Corruption and the financial performance of firms in
transition economies.
Bachelor thesis
Author: Maaike Kistemaker Student number: 10398910
Study: Economics and Business (specialization: Business Studies) Date of submission: 10-‐07-‐2015
First supervisor: E. Dirksen MSc.
Second supervisor: R.H. Kleinknecht MSc.
Statement of originality
This document is written by Maaike Kistemaker, who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
The purpose of this research is to identify the relationship between corruption and firm performance in transition countries, both directly and indirectly via the variable competition. The corruption perception index is being used to measure corruption, while the other variables are measured using results from the BEEPS survey. The outcomes include an analysis on the differences between industries. The results of the regression analysis show that for firms in the manufacturing industry, corruption influences firm performance in a negative way, while in the other industries, this relationship is contradictory. The indirect relationship via competition cannot be proved. The findings may be useful for firms that want to expand or start up a business in one of the transition countries.
Table of contents
1. Introduction ... 5
2. Literature review ... 7
2.1. Definition of corruption ... 7
2.2. Earlier studies on corruption and firm performance ... 8
2.3. Studies on competition ... 11
2.4. Hypotheses ... 13
3. Methodology ... 15
3.1. Measures of corruption ... 15
3.2. Measures of firm performance ... 18
3.3. Other measures ... 18
3.4. Limitations ... 20
4. Results ... 21
4.1 Descriptive statistics ... 22
4.2. Results and analysis ... 23
4.3. Comparison between industries ... 26
5. Discussion ... 28
5.1. Interpretation of the results ... 28
5.2. Contributions to theory and practice ... 29
5.3. Limitations and suggestions for further research ... 30
5. Conclusion ... 32
Bibliography ... 33
Appendices ... 36
Diagram 1: Visualisation of the hypotheses ... 14
Table 1a: Descriptives competition and sales growth ... 22
Table 1b: Descriptives industries ... 22
Table 1c: Descriptives CPI ... 22
Table 2: Control variables ... 23
Table 3: Summary of the regression results ... 25
1. Introduction
In today’s fast growing business environment, the challenge for firms is to broaden their horizon and expand to new markets. When moving to a new country or starting a new business somewhere, companies are looking at the attractiveness of a market. Attractiveness can be influenced by a factor such as riskiness. The degree of development of a country’s economy produces different degrees of risks. Examples of risks are an instable economy, political issues or corruption. These factors can make it harder to estimate the business environment. Corruption is being elaborated upon in this research. Research shows that the scale of corruption is estimated at more than US $1 trillion dollars per year. However, the scale of corruption varies a lot between countries, which makes it an interesting phenomenon to study (Website World Bank). Research on the topic of corruption is already widely available. There has been done research on the relationship between corruption and the willingness to invest in a country. Mauro (1995) has found a negative relationship between these factors, as well as Brouthers et al. (2008). Corruption is also negatively related to economic growth (Mauro, 1995). Researchers have spent time on the issue of corruption, but most studies tend to focus on country-‐level effects only (Asiedu and Freeman, 2009). However, in a market of products and services, the main players are firms and consumers. Depending on the market, the government can also play a big role. The focus on firms in this context has not been investigated extensively yet. When firms will be affected by corruption, consumers will indirectly be involved, because they are dependent on firms for their demands. Corruption may have a distinctive effect on different types of firms.
Corporations have to deal with corruption and some will be able to do this better then others. Firms operate to ensure continuity; therefore they strive to optimize their financial performance. The performance of these firms might be influenced by corruption, but there is not yet much investigation done on this part of the topic. When companies would know what influence corruption has on their performance, they could take this into consideration when deciding where to expand or start up their business (Rodriguez et al., 2005). One of the few papers that concentrate on the effect of corruption on firm performance, is the paper by Gaviria (2002). This paper concentrates on the case of Latin America and concludes that firm performance can be affected by corruption, if managers see corruption as an obstacle. However, this research focuses on
the case of Latin America and is presumably not generalizable for other parts of the world.
The situation in the so-‐called transition economies could be different. These countries are moving from a centrally planned economy to a market economy (Feige, 1994). Shleifer and Vishny (1993) believe that a political modernization, which includes the transition from an autocracy to a democracy, is accompanied by growing corruption. The main reason for this is that the institutions are underdeveloped. A state of transition is impacting the environment for firms. A free market is introduced and therefore the dependence on the government is reduced. Firms that existed in the time of the planned economy are less used to competition, which could make it harder for them to survive in a free market. The absence of institutions as well as the redistribution of wealth from the government to the public sector could foster corruption (Website World Bank 4). The transition countries are primarily located in Eastern Europe and Central Asia. In the past decades, internationally producing and trading has become easier, especially for countries within the EU. Therefore, in the last years, it is becoming more common for European companies to move to Eastern Europe, mainly because producing products in this area is cheaper than in the rest of Europe (Website European Commission). The countries in Central Asia are also cheap for producers from the Western world. However, if corruption in these countries appears to have a strong negative influence on firm performance, it can be a stimulant for these companies to relocate their business. Therefore, transition countries in this context form an interesting population to study. This paper contributes to the existing literature by investigating the link between the financial performance of a firm and the corruption in the country where it operates. The research question therefore is: How does corruption impact the financial performance of firms in transition economies?
The next section will contain a review of literature. In this review, the most relevant literature will be discussed and there will be elaborated on the most important concepts. After that, the hypotheses are being developed. The subsequent chapter will describe the research design, endeavoring to provide a comprehensive overview of how the research question is going to be answered. The hypotheses will be tested by the use of a regression analysis and the thesis will finish with a section presenting the results and the implications of this research. Finally, an answer to the research question will be provided in the conclusion.
2. Literature review
In order to introduce the topic of corruption, this chapter reviews the existing relevant literature. There are different publications on corruption, all focusing on different aspects or different relationships; these will provide a solid background as the start of this thesis. This chapter starts with a description of the concept corruption, followed by a discussion of the literature on the topic of corruption. After that, the concept of competition is being discussed. An indirect effect via competition on the relationship between corruption and firm performance is tested.
2.1. Definition of corruption
Literature provides different definitions of what corruption actually is. Many papers cite the definition of corruption in the way it was formulated by the World Bank. The World Bank defines corruption as ‘the abuse of public power for private benefit’ (The World Bank, 1997). Shleifer and Vishny (1993), define corruption as ‘the sale of government property by officials for personal gain’ (p.599). Here, the focus is also on the government and the public officials. Transparency International defines corruption as the abuse of entrusted power for private gain (Website Transparency International). According to this source, corruption hurts everyone who depends on the integrity of people in a position of authority. Although I agree with the definition used by Transparency International, for this research, the definition by the World Bank is going to be used. This definition is used by many other researchers. On top of that, there are more exact measures provided of corruption in the public sector.
A widespread form of corruption is the use of “facilitating” payments (Argandona, 2005). An important question is, where the boundary lies between corruption and “being polite”. In the case of bribery, it is difficult to distinguish if firms are forced to pay bribes, or that they do payments because it is considered to be a habit to give presents or gifts out of politeness. Tanzi (1998) makes a clear distinction between these two: his reasoning is that a bribe suggests reciprocity, while a gift is non-‐binding. However, this author admits that it is sometimes difficult to see whether something is a gift or a bribe. Circumstances like the size of the gift and the prevailing culture are important factors (Tanzi, 1998). Argandona (2005) elaborates on this, by stating that gifts are explicitly an expression of appreciation or good will, with the purpose to create a friendly
atmosphere. Also, the receiver should not feel forced to return anything to the giver. This confirms that, although there is a distinction between gifts and bribes in theory, it is not so easy to distinguish them in practice. It can be difficult to find an exact measure of corruption, because not all people will define corruption the same.
2.2. Earlier studies on corruption and firm performance
There has been done a lot of research on the different sides of corruption, on the origin as well as on the consequences. Some of the researchers have already related it to firm performance. Romney et al. (2012) dive deeper into the question who perpetrates fraud and why. In their book, they mention the so-‐called ‘fraud triangle’, which was created by researchers that observed psychological and demographic differences between white-‐ collar criminals and the public. This model claims that there are three conditions present when fraud occurs: pressure, opportunity and rationalization. An example of pressure could be an extreme dependence on lifestyle, for instance when a person uses drugs or gambles a lot and is in “need” of money. Opportunity exists when someone is able to commit and conceal the fraud, as well as convert the theft to personal gain. Rationalization allows the perpetrator to justify its illegal actions. People can think that they ‘owe’ something or that they don’t have to behave according to rules and regulations. This model demonstrates that corruption is prevalent because of different factors and that it is not so easy to fight it. The mentality seems very important, but also the rules and regulations, which can create an opportunity for potential fraudsters. There are also other authors that focus on the origin of corruption. Johnson et al. (1998) claim that a higher tax rate causes an increase in the size of the unofficial economy. However, countries that are considered to have a low amount of corruption, for instance Denmark and Sweden, have a really high tax rate. Therefore, this conclusion does not seem to be true for all countries. Friedman et al. (2000) state that it’s not the tax rate that causes a large amount of unofficial activity. If the application of rules and regulations causes a high amount of bureaucracy, unofficial activity tends to increase. This bureaucracy drives firms underground. The government will see its tax revenues decline, and in this way governments become smaller and weaker and the underground sector bigger and stronger (Friedman et al., 2000).
The reason why there is more corruption in some countries than in others can be dependent on the institutions (Tanzi, 1998). Jain (2001) sums up the different
considerations that are taken before taking part in corrupt activities. First of all, the probability of being caught is an important factor. When there is no policy to search actively for corrupt practices, it will become easier and more attractive to engage in such practices. The second consideration is the independence of judiciary from politicians. In countries with a high amount of corruption, politicians have relatively much control of the judicial system. Thirdly, the amount of law-‐enforcing officials that are corrupt will determine the effectiveness of the system and the amount of corruption. Lastly, equal access to the law for everyone is likely to be good against corruption.
When countries experience a high level of corruption, the government can decide to set up a policy to fight corruption. Improving one of the four things mentioned above by Jain (2001) could help. Imposing high penalties could be an aid to fight corruption, because a penalty plays an important role in deciding whether or not to commit a crime (Tanzi, 1998). However, Shleifer and Vishny (1993) believe that penalties change the level of bribes, but don’t address the essence of the problem. Officials will change the bribes according to the penalty system that will be used. When penalties are increasing with the height of a bribe, officials will tend to reduce the bribe, but increase the amount of bribes. This process will be reversed when penalties will increase with the number of bribes. Moreover, in a lot of cases, there is a large gap between the penalties stated in the law and the penalties that are enforced in practice (Tanzi, 1998). Penalties are a good way to fight corruption in theory, but in reality this is more complicated. Research on the country-‐level effects of corruption is performed by Mauro (1995). A negative relationship between corruption and investment as well as GDP growth is proved. The paper by Lambsdorff (2005) confirms this, showing that corruption makes a country less attractive for foreign as well as domestic investors. The cost structure in a country has a high chance of getting distorted due to corrupt payments. When a high amount of corruption is present, prices tend to increase, because there needs to be accounted for the charges or rents that are being extracted later. Therefore, according to Kwok and Tadesse (2006), corruption distorts the efficient allocation of resources. Mo (2001) investigates in what particular way corruption affects the economic growth in a country, and finds that 53% of this is caused by political instability. This means that corruption is for a large part being caused by the institutions prevalent in the country, for instance a weak juridical system, or other institutional inefficiencies. Referring back to the fraud triangle (Romney et al., 2012), this means that
“opportunity” is seen here as the main reason for corruption. Mauro (1995) advocates that both corruption and bureaucratic efficiency are having a negative impact on GDP, which could suggest that these two are strengthening each other. Overall, on country level, corruption seems to have a negative influence on economic factors. Gaviria (2002) has elaborated further on the efficiency within a firm and concludes that the time being wasted on bureaucracy will be lower in firms that pay bribes. This means that firms paying bribes allocate their time more efficiently than firms that don’t engage in these corrupt practices. An example of this is the positive relationship between corruption and effective bureaucratic interference. Bureaucratic interference can be described as the part of senior management’s time that is being spent on dealing with government officials, for instance to discuss the application and interpretation of regulations and laws (Gaviria, 2002). A comparison between paying high and low bribes is also being made: when higher bribes are paid, bureaucratic interference becomes even more effective and the time consumed by communicating with public officials lowers, which has a positive effect on the budget (Gaviria, 2002). This can be a positive side of corruption. Kaufmann and Wei (2000) also dive deeper into the topic of efficiency. They investigate the ‘efficient grease’ hypothesis: this suggests a negative correlation between corruption in the form of bribes, and the time wasted by officials. However, they don’t find strong evidence for this hypothesis. In fact, what they find is actually a positive correlation across the firms in their sample. Rose-‐ Ackerman (1997) takes an in-‐between view on this: according to this author, tolerating corruption in small portions may ‘smoothen the rough spots’ in the system (p.33). However, small amounts of corruption tend to provoke pressure and opportunity to increase the amount of bribery. Here it is shown that researchers recognize another part of the fraud triangle: pressure. All in all, there is no conformity on this part of the topic. Kwok and Tadesse (2006) decide to look at the issue from another perspective. According to them, companies should not try to adjust to the country where they want to settle. They claim that the different firms, which are active in a country, are shaping the environment. This means that the prevailing business climate and the behaviour of the firms is the main reason for the amount of corruption existing in a country, not the government or other institutions. Rodriguez et al. (2005) don’t agree with the conclusions of the previous study. They have the opinion that firms should understand and appreciate the essential characteristics of corruption in a country, before they can
adapt and perform well in the new environment. They find this especially important for multinational companies that are willing to expand to new territories, because these companies mostly don’t have any previous experience with the culture and habitats of the new country. According to them, the performance of these firms depends therefore on the willingness to adapt and learn.
An example of authors that already dealt with the relationship between corruption and firm performance are Athanasouli et al. (2012). Their results show a heterogeneous firm engagement in corruption, which means that firms of different sizes are affected differently by corruption, whereby smaller firms are more likely to be engaged in corruption. Moreover, they suggest that firms choose their own optimal level of corruption, which allows them to maximize their profits. This implies that some levels of corruption have a positive impact on a firm’s profit. Brouthers et al. (2008) do not agree with this, by claiming that corruption will always cause lower profits. According to them, corruption will lead to higher consumer prices due to additional costs. This is in line with the theory Kwok and Tadesse (2006) provided about efficient allocation of resources before. New companies will mostly use a ‘follow-‐the-‐leader’ strategy, and set similar prices (Knickerbocker, 1973). In this way, consumers pay for the corrupt practices, their demand lowers, this leads to lower market penetration and this will eventually lower profits in the particular sector (Brouthers et al., 2008). Gaviria (2002) also discusses the negative relationship between corruption and firm performance and concludes that for the countries in Latin America, the effect of corruption on sales growth is negative, but limited. There is no conformity on this part of the topic, but most papers point at a negative correlation between corruption and firm performance.
2.3. Studies on competition
Another concept that has been related to corruption before is competition. Competition can roughly be divided in two types: price-‐ and quality competition. Price competition exists when consumers make their purchasing decision on the basis of the price and companies therefore compete by trying to offer their product for the lowest price. This is mostly the case when products have homogenous attributes. Examples of these products are commodity goods and raw materials such as oil or wheat (Baldwin & Ito, 2008). A typical example in everyday-‐life is the decision where to buy your groceries, when supermarkets sell the exact same brands.
A different strategy for companies is to compete on quality. Quality can exist in different forms. Companies can decide to excel at the service they provide, for instance a 24/7 helpdesk or an extended warranty policy. Customers may decide to pay more, in exchange for these types of services. In the example of the supermarket, prices may not be the only factor influencing the purchasing decision. A dirty store or long lines at the cash desk may cause the consumer to pay more for the same product in a convenient store. Firms can also make the products more robust and from a better quality. This can persuade customers to pay more, because they assume that the product can better fulfill their needs or that the product will have a longer life than an alternative having lower quality. Schott (2008) illustrates this principle by showing that consumers are willing to pay more for a “made in OECD” product, than for a “made in China” product. In short, if consumers care enough about the quality of a product, the higher priced goods will be more competitive. This is opposite with homogeneous goods, where the price is almost the only way to differentiate from competitors. Here, consumers will prefer the goods with the lowest price (Baldwin & Ito, 2008). In this research, the type of competition used will be price competition. The main reason for this is that price competition is easier to measure, because prices can be quantified. Quality, on the other hand, is really difficult to express in quantities. Moreover, the literature on this topic is mainly about price competition.
Shleifer and Vishny (1993) estimate the effect of corruption on different levels of competition. Their conclusion was that with an independent monopoly, most corruption and inefficiency occurred. An oligopoly or joint monopoly already made the situation better, but competition in a market was the most effective to fight corruption. This can be explained by the fact that in a market with more competition, prices will end up lower and in this way a larger amount of products will be sold and the production will be most efficient (p. 608). This can also be applied to the case of government goods. When competition is allowed in the provision of goods for the government, bribes and corruption will be driven down (p. 610).
Bliss and Tella (1997) study the relationship between corruption by officials and competition. They create a model in which corrupt officials can maximize the bribes they extract from firms, by choosing such a height of the bribe, that some firms will leave the market, because they can’t generate enough profits anymore. This will reduce
competition. However, according to them, when there will be growing competition in a market, this will not lower corruption, nor change the demands of the officials.
Another author that mentions the relationship between competition and corruption is Mauro (1998). He states that large bribes are more often available in markets with a low degree of competition, which is in line with the research of Shleifer and Vishny (1993). However, the concept of competition is not yet brought together with sales growth of firms in different countries. The link between corruption and sales growth can be influenced by the degree of competition in a market. Therefore, on top of studying the relationship between corruption and firm performance, competition is a third variable that can be brought together with the other two. Whereas these concepts are not yet investigated from this perspective, this research contributes to the existing literature.
2.4. Hypotheses
From this theoretical framework, four hypotheses are being derived. As mentioned in paragraph 2.1, the boundary between informal gifts and bribery is not clear. Bribery is a part of the concept “corruption”, while gifts are used to create a friendly atmosphere (Argandona, 2005). First of all, to check if the measurement of corruption is related to the perceived amount of bribery and facilitating payments, the following hypothesis is tested:
H1: The amount of gifts and informal payments within a company is positively related to corruption.
Secondly, the authors who performed studies on the relationship between competition and corruption agree for the largest part on the effect of competition on corruption: a market with a low degree of competition has a higher chance of facilitating corrupt practices. Therefore, a negative relationship is expected between these variables. The effect of competition is included in the hypotheses, because there will be investigated if there is an indirect effect via competition on the relationship between the independent variable corruption, and the dependent variable firm performance. Therefore, the second hypothesis is:
Thirdly, competition should also be tested for a relationship with the dependent variable, firm performance. Generally, economic theory teaches that in a homogeneous market with a lot of competition, prices will decrease till they reach the level of production costs, which will set the profits of the producers at a minimum level (Pindyck & Rubinfeld, 2005). However, firm performance will not measured as profit, but as sales growth, which will be elaborated upon in the methodology section. Lower prices will increase the size of the potential market, because buyers of related substitutable products might move to this cheaper market. This shows that high price competition can attract new customers, and in this way this can foster sales growth of the companies active in that market (Pindyck & Rubinfield, 2005). Therefore, the third hypothesis expects the following relationship:
H3: There is a positive correlation between price competition and firm performance.
The relationship between corruption and firm performance has been investigated before, as shown in the literature review (section 2.2). Most authors point in the direction of fewer sales in an environment which is highly corrupt, because corrupt practices extract some part of the revenue. Therefore hypothesis four is:
H4: There is a negative correlation between corruption and firm performance.
The hypothesized relationships between hypotheses 2, 3 and 4 are visualized in diagram 1.
Diagram 1: Visualisation of the hypotheses
3. Methodology
To be able to answer the research question and to accept or reject the hypotheses, appropriate data should be collected. For this, a quantitative research method will be used. The data used to approach this research will be secondary and will be collected from one of the databases described in the subsequent section. All these databases can be accessed easily and contain all information necessary to answer the research question. Secondary data refers to data that is originally collected and analyzed by someone else, and meant to serve a different purpose. An advantage of using secondary data is that it saves time, money and effort, as the data are already collected (Saunders et al., 2012). A disadvantage of using secondary data is that the user has no control over the quality. However, the quality of the secondary data used in this research can be examined by looking at the methodology of the researchers who collected the data in the first place. With the collected data, a regression analysis is going to be performed to test the hypotheses. From this analysis, conclusions can be drawn and the results will be elaborated further upon in the discussion (Saunders et al., 2012).
3.1. Measures of corruption
Corruption is defined in the previous chapter, but it is also important that there is a reliable measure of corruption available across different countries. Corruption is a very big ‘industry’. Only until recently, there were not many estimates available on corruption (Website World Bank). Today, more and more effort is being made to collect information about the size of corrupt transactions. Most papers about corruption base their research on an index. The question is if these indices are really accurate, because the data in these indices do not measure the factual amount of corruption, but only the perception towards the corruption-‐level (Knack, 2006). This means that these perceptions are mostly not based on any direct knowledge. This can possibly cause biased results (Treisman, 2007). Even though most indices use distinctive methodologies, all of them measure the same target variable. The findings of different institutions are showing high correlations (Knack, 2006). Two measures of corruption will be discussed next. By weighing the advantages and disadvantages, a decision will be made about which index to use.
Transparency International: corruption perception index (CPI)
The CPI is widely known and used. It is being published every year by Transparency International. CPI ranks countries based on the level of corruption in the public sector. The rating reaches from 0 (highly corrupt) to 100 (no corruption). No country has a perfect score; the highest in 2014 was Denmark, with a score of 92 (Website Transparency). The information used to compile this index is retrieved from 12 different institutions, including the World Bank and the African Development Bank (Website Transparency). CPI tries to reduce the measurement error in its index by averaging the outcomes of different sources (Treisman, 2007). These 12 institutions measure corruption in different countries, but a lot of the institutions show an overlap in their results, which make them more robust (Website Transparency). The measures include the bribery by public officials and the effectiveness with which corruption is exterminated in the public sector. The scores are based on expert’s assessment.
A disadvantage of the CPI is that the methodology for the construction of the index has changed over the years (Treisman, 2007). This means that the fluctuations of a country’s score might be partly caused by the different methodology used, and not by an actual difference in the level of corruption. When there would have been performed a comparison of CPI scores over different periods, this would have caused problems for the reliability. CPI data from only one period is used for this research, which solves the problem of reliability.
World bank: control of corruption
Another measure of corruption is published by Kaufmann and his team at the World Bank. The Worldwide Governance Indicators project reports for six dimensions of governance. One of these dimensions is control of corruption. The ranking is a percentile rank among all countries, ranging from 0 to 100, which shows how a country performs in comparison to the other countries. A variety of institutes have provided the information to compose this index and the index is updated every year, which doesn’t differ from CPI (Website World Bank 2).
The CPI and control of corruption measures share some characteristics. They both gather their information from several sources, and by averaging these different sources they reduce the measurement error (Treisman, 2007). Furthermore, they both show margins of error when presenting a country score, to remain as transparent as
possible and to show that some measures contain more uncertainty than others. Both these indices have changed the sources they used over the years (Treisman, 2007). This means that fluctuations in the level of corruption can be caused by the use of different institutions providing information. However, the choice to do it this way can also be viewed from another perspective. Both Transparency International and the World Bank keep trying to look for the best possible information to compose their index, and sometimes better or newer sources have more complete information.
A difference between the measure of Transparency International (TI) and the World Bank (WB) is that TI only takes into account countries that can provide information for every part of the rating, while WB already includes a country when it can provide information on one single part of the rating (Treisman, 2007). Looking at this difference, the measure by TI seems more solid and less biased. In the measure by WB, a country can provide information on only one aspect in the total rating, but it is uncertain if this information is representative.
A disadvantage of using one of these indices is that they are based only on perceptions, not on factual information. The question remains if these figures are really exact, but on the other hand, it is really difficult, maybe even impossible, to collect accurate information about corrupt transactions. Lambsdorff (2005) states that the data available on the topic of corruption are mostly based on subjective perspectives, but are considered to be good indicators of the actual levels of corruption, which makes an index an eligible measure for quantifying corruption. On top of that, the perception of corruption can have substantial effects, even when it is not matched by reality. When people perceive corruption in a country to be high, even though it might not be the actual situation, foreign investment is reduced (Treisman, 2007). The study by Gaviria (2002) looks at the perception of corruption managers have. When managers have the perception that corruption is an obstacle to doing business, the performance of this firm is on average worse. These perceptions of the managers might not be in line with the real amount of corruption, but this study confirms that perceptions do matter.
To measure corruption in this research, the most recent results of the corruption perception index will be used (2014). The indices have similarities and their measures correlate for the largest part, but the CPI measure appears to be more reliable, due to the higher requirements set by Transparency International for a country to participate.
3.2. Measures of firm performance
The measurement of firm performance can be done in different ways. Profits can be compared, market share can be measured or a growth of sales measure can be used. For this thesis, the last measure seems most suitable, because growth of sales is a relative measure, which makes it easy to see if a firm made any progress in the previous year. On top of that, this data is available in the used database.
To measure the firm performance of these transition countries, data from the Business Environment and Enterprise Performance Survey (BEEPS) will be used. BEEPS is a survey for firms in Eastern Europe and Central Asia, which contains a representative sample of companies in a country. The purpose of the survey is to collect information about the business environment in a country, including topics such as competition, crime and corruption. The respondents are mostly top managers or business owners (Website BEEPS). In 2012, it was the fifth time BEEPS gathered data from all these countries. The 2012-‐2014 round includes information from 15.883 companies in 30 countries, all in Eastern Europe and Central Asia (BEEPS website). The countries included in the research sample will be these transition countries. These countries are: Albania, Armenia, Azerbaijan, Belarus, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Estonia, FYR Macedonia, Georgia, Hungary, Kazakhstan, Kosovo, Kyrgyz Republic, Latvia, Lithuania, Moldova, Mongolia, Montenegro, Poland, Romania, Russia, Serbia, Slovak Republic, Slovenia, Tajikistan, Turkey, Ukraine and Uzbekistan. The companies included in the survey are all formally registered and from divergent sectors, forming a representative sample of the firms active in that country (BEEPS website). To measure sales growth, the difference between the outcomes of questions D2 and N3 will be used (Appendix 1A). These questions ask about the sales in the year of the survey, as well as the sales three years before that. The difference between these answers represents the sales growth of the firm.
3.3. Other measures
For the different hypotheses, different parts of the BEEPS survey are going to be used. To test hypothesis 1, the outcomes of questions G4, J5, J12, and J15 will be tested for a significant correlation with the levels of corruption in the countries (Appendix 1B). These questions are related to the expectation of an informal gift or payment in different situations. For the measurement of corruption, the CPI will be used for the participating
countries in BEEPS. To measure competition, question E2 of the survey is relevant, because this question relates to the degree of competition which managers’ experience (Appendix 1C). Also, there is being asked if this competition is considered too intense. More data to measure competition would have made the measure more reliable. However, in the BEEPS survey, there was no more information provided about the competition the respondents experience. It is a limitation that competition is not measured on the basis of more data. However, other sources providing valuable information could not easily be accessed. The questions in the BEEPS survey are asking for very specific information, which makes the data still useful.
Control variables will be used when performing the regression analysis. The control variables in the regression are firm specific. The firm specific control variables will be the age of the firm (Appendix 1D), size of the firm (Appendix 1E) and whether the firm exports or not (Appendix 1F). All these characteristics have been provided by the respondents in the survey so they can be used in the regression model as control variables. There is also one country-‐control variable implemented in the regression: GDP per capita. The GDP per capita measure is composed by the World Bank and is measured over the period 2010-‐2014. The average calculated over four years will make the GDP per capita measure more reliable and less sensitive to disruptions (Website World Bank 3).
When analyzing the results, the differences of the effects between distinctive sectors will be considered. In question A4 of the questionnaire, the respondents are being asked what sector their firm is active in (Appendix 1G). There is a list of options provided, but for this research, it is decided that the sectors are going to be divided in three parts. The first one is manufacturing, the second one is going to be retail and the last part contains firms active in other services. In the results, the different sectors will be considered, to verify if the there are any significant differences between firms active in the retail-‐, manufacturing-‐, or other services sector. This can be useful for firms wanting to start a business in one of the countries in the sample. By creating more specific outcomes, the results can be of more use, because they are explicitly focused on one type of company.
3.4.Limitations
A limitation of using BEEPS is that on some questions, the possible answer can be “don’t know”. These results must be excluded from the research, as they cannot contribute in rejecting or confirming the hypotheses. Respondents that answered “don’t know” on one of the relevant questions for the research might do so because of a confidentiality policy, because of lacking administrations or because they don’t want to share all information. Exclusion of these firms might cause an underestimation of the impact of corruption. Also, firms younger than three years are being excluded from the research, because the sales growth is being measured over a period of three years. Therefore, this survey is unable to measure sales growth for young firms. Therefore, the average age of firms in the sample will increase, which might have an effect on the results. It is believed that firms, which are longer active in the industry, are more involved in corruption, so this limitation is likely to increase the estimated influence of corruption on sales growth. Moreover, the results of this research will not be generalizable, as the sample of countries is not representative for other parts of the world. This should be taken into consideration when interpreting the results.
4. Results
Paragraph 4.1 presents the descriptive statistics. After that, the results of the regressions will be provided in paragraph 4.2. A few regressions have been performed to test the hypotheses. (Table 3). This will be followed by an analysis of the results, including a specification per industry.
Some of the variables are recoded into different variables, to make it easier to work with them in the regression. When transforming the data, problems with outliers or distributional problems are eliminated (Field, 2013). The dependent variable (sales growth) is transformed into a categorical variable. To measure sales growth, the respondents were being asked to provide information about the amount of sales in the current year and the amount of sales three years before that. This is being transformed to the relative difference, expressed in percentages. When respondents couldn’t give an answer on one of the two questions, the results of this firm are left out. This is also the case when respondents state that the company exists less than 3 years. These percentages are being converted into different categories: category 1 contains the firms, which have a sales decline of more than 15% in the last three years. Category 2 includes all firms, which have a sales decline that is smaller than 15%, or firms that have a sales growth of less than 15%. The sales of these firms have stayed relatively stable. The last group contains the firms that have experienced sales growth larger than 15%. Hypothesis 4 is being tested for the entire sample, but on top of that there is done an analysis on the difference between industries. The variable industry is being coded as 1: manufacturing, 2: retail and 3: other services. In this way, results across the different groups can be compared. This analysis can provide companies with more detailed information. For all information concerning the coding of the remaining variables, appendix 2 can be consulted.
4.1. Descriptive statistics
First, a summary of the key characteristics of the variables used in the analysis is offered. Some of the cases are excluded from the results, because the data was not able to provide adequate information. For example, several cases are excluded when measuring sales growth, namely 6.240. This is equal to 39% of the total cases. The reason for this figure to be so high can be that companies don’t exist for three years yet.