𝛼

1 − 𝛼ln(𝑠) − 𝛼

1 − 𝛼ln (𝑛 + 𝑔 + δ)

(1.23)

Based on the determinants of the growth rate of output in the Solow model (output per worker, the preliminary technology, the technological progress, saving rate, growth rate of the working-age population, depreciation rate and capital share) Solow concludes, that output per worker is positively affected by high saving rate and on the other hand, is negatively determined by a high growth rate of working-age population.

Because in the steady state all variables grow at constant rates, the growth rate of output in steady state is determined by the growth rate of technological progress.

(16)

1.2 Alternative Perspectives on Population Growth

In the Solow model, population growth plays an important role in determining the economy steady-state. With increasing population, the steady-state decreases, because the capital has to be spread among more people. This means, that countries with higher population growth have less capital stock per worker and thus lower income. In this section, I will present two alternative theories concerning the effect of population growth on natural resources and interaction of population with technology, respectively.

The first model comes from Thomas Robert Malthus, who believed, that all living species tend to grow exponentially, while available resources such land used for growing crops are limited and according to Malthus, the human civilization will come to its destruction one day due to not sufficient supply of food. As one might see, Malthus was not very optimistic about the growing population and that lead him to several measures to prevent such growth.

The second model was developed by Michael Kremer, who had a different view on population growth than Solow and Malthus. From his point of view, the growing population was a key driver of economic growth. His claim is based on assumption, that more people raise the chance of discovering new ways of production and thus the positives coming from the new innovation outweigh the negatives from a bigger population.

1.2.1 The Malthusian Economy

The Malthusian Economy was developed by a classical British economist Thomas Robert Malthus, who was born into a wealthy family in 1766 and served as a professor at Cambridge. Malthus in his Essay on the Principle of Population, published in 1798, questioned the positive view of philosophers of the 18th and early 19th centuries which believed, that the development of society will never stop. In his opinion, society due to its population growth will suffer from the lack of food, diseases and high rate of mortality.

Despite the fact, that humans are rational, they are not able to break out of the laws of nature. Due to this incapability, humans have a necessary need to eat and

(17)

reproduce. Malthus was not willing to ease from the assumptions because no such changes were noticed in the past (R. Malthus, 1798).

Malthus stated, that the population is growing in geometric progression, while the food production increases in an arithmetic progression. Due to the increasing gap between the population growth and the growth of food production, one might see, that this will lead to a shortage of food and this so-called “Malthusian catastrophe”. will lead to the population which is back at its “sustainable level.” Another problem arises when it comes to the land used for living. Malthus supposed, that due to the technological progress will increase food production, which will consequently increase the population level. Because the agriculture land will be used for living, the society would get into so-called

“Malthusian Population Trap (R. Malthus, 1798).”

The population decline would have to forms. A first type is a form of Natural Checks. Malthus supposed, that the imbalance between food supply and population growth will be equalized by natural disasters. The second type is of the correction would be a form of preventive control. In the second type, people would voluntarily put off the marriage until they are able to take care of themselves, families would be planned before and in some cases, people would have to keep celibate (R. Malthus, 1798).

The theory stumbles in several areas. The first problem is in the mathematical form of the theory, which was only in the first edition of this Essay. The real data did not prove Malthus claims. When looking to the past, the food production grew faster than arithmetically and the rate of population growth was slower than geometrical. Malthus also did not count with opening up new areas of the United States, Argentina and Australia. Due to that was food more affordable not only in those countries but also in Europe thanks to improvements in transportation. Malthus also did not expect the increase in agricultural techniques and scientific knowledge in the future. As a result, food production was faster than he expected even in developing countries.

We might say, that the problem of The Malthusian Economy lays in its static nature. The overall benefits streaming of a growing population are exceeding its costs.

(18)

1.2.2 The Kremerian Model

American economist Michael Kremer sees the population growth not as a threat, but as a key driver for prosperity. According to Kremer, more people rise the chance of bigger technological progress, because within one hundred thousand people will be more scientist, engineers and so on than within a hundred people (Kremer, 1993).

He argues, that over history, world growth rates have increased together with population growth. This evidence confirms his arguments, that with more people come more ideas and the world is better in looking for new paths towards growth through technological progress (Kremer, 1993).

In his second argument, he says, that the productivity might be depended on the size of the population. In the past was the world separated on several regions for thousands of years. When looking at the numbers, Kremerian assumptions is correct, because countries with higher population experienced higher productivity and thus income per capita (Kremer, 1993).

1.3 Endogenous Growth Theory

Endogenous growth theory attempts to explain technological advance. In the Solow model, technological progress was just assumed as an exogenous process. The endogenous growth theory, the technological advance has endogenous characteristic and thanks to that attempts to clarify the cause of the advance in technology and suggests policies, that might be beneficial. Compare to the other theories, the endogenous growth theory incorporates capital also knowledge, which takes a form of new technologies and education (Aghion & Howitt, 1997).

The specifics in the endogenous growth theory model lay in its production function. In the Solow model, the production function exhibits the property of diminishing returns to capital. In other words, every additional unit of capital raises the product less, than the previous unit. In the endogenous growth theory, every unit of capital raises the output of the same amount(Aghion & Howitt, 1997).

Because the change in the capital stock is a result of the difference between investment and depreciation, the new production function with constant returns to capital

(19)

shows, that as long as sA > δ, the growth of economy’s income never stops (Aghion &

Howitt, 1997).

Thus, the model states, that investments cause a higher rate of production than was assumed in the earlier theories. Because capital investments into physical capital and human capital create positive externalities, investments of the investing company are increasing productivity also to other companies, even though that there exists protection of intellectual property, because it is not possible to keep all the new technologies of production in secret. Thus, from this perspective, we might say, that new knowledge and technologies are characterised by increasing returns. This is in contrast to the Solow model, in which economies once reached its steady-state level and stayed there.

Considering the increasing marginal product of human, the economies do not have to stop in their steady-state anymore (Aghion & Howitt, 1997).

After considering positive externalities from investments to capital, the higher technology process and its externalities might outweigh the decreasing marginal product of physical capital. Thus, a higher rate of saving might lead to a permanent increase in labour productivity and thus income (Aghion & Howitt, 1997).

Due to the connectivity between national saving and the rate of GDP growth, the endogenous growth theory implies ways of designing government policy. Due to the importance of new production technology, the economy can experience unlimited growth and contrast to the Solow model, less developed economies do not always have to experience the catching-up effect. And because according to the endogenous growth theory innovations are the driver of the economic growth, it contains implications for setting up the economic policy (Aghion & Howitt, 1997).

The economic policy should besides its other public investments to education and research include: 1) Creating tax incentives for expenditure on research and development, 2) support of investments to human capital leading to a higher level and quality of education, 3) increase of investments to infrastructure, 4) lowering the level of budget deficits (due to their crowding out of private investment effect), 5) creating tax incentives for an increase of the investment activity, 6) limiting the crating of new and removing the existing regulation of economic processes(Aghion & Howitt, 1997).

(20)

The described endogenous growth theory attempts to answer questions, which the Solow model was not able to answer. The theory sees innovations (creation of new ideas, technologies, etc., that leads to higher productivity) as an essential part of the capital and a determinant of economic growth. In order to maximize the economic prosperity, the government should create a motivational environment, that creates incentives leading to new ideas. At the end of this chapter would be appropriate to say, that economies do not distinguish only in the capital per worker, but also in the institutional framework that is behind the creation of new ideas(Aghion & Howitt, 1997).

1.4 Institutions

This chapter concerns institutions and its importance for economic growth.

Firstly, I will take a look at Douglas North view on institutions and then move forward to empirical studies, that attempt to analyse the role of institutions in the economy.

1.4.1 Douglas North

Douglas North in his paper Institutions written in 1991 defined institutions as socially created constraints that design political, economic and social interaction.

Institutions might be separated on informal institutions, such as sanctions, taboos, customs, traditions and formal institutions, such as constitutions, laws and property rights.

Both, formal and informal institutions, aim to create order and reduce uncertainty in exchange (North, 1991). Together with the constraints of economics, institutions characterize the choice set and thus determine transaction and production costs and subsequently the profitability and feasibility of engaging in economic activity. The economic structure then determines the shape the direction of economic change towards growth, stagnation or decline (North, 1991).

North argues, that it is necessary to constrain human interaction with institutions.

As shown in a game-theoretic context, because people who maximize their wealth function usually find it worthwhile to cooperate with other individuals when the game is repeated. This applies especially for repeated games with a small number of players who knows the complete history of the other players actions. And on the other hand, cooperation becomes more difficult when there is a large number of players with

(21)

incomplete information about the actions of the others and the game is not repeated (North, 1991).

In reality, those, who engage in production seek to maximize profit, has to face increased transaction costs due to unperfect information. It is pretty observable that the economic performance is determined by institutions and North argues, the role of institutions is to make transactions smoother. Therefore, special attention should be paid to ensuring the effectiveness of enforcement. Only when the effectiveness of enforcement is ensured, one can realize the potential gains from trade. Due to this huge importance of institutions, there should be set the right mix of economic and political institutions which will through an adequate economic environment induce economic growth. But how does an economy make its way to the right mix of institutions? North tries to find the answer in the history of those who succeeded, and those, who did not (North, 1991)

The economy had to go through several stages before reaching its current state.

From the simplest form to the most complex we know so far. In the beginning, the first phase represented only a very simple mechanism. The goods were made by using a very simple technique and was exchanged on local markets based on informal constraints. Due to the small extend of trade, the transaction costs were very low at this phase. Later, when the trade gradually increased and the markets broadened, the transaction costs increased as well (North, 1991).

When the long-distance trade developed, two serious transaction costs problems arose. Sending relatives with the goods to negotiate the conditions of the trade and to obtain return goods was the first problem. Contract negotiation and enforcement in alien parts of the world, where there is not easy to achieve agreement and enforce contracts, was the second problem. As the complexity of the trade was increasing, the less reliable and effective the personal ties were (North, 1991).

Only the modern western societies were able to define the choice set and determine the transaction costs and profitability of engaging in economic activity. As a result of increased work specialization across the labour market, the demand for institutions that would keep an eye on the observance of property rights increased. This allowed the capital market to grow efficiently (North, 1991).

(22)

But why more complex forms of organizations do not come into existence in the primitive tribal economies? According to North, no matter what the exchange system is like, economic agents are willing to invest more of their time in knowledge and skills, that will lead to their material status improvement. But in economies, where the institutions are on a very low level, the new knowledge and skills will not result in any institutional evolutions that would make the economies more productive and efficient (North, 1991).

In tribal societies, where the survivals depend on the group, innovations are seen as a threat for its survival. Those economies see the exchange acts as a single activity, which is virtually undifferentiated and cannot lead to the emergence of institutions, that would set the choice set. (North, 1991).

On the other hand, stays the modern western world, which is based on more complex organizations. The arise of long-distance trade in modern Europe between the 11^thand the 16^th century helped to create several innovations lowering transaction costs of engaging in exchange. During those times were developed mechanisms, which ensured effective enforcement. Variety of courts took care of business disagreements and an internal code of conduct developed by guild merchants themselves. In the background of all these changes that occurred within modern Europe, was the state. The state was always trying to reach a balance between its fiscal needs and its credibility in its relationships with merchants (North, 1991).

North provides us with three answers, that attempts to answer the question, why these institutional changes occurred only in Europe and nowhere else in the world. Firstly, in former times, fragmented political units within Europe forced the rulers to seek more revenue to survive. Secondly, the difference in gains acquired from new knowledge and skills in Europe and the rest of the world. Thirdly, path dependence was experienced rather in Europe than in the less developed countries (North, 1991).

1.4.2 Studies of the Effect of Institutions on Economic Growth

A study written by Knack & Keefer (Knack & Keefer, 1995) examines the impact of property rights on economic growth by using indicator provided by country risk evaluators to potential foreign investors. The country risk indicators consist of enforceability of contracts and the probability of being expropriated. The founding of this

(23)

study indicates that property rights have even greater importance in determining frequencies of revolutions, indexes of freedom and political disputes. When such property rights variables are put into growth regressions, the rate of income convergence to the United States raises up. The obtained results are statistically robust.

Bills & Klenow in the paper Does Schooling Cause Growth or the Other Way Round (Bils & Klenow, 1998) founds, that growth and schooling are highly correlated across countries, with each additional year of 1960 enrolment associated with approximately 0.6 % per year faster growth in per capita GDP from 1960 to 1990. They incorporated into their model finite-lived individuals who choose to obtain education.

The model showed, that growth might be influenced by: 1) schooling, 2) technology progress, which increases the return on investment in education. When analysing the influence of those two channels, the results showed, that countries that experienced higher GDP growth are more likely to have modestly flatter cross-sectional experience-earnings profiles. After employing UNESCO attainment data to the model before 1960, schooling did not prove its positive effect on growth, when it did not generate even one half of Barro’s coefficient. Thus, the study results conclude, that growth affects schooling and not the other way round.

The next paper (Hall & Jones, 1999) called Why Do Some Countries Produce So Much More Output per Worker Than Others attempts to answer the questions, why output per worker varies across countries. According to the results, the difference among countries can be only partly explained by differences in psychical capital and educational attainment. They found out, that the different accumulation of capital and productivity and subsequently output per worker is determined by differences in institutions and government policies. The differences in institutions and government policies were determined historically by location and other factors and were treated as endogenous variables.

A paper The Quality of Government (Porta et al., 1999) analysis the determinants of the quality of governments and assesses government performance. For analysing the government's performance were used variables such as the efficiency of the public sector, interventions of government, its size, public good provision and political freedom. The results indicate, that countries, which experienced inferior government performance in

(24)

Another finding of this study is, that governments of bigger size are more likely to perform better over time than those of smaller scale.

Another study by Acemoglu (Acemoglu et al., 2001) estimates the importance of institutions on economic growth by examining differences in mortality rates in the European colonies. Because European colonies adopted various manners, different institutions took place. In the used model played the mortality rate the key role.

Institutions, which were meant for extracting the country sources, were more like to set up in countries with high mortality rates. An extreme case of this might be the Gold Coast or Congo. Institutions in those countries were harming investment and thus economic growth. On the other hand, as in case of the United States, Australia or New Zealand, where the mortality rate was significantly lower than in the previously mentioned countries, institutions which encouraged investment were set up. The model estimates, that the income per capita is largely affected by institutions, which causes the differences in income per capita across European colonies. It would be worth notice, that the results do not imply, that the early institutions prevailed until today and cannot be changed.

Improving institutions is correlated with better economic performance, as one might see on an example of Japan during the Meiji Restoration or South Korea in the 60s.

A study from Clemson University (Vijayaraghavan, 2001) was analysing the relationship between institutions and economic performance in a model, which includes 43 nations within years 1975-90. To estimate the relationship between those two variables, the model consisted of a set of institutional variables such as governance, the security of property rights, political freedom and size of government. This set served as the proxy for the institutional framework. According to this study, the most important institution related to economic growth is the security of property rights. More secure property rights lead to allocative efficiency and a thus higher level of GDP per capita.

When property rights are well defined, the capital is used more efficiently, and the economy is using its potential. The study also indicated that governments of a smaller scale are better.

In the paper Institutional differences as sources of growth differences (Ali, 2003) was examined the importance of institutions on growth and development and evaluated the empirical results on the effect of institutions on growth and investment. This paper shows, that institutions are an important determinant of economic growth. Countries,

(25)

which are characterized by a high level of institutions, such as low level of corruption, effective bureaucracy, protected private property, judicial efficiency and high levels of economic freedom, are experiencing high levels of economic growth. Furthermore, economic freedom is very important for establishing good institutions, which then lead to higher growth and investment. As one might notice, freedom and high-quality institutions are stimulating better economic performance and are closely knitted. Without good institutions, freedom is not sustainabled.

A study by Alcala & Ciccone (Alcala & Ciccone, 2004) was examining the economic effects of international trade on aggregate productivity across countries. The model uses real openness as a measure of trade. Due to the fact, that openness is disturbed by differences across countries in a way of non-tradable goods, the authors argue, that real openness is a better variable for estimating the effect. Because the production of non- tradable between countries is calibrated, the relative price of this type of goods does not affect real openness. The obtained results indicate a statistically and economically significant effect of trade on productivity across countries. Productivity is also affected by the size of the countries when international trade is taken into account. The average labour productivity is affected through the total factor productivity when the economy opens to international trade.

Another study written by Tridico (Tridico, 2007) analysed the determinants of economic growth among emerging economies. There were made two analysis. Firstly, was run a cross-country analysis if a group of emerging and transition economies within period 1999-2005. Results of this analysis should put a light on growth determinants in this type of economies. After that was made a comparative analysis, in which each state obtains a specific classification according to its financial structures and ownership control over firms. When taking that into accounts, the impact of the socio-economic model on growth was analysed. Results showed, that one of the main determinants of economic growth is human capital, but it is necessary but solely not sufficient in transition countries because appropriate institutions are needed as well.

A study from Nawaz (Nawaz, 2015) was examining the growth effects of various institutions on economic growth while using 56 developing and developed countries. The examined institutions included government stability, investment profile, corruption-

(26)

growth. The effect of these institutions was also examined concerning the type of economy. The study shows, that effect of institutions on economic growth is positive, even though it is greater in developed economies relative to developing economies. In low income, economies are more growth-enhancing investment profile compare to developer economies. According to the results, different countries require different sets of institutions for ensuring long-term economic growth.

(27)

2 B ULGARIA ' S E CONOMY T RANSFORMATION

In this section, I will attempt to familiarize the reader with the modern Bulgaria history, during which the Bulgaria economy came through many changes. The first section of this chapter starts by the beginning of reforms towards liberalization after the end of the Stalinist era in 1953 when the communist regime was in charge. During the communist regime, disagreements against the government grew, resulting in massive protests and the subsequent disintegration of the Communist Party in 1990. The second section was devoted to the post-communist era. The provides an overview of the results of elections until 2017 and information about the economic situation of the country.

2.1 The Collapse of the Soviet Regime

Before the collapse of the Soviet Union, the communist countries with centrally planned economy struggled with low effectiveness compare to countries with a market economy. This led to social, political and economic reforms in the Eastern Bloc with Mikhail Gorbachev in leading. The reforms towards liberalization together with the release of tightening after the end of the Stalinist era at 1953 increased expectations for freedom and democracy in the society (Curtis et al., 1993).

Despite the attempts to the system liberalization, the communist regime did not get the support of Bulgaria’s intellectuals, who showed support of several members of the Turkish minority, who decided to hold a hunger strike in 1989. The leader of the People’s Republic of Bulgaria, Todor Zhivkov, to avoid possible conflicts, offered the Turks a departure to capitalist Turkey (Country Profile: Bulgaria, 2006). The number of emigrants for increasing so quickly, that the Turkish government decided to close the borders with Bulgaria. The total number of Turks who decided to leave Bulgaria was approximately three hundred thousand. The outflow of Turks meant a further deterioration of the international position for Bulgaria (Curtis et al., 1993).

In 1989, the support of the opposition was increasing, and the opposition was organising events against the expulsion of the Bulgarian Turks. Besides that, ecological associations were formed, that at the end of the year 1989 organized demonstration on ecological issues, which took place in Sofia. The goal of these demonstrations soon

(28)

expanded to demand a reform of the entire political system. An opposition group led by Petăr Mladenov was formed also in the communist party, which wanted to reform the political organization based on Marxist-Leninist doctrine. They wanted to replace the obsolete doctrine by socio-democratic principles. As a reaction to the demonstrations, communist politicians voted for removing Todor Zhivkov, who officially left his position on the 17. of November 1989 (Country Profile: Bulgaria, 2006). Zhivkov was replaced by Petăr Mladenov. Despite the change in the leading of communist party, the pressures to reform the party came back very soon. In February 1990, the communist party renounced Marxism-Leninism, announced its social-democratic orientation and in June 1990, the first elections were held since 1931. On the 12. July 1991 was officially announced Bulgaria republic with a market economy (Curtis et al., 1993).

2.2 The Post-Communist Era

Bulgaria fully started making its path to democratization after the collapse of the Soviet bloc in 1989, when a member of the Union of Democratic Forces, Thelyu Zhelev, was elected as the first non-Communist prime minister (Country Profile: Bulgaria, 2006).

The democratic political party Union of Democratic Forces, which was in charge in years 1991 and 1992, had to face many troubles when making the reforms such as privatization of agricultural land, properties and industry issuing shares in government enterprises to all citizens. Due to fact, that the Council for Mutual Economic Assistance was dissolved in the mid of 1991 and Bulgaria didn’t participate in any new regional or world trade organizations, the poor state of Bulgaria’s industry came up, when it was not able to face competitors from abroad. This led to unemployment, that exceeded 10 % and which has never been seen before during the collectivist doctrine (Crampton, 2008).

The rise in unemployment was seen as a chance by the Socialists, who were presenting themselves as the protectors of the poor against losing their jobs and other factors, that free-market economy brings. As a reaction to the instability in society, Zhan Videnov, member of the former Communists, got the most votes in parliamentary elections in 1994. Due to the lack of experiences, his weakness soon came up and people from his close surrounding used Videnov for advancing their interests. This pitiful state of the Bulgarian government even deepened the poor Bulgarian economy, which experienced hyperinflation, that was more than 1 000 % in 1997. At the end of the year,

(29)

1997 was Petar Stoyanov, a candidate of the Union Democratic Forces, elected as the new president. The Bulgarian Socialist Party, which was dissolved in 1997, was replaced by the Democratic government (SDS) led by Ivan Kostov. Despite the economic improvement, the BSP failed to solve some of the core issues in the country and was later suspected of corruption. In the presidential elections in 2002, Kostkov failed to retain his position and was replaced by Georgi Parvanov, who claimed to be independent (Crampton, 2008)

People of Bulgaria were unsatisfied with both, BSP and SDS, which graduated in 2001 when Simeon Sakskoburgotski (son of Tsar Boris III of Bulgaria) became Prime Minister of Bulgaria with his political party National Movement Simeon II. During the regime of Sakskoburgotski, GDP per capita improved from 1 757 USD (in 2001) to 3 869 USD (in 2005). Despite the fact that the political situation improved as well, some of the key problems were not solved, such as high unemployment and emigration, corruption, organized crime and poor health care (Financial Times Limited & Dorling Kindersley Publishing, 2011).

A month after the next parliamentary elections in 2005 was created a coalition consisted of the Bulgaria Socialist Party, National Movement for Stability and Progress and Movement for Rights and Freedoms (Spirova, 2006). Despite the different ideology of each of the political parties, their cooperation was enhanced by the common goal, which was the necessary reforms for joining the European Union in 2007.

In the elections of 2009, the political party Citizens for European Development of Bulgaria, which claimed that it ideologically belongs to the centre-right spectre, won 117 seats in the Bulgarian Parliament (Bulgaria Opposition Wins Election, 2009). These times between 2009 and 2013 were times of many reforms. The new prime minister Boyko Borisov was planning on liberalizing the educational system and fiscal discipline, that led to reducing the budget deficit. The hurt Bulgarian economy by the Great Recession experienced another shock by the government spending cuts and increase of taxes. That led to dissatisfaction among the people and protests, which ended by fell of the Borisov government in February 2013 (Brunwasser & Bilefsky, 2013).

The following leading political party led by Plamen Oresharski from May 2013

(30)

the appointment of an oligarch as the director of the State Agency for National Security.

The protests soon extended to the broader political issues in Bulgaria. The protests ended with the resignation of the government in the mid of 2014 (‘Bulgaria’s Prime Minister Resigns with Bank Crisis Unresolved’, 2014).

The upcoming government in 2014 was formed by Citizens for European Development of Bulgaria (the winning party), Reformist Bloc, Patriotic Frond and the Alternative for Bulgarian Revival. The party resigned after the presidential election in 2016 (‘Bulgarian Parliament Approves Resignation of Center-Right Government’, 2016).

In the parliamentary election in 2017 won a conservative Bulgarian political party led by Boyko Borisov, receiving 32,65 % of the popular votes (‘Ex-Premier’s Party Wins Bulgaria Vote to Boost Pro-EU Stance’, 2017).

(31)

3 A CCESSION OF B ULGARIA TO THE E UROPEAN U NION

The first section of this chapter aims to familiarize the reader with Bulgaria's way towards being accepted to the European Union. Starts with the program Phare in 1989 followed by unsuccessful entering in the year 2004 and successful accession in the year 2007 together with Romania. The second section takes a look at the transition period and the special measurements, that had to be made from the side of the newly entering country such as Bulgaria was.

3.1 The Path Towards Accepting

The first official relations between Bulgaria and the European Economic Community were established on the 8th of August 1988. In the following year was created a program Phare, which should have been originally used as a financial help during the renewals of economies of Hungary and Poland. In 1990 was the program offered to other transforming countries of the middle and western Europe for reforming its political systems. Bulgaria in 1990 signed the Convention on Trade and due to that, Bulgaria was able to get money from the Phare program as well (European Commission, 2002).

The first negotiations about entering the European Union began in 2000. The European Council in 2002 decided, that Bulgaria was not ready to start the official negotiations about entering into the EU (European Commission, 2002). The major factors were a high level of corruption, prevailing instability, frequent changes of its government, criminality, etc. Besides these reasons, the European Council criticised not-protecting human rights and lack of environmental protection (Noutcheva & Bechev, 2008).

The preliminary date for Bulgaria entering the European Union was set on 2007.

The accessing document of the European Council from 2002 evaluated Bulgaria as a country with a high level of macroeconomic stability and market economy in a good condition. Highlighted was especially improvement in the fight against corruption and organized crime. On the other hand, the European Council found some difficulties in controlling and veterinary services (Noutcheva & Bechev, 2008).

Bulgaria officially ended its entering talks in December 2004 at the meeting of the

(32)

Luxemburg together with Romania. Signing the Treaty of Accession did not fix the date of the enter, because the European Council had the right to postpone it off one year in case if Bulgaria will not meet some of the required requirements. Bulgarian ratified the Treaty of Accession in May 2005. The final word, that postponing of Bulgaria’s accession is not needed, was said in an official document in September 2006 (European Commission, 2002)

The document released by the European Council was positive about Bulgaria’s preparation on entering the European Union but also mentioned the prevailing problems.

The European Council in a positive way stressed Bulgaria’s fight against organized crime, corruption and reforms in the jurisdiction system. Some issues were found in Bulgaria’s constitution and independency of the jurisdiction power. In terms of government, the economy is necessary to lower the negative trade balance and in terms of nuclear power, the nuclear power station was shut down. Critical issues were found in air traffic control and the food industry (European Commission, 2006b).

3.2 The Transitional Period

The transition measurements concern especially with: (1) Free movement of workers, (2) purchasing of land, (3) road transportation, (4) natural environment, (5) agriculture and (6) food safety (European Commission, 2005).

(1) Despite the fact, that the free movement of persons is one of the 4 fundamental freedoms guaranteed by the single internal market, with Bulgaria were made some exceptions. The older members of the European Union were afraid of the inflow of low wage workers from abroad and as a reaction to that, the European Council said, that in the first two years of Bulgaria's membership in the European Union, the access of the Bulgarian workers to labour markets of the older EU members will depend on the national law and mutual agreements, if there are some (European Commission, 2005).

(2) When Bulgaria entered the EU, many factories did not meet the EU standards and due to that were not able to export meat (26 out of 359) and milk (34 out of 391) to other states of the European Union in the next three years. Producers who were not able to improve their processes were forced to quit their activity (European Commission, 2005).

(33)

(3) The transition period was in terms of natural environment concerned especially with wastewater treatment, storage and recycling waste, pollution control and emissions (European Commission, 2005).

(34)

4 S YNTHETIC C ONTROL M ETHOD

In this section, I will start by introducing the synthetic control method and present its advantages over the traditional comparative case study methods, following its theoretical background. Later, I will present its application in the core and recent empirical studies using the synthetic control method.

4.1 Description of the Method

The synthetic control method (SCM), which was firstly illustrated by Abadie &

Gardeazabal (2003), is well-suited for usage in comparative case studies. This method attempts to estimate the effects of external shocks such as policy interventions appearing on large aggregated units (countries, cities, regions, etc.) and affecting rather smaller units. (Abadie & Gardeazabal, 2003).

Using the SCM, we compare two groups (the treated group vs. the control group).

The treated group represents unit(s) affected by an event or intervention and the control group represents unit(s) non-affected by an event or intervention. SCM creates a synthetic control unit which is then used for estimating the effect on the treated group in the absence of treatment (Abadie et al., 2015).

The method aims to create the synthetic control units, which was defined as a weighted average of available control units, that approximates the most relevant characteristics of the treated unit before the treatment. According to this definition of the synthetic control method, the weights are chosen such that the method best approximates the relevant characteristics of the treated unit during the period before treatment. The synthetic control method then simulates the development of the treated unit in the absence of treatment. The final effect is the result between the synthetic unit and the treated unit (Abadie et al., 2015).

4.1.1 Theoretical Background

Suppose, that we examine data for J + 1 countries. Without loss of generality, we assume that only the first country (J = 1) is exposed to the intervention of interest. The

(35)

remaining J countries, which were not exposed to the intervention of interest, might serve as potential control units (donor pool) (Abadie et al., 2015).

The observed data are for countries i, where i = ∈ {1, …, J + 1} at time t, where t

= ∈ {1, …, T}. The pre-intervention period is denoted by T0 and the post intervention periods are denoted by T0 + 1, T0 + 2, ..., T (Abadie et al., 2015).

YitN (unobserved) represents the potential outcome in the absence of the intervention for country i at time t and YitI (observed) stands for the outcome for the country exposed to the intervention (Abadie et al., 2015).

Constructing the set of control units (countries) has to meet some requirements.

Firstly, units that experienced the intervention of a similar characteristics shouldn’t be included in the control group. Due to this condition, Romania was excluded from the donor pool, because Romania joined the European Union together with Bulgaria.

Secondly, units which adopted a certain shock to the outcome of interest during the examined time shouldn’t be included in the donor pool if this shock would not have affected the treated unit in the absence of treatment. Thirdly and finally, units included in the donor pool should be of similar characteristics to the treated unit to prevent interpolation biases and overfitting (Abadie et al., 2015).

The synthetic control unit is defined as a (J x 1) vector of weights W = {w2, …, wJ+1}’, where 0 ≤ wj ≤ 1 for j = 2, …, J + 1 and

a 𝑤_c = 1

de,

fgh

(4.1)

Given these restrictions, it will prevent us against extrapolation (Abadie et al., 2011).

The effect of intervention of interest on the treated unit is given by the difference between the treated unit and the synthetical unit, denoted by:

𝛼i_,j = 𝑌_,P− a 𝑤_c^∗𝑌_cP

ke,

cgh

(4.2)

(36)

When implementing the synthetic control method, the difference between YitI and YitN needs to be defined. Let X1 represents (k x 1) vector of the country exposed to the shock and X0 represents (k x J) vector represents the values of the same characteristics of the control countries. With respect to the weight constraints, the synthetic control method aims to minimize a distance between X1 and X0W, ||X1 – X0W||, by choosing the optimal vector W*, that is in its full form denoted by:

‖𝑋_,− 𝑋_D𝑊‖_o = p(𝑋_, − 𝑋_D𝑊)′𝑉(𝑋_,− 𝑋_D𝑊), (4.3) where v = ∈ {v1, v2, …, vk} assigns different weight to the predictors depending on their importance. The assigned weights minimize the mean square error of the synthetic control estimator in the preintervention period.

4.1.2 Inferential Techniques

To evaluate the significance of the results obtained by the synthetic control method, we conduct falsification exercises. Abadie & Gardeazabal (2003) initially run two types of placebo tests. The placebo tests assume, that the confidence estimated by synthetic control would be significantly changed if we obtained effects of comparable or even greater scope for units, where the intervention did not occur (Abadie et al., 2015).

The first method applies the synthetic control unit to dates in the treated unit, when the intervention did not occur. These tests can be applied for units, where is a data period large enough when no structural intervention to the treated unit appeared. If there is a large effect also before the event of interest, the confidence that the effect estimated for the event of interest would significantly decrease (Abadie et al., 2011).

The second method constructs synthetic control method to every control unit of the donor pool. The credibility of results would greatly diminish if an effect of similar or larger scope would be estimated for units, which were not exposed to the intervention (Abadie et al., 2015).

When the synthetic control method is iteratively applied to each state in the donor pool and we obtained a distribution of placebo effect, we compare the gap (RMSPE) for the treated unit to the distribution of the placebo gaps before and after the intervention of

(37)

interest. Then is created a pseudo p-value based on permutation tests, which is defined as a ratio of RMSPE before and after the shock.

RMSPE for the prior-intervention period is defined as follows:

𝑅𝑀𝑆𝑃𝐸 = w(1 𝑇_Da(

y_z

Pg,

𝑌_,P− a 𝑤_c^∗𝑌_cP)^h)

ke,

cgh

(4.4)

and the ratio for the post-intervention period:

𝑅𝑀𝑆𝑃𝐸 = w( 1

𝑇 − 𝑇_Da(

yz

Pg,

𝑌_,P− a 𝑤_c^∗𝑌_cP)^h)

ke,

cgh

(4.5)

Thus, the ratio of the preintervention and postintervention period is described by:

𝑅𝑀𝑆𝑃𝐸_{|P}~ = 𝑅𝑀𝑆𝑃𝐸_•~€P

𝑅𝑀𝑆𝑃𝐸_••‚ (4.6)

The model credibility can be deducted by the p-value of the treated unit compare to the control units. The higher the value of the p value, the more robust the model is.

4.2 Applications of the Synthetic Control Method

This section aims to familiarize the reader with the application of the synthetic control method. As mentioned above, the synthetic control method was introduced by Abadie & Gardeazabal (2003), which was later used in several empirical studies.

Abadie & Gardeazabal (Abadie & Gardeazabal, 2003) in The economic costs of conflict: a case study of the Basque country, where they investigated the impact of the terrorist conflict in the Basque Country. After comparing the Basque country (treated unit) with its synthetic control unit, authors find that the average gap between Basque per capita GDP and the per capita GDP of a synthetic Basque is about 10 % over the period of 20 years.

The next paper by Abadie et al. (Abadie et al., 2010) using the SCM is Estimating the Effect of California’s Tobacco Control Program, which estimated that the number of

(38)

of Proposition 99. Authors in this study extended the synthetic control method by using new inferential methods that involve uncertainty about the validity of the control unit.

Using the synthetic control method, Lee (Lee, 2011) examines the efficiency of inflation targeting in emerging economies. He finds that inflation targeting was effective in reducing the inflation rate in Colombia, the Czech Republic, Hungary and Poland, but no significant effect was found in countries which adopted later such policy.

Billmeier & Nannicini (Billmeier & Nannicini, 2013) used the synthetic control method in their empirical study that was assessing economic liberalization consequences.

The results showed that there was a positive economic effect of liberalization in most economies but notice, that there was found no significant impact of the more recent liberalization mainly in Africa.

Acemoglu et al. (Acemoglu & Robinson, 2013) present the connectivity of financial firms and politics. In November 2008 was Timothy Geithner announced as the nominee for Treasury Secretary in the United States. The cumulative return for financial firms was approximately 6% after the first full day of trading and about 12% after ten days of trading.

In the next paper by Campos et al. (Campos et al., 2014) were analysed the economic benefits from membership in the European Union. They estimated the effect of membership for countries from the 1980s and 2004 enlargements. The results indicate, that joining the EU led to approximately 12% higher GDP per capita in average.

Abadie et al. (Abadie et al., 2015) in an empirical study Comparative Politics and the Synthetic Control Method used the synthetic control method as a bridge between qualitative and quantitative approaches in empirical studies. The authors estimated the effect of German reunification in 1990 on West Germany. The results show that over the entire 1990-2003 period was the real GDP per capita reduced by about 1,600 USD per year on average.

Žúdel & Melioris (Žúdel & Melioris, 2016) examine the economic effect of Slovakia’s euro adoption in 2009. They find that entering the EMU and adopting euro had a positive effect on its economy, which led to about 10% higher real GDP per capita by 2011.