The impact of competition on bank performance

Hele tekst

(1)University of Amsterdam - Amsterdam Business School MSc Business Economics, Finance track. The Impact of Competition on Bank Performance Master Thesis July 2014. Popescu Vlad-Alexandru (10622802) Supervisor: R.I. Todorov.

(2) Abstract In this thesis I examine the impact of concentration in the banking sector using a sample of 21,240 banks from 101 countries during the 2001-2012 period. Competition in this sector is an important subject of public debate. Profitability and stability of banking institutions are two important fields influenced by competitive forces from the banking sector. Regulatory and monitoring institutions often believe that competition is a detrimental factor to bank stability. I find that stability increases when concentration increases in highincome countries, while an opposite relation can be observed in countries with lower income. Concentration is also related to the level of profitability. There might be a fear that an increase in concentration may be favorable to banks which could see their profits increase due to their growing market power. However, this is not necessarily the case, as the results in this thesis show. I find a negative relation between concentration and profitability of banks. However, banks from Scandinavian Law countries exhibit a different pattern from this point of view. This thesis will extensively examine on one hand the relation between competition and profitability of banks and on the other hand the extent to which stability and competition are related to each other, under the important assumption that concentration and competition are positively related to each other.. Keywords: competition, Herfindahl, return on assets, stability, panel data. 2.

(3) Contents I.. Introduction ........................................................................................................................ 4. II. Literature Review ............................................................................................................... 6 III.. Methodology and Hypotheses....................................................................................... 11. Hypotheses ........................................................................................................................... 13 IV.. Data and Descriptive Statistics ..................................................................................... 14. V. Results .............................................................................................................................. 18 Profitability and Concentration ............................................................................................ 18 Legal Origin ......................................................................................................................... 19 Types of Banks ..................................................................................................................... 21 Type of Ownership ............................................................................................................... 21 Mergers and Acquisitions..................................................................................................... 22 Concentration and Stability .................................................................................................. 23 VI.. Robustness Check ......................................................................................................... 23. Other measures of profitability ............................................................................................ 24 The Four-Bank Concentration Ratio .................................................................................... 24 Smaller Samples ................................................................................................................... 25 Fixed Effects ........................................................................................................................ 25 VII.. Conclusions ................................................................................................................... 26. References ................................................................................................................................ 27 Appendix ............................................................................................................................................... 31. 3.

(4) I.. Introduction The banking sector has been subject to important structural changes over the last. decades as a result of deregulation, mergers, financial crises and financial innovation. The banking consolidation in the recent years has enhanced the public debate regarding the impact of competition and concentration on bank performance (Berger et al, 1999 and 2004). There are several important aspects related to competition in the banking sector, but probably two of them receive the highest level of attention from the literature.. The first of them is banks’ profitability. In many countries, in order to respond to a decrease in performance, banks tried to increase the magnitude of their operations. This has been done through mergers and acquisitions or through the bank’s own efforts to increase its market share. This led to an increase in the concentration ratios in banking sector which could have weakened competition in this sector (Scholtens, 2000).The increase in the number of mergers and acquisitions, especially in the mid-1990s, increased the fear that monopolistic tendencies will occur in the banking sector. Regulatory institutions feared that banks could develop into huge institutions that would be able to distort the market and abuse their power through squeezing more profits from borrowers. This increased power was thought to increase banks’ profitability.. In order to find out whether this belief is correct or not, I will regress banks’ profitability, quantified through several measures as return on assets, return on equity and net interest margin against several measures of concentration (Herfindahl-Hirschman Index and four-bank concentration ratio). Regarding the relation between competition and concentration, there is no consent in the literature. However, Claessens and Laeven (2012) find that the H-statistic, a measure of competition, is positively related to concentration in the banking sector. 4.

(5) The second aspect related to competition in the banking sector that will be discussed in this thesis is stability of banks. Regulators and monitoring institutions often look at competition as a hampering factor to stability. However, as the extant literature already showed, this is not necessarily the case. In fact, regulations were found by several authors to be responsible for the increase in bank fragility when competition intensified. The measure used in this thesis to quantify bank performance is z-score. This will be regressed against the previously reminded measures of competition.. This research would contribute to the existing literature by analyzing the relation between competition and performance using several measures of banking competition and performance. I expect my results to show that the differences in performance of banks can be explained to a large extent by the factors included in my empirical analysis which generates new insights for a better understanding of the impact of competition in the banking sector. The results of this thesis could be useful to banks in order to better understand the competitive forces that drive the banking system, as well to regulators and monitoring institutions that need to reconsider the importance of competition for assuring the stability and soundness of the financial system.. This thesis is organized as follows. Section II presents an extended literature review related to concentration, profitability of banks and stability. Section III illustrates the methodology used in the thesis. Section IV introduces the dataset used and provides some information related to descriptive statistics. Section V presents the main findings of this thesis, while in section VI, I run several robustness tests. Section VII concludes.. 5.

(6) II.. Literature Review Competition between banks is intensifying (Weill, 2013) with both positive and. negative results for the stability of banking system. Bikker and Groeneveld (1998) find that banks in the European Union face monopolistic competition, which encourages the view that competition is strong in the banking sector. Thanks to the liberalization and deregulation in the last two decades, banks from developed countries are now able to enter emerging markets which were previously regarded as inaccessible (Ferraris and Minetti, 2013).. There are two opposing hypotheses concerning market concentration and its impact on bank performance: the Structure-Conduct-Performance (SCP) hypothesis and the Efficient-Structure (ES) hypothesis. The former states that a higher level of market power (due to barriers to entry for foreign banks for example) yields monopoly powers and leads to concentration and a higher profit (Short, 1979, Molyneux et al., 1996, and Samad, 2005). This could be detrimental to consumers. The efficient-structure hypothesis asserts that the performance of a bank is not caused by market structure. So market concentration is not the cause of superior proﬁtability, but more efficient banks gain market share at the expense of less efficient banks, thus making the market more concentrated (Demsetz, 1973, Evanoff and Fortier, 1988). Conflicting predictions regarding the relationship between stability and competition which can be found in the existing literature can be attributed to the manner in which competition is assessed in the studies which stick to the Structure-ConductPerformance theory and that take as granted a negative relationship between concentration and competition. From this point of view, a highly concentrated market is considered to provide strong incentives to banks to abuse their market power. However, there are several studies in the literature that point out at the possible positive relation between the two (competition and concentration). Claessens and Laeven (2012) find that competition and 6.

(7) concentration are positively related to each other for a sample of banks from 50 countries. Baumol (1982) develops the theory of contestability, arguing that potential competitors are able to quickly enter or exit a market without the fear that they will see their capital lost. Moreover, they have the same cost functions as the firms which already have a presence in the market. These views encourage the creed that competition and concentration are positively related to each other. This thesis is written under the important assumption of a positive relation between concentration and competition. As time passes, less efficient banks will be driven out of the market by competitive forces. More efficient banks will remain in the market and will face stronger and stronger opponents since the less efficient banks will leave the market. This will increase competitiveness, despite the fact that concentration may also increase.. Gilbert (1984) surveys different empirical studies to find a significant influence of market structure on the performance of banks. Short (1979) uses a sample of 60 banks to find that a higher rate of concentration leads to higher profits for banks. He also finds other explanatory variables for bank profits, among which the form of ownership (private vs public). However, Scholtens (2000) investigates the relation between concentration and bank profitability in several developed countries and finds a weak relation between bank profitability and concentration, so a higher level of concentration does not necessarily mean a growth in profit. So there is a lack of consent in the literature concerning the relation between profitability and concentration. My thesis will provide new insights into how profitability is related to concentration and competition.. But not only profitability of banks is important from the point of view of banking competition. Institutions regulating banks fear that intensifying competition among banks will lead to an increase in the vulnerabilities of the banking system. From the existing literature we find that different authors have different views regarding the impact of 7.

(8) competition on the stability in the banking sector. For example, Keeley (1990) argues that there is a connection between the increase in competition and the increase in instability, due to the fact that risky behavior becomes more attractive to banks when competition increases. However, Beck, Demirgüç-Kunt and Levine (2006) find that competition has a positive role on bank stability and that countries that hamper competition increase the chances of bank failure, while Honohan and Stiglitz (2001) argue that competition may threaten stability mostly in developing countries. Boyd and de. Nicolo (2005) show that, although regulators around the world believe that a serious increase in competition between banks will finally lead to financial instability due to the fact that banks take more risk, this is not always the case. In fact, they discover a mechanism that makes the banks to behave in a different manner. So, contrary to the usual belief, when bank competition goes down, banks will be stimulated to charge higher interest rates for the loans they grant to borrowers. This means that borrowers face a greater default risk, thus increasing the risk faced by banks, because the probability that they will not recover their money goes up. So this would be a positive aspect of bank competition.. An interesting fact is that there are ambiguous predictions with respect to the relationship between bank stability and competition when empirical studies focus on individual countries. However, studies focusing on cross-country analysis found a positive relation between competition and bank stability, so there is no need for monitoring and regulatory institutions to fear competition. Where this was not the case, regulatory failures were in fact regarded as the culprit (Beck, 2008). Beck, de Jonghe and Schepens (2013) conduct an empirical cross-country analysis and find a high level of heterogeneity in the relation between competition and stability across countries. An intensifying level of competition will affect more the stability in countries with stricter restrictions, more financial development of stock exchanges and more efficiency in credit information sharing. Schaeck, Cihak and Wolfe (2009) use the H-Statistic (Panzar and Rosse) for a sample of 38 countries 8.

(9) in the period 1980-2003 to conclude that a higher level of competition in the banking sector will indeed provide more stability.. Caminal and Matutes (2002) showed that although banking deregulation, by increasing competition, is seen as a major factor that leads to a higher probability of crises, this is not the case. They compare a monopoly system with a competitive one and assert that even a monopolist bank could go bankrupt due to the fact that it takes more aggregate risk when engaging agency problem, so a monopoly is not necessarily a good way to increase banks’ profits.. The level of regulations and economic freedom is also a relevant aspect to my analysis. Banks that operate in countries with more economic freedom and better regulations could face better conditions than their counterparts from other countries. Claessens and Laeven (2003) use a sample of banks from 50 countries and the Panzar and Rosse model (1987) to find that the most important factor which drives competition is the contestability of the local market. In other words, a market which is highly contestable can be easily entered by a foreign bank, unlike a market with a high level of governmental regulation. As a consequence, more liberalized markets are more competitive. So, the level of economic freedom could be taken into account as a determinant of banks’ profits.. Berger et al. (2003) take into consideration foreign-owned and state-owned banks from developing countries to show that different types of banks are differently related to competition and any empirical model aiming to find a link between competition and performance of banks must take this into account. Furthermore, the level of bank concentration seems to have smaller effect in countries with a smaller level of governmental regulation of banking sector and more foreign bank entries. Berger’s paper provides evidence. 9.

(10) in favor of using the form of ownership (public or private) and the nationality criteria (domestic or foreign) as factor variables for banks’ performance.. Bank size was considered in several papers as a driving factor of performance. Shehzad, de Haan and Scholtens (2013) do not find enough evidence to reject the hypothesis that bank size is not related to profitability. Some academics argue that bank size is positively related to profit, while others say that, on the contrary, the two variables are negatively correlated. Hameetemam et al. (2000) find bank size to be negatively related to net profit before income tax. Scholtens (2000) finds evidence on the relation between bank size and profitability. In his view, the profits of banks are negatively related to bank size, measured by total assets. Fadzlan and Chong (2008) also find a negative relation between size and profitability of Philippines banks. Molyneux and Wilson (2004) find that even though there are some significant relationships between size and profit in some estimation, there is no overall convincing evidence for a consistent size-profitability relationship.. Legal origin of countries may impact the performance of banks. La Porta, Lopez-deSilanes and Shleifera (2000) argue that in common law countries, investors are better protected than in civil law countries. This relationship is even stronger when only French Civil Law countries are taken into account since German Civil Law and Scandinavian Civil Law are somewhere between French Civil Law and English Common Law. Better investor protection will lead to better financial results. Hence the legal origin could be taken into account as a factor influencing the performance of banks. I hypothesize and later prove that legal origin has a significant impact on bank performance and that competitive forces vary across countries with different legal origin.. 10.

(11) III.. Methodology and Hypotheses The main objective of my thesis is to analyze if bank competition is a strong. determinant of performance of banks. I will use panel data for 21,240 banks in 101 countries for a period of 12 years between 2001 and 2012. I am going to regress banks’ performance, quantified through profitability and stability against two measures of competition to see whether the relation between the two is significant. Profitability will be measured as return on assets, return on equity and net interest margin. Stability will be quantified by z-score.. The Return on Assets will be calculated using the following formula:. , =.

(12) , ,. (3.1). The Return on Equity will be computed using this formula:. , =.

(13) , ℎ ℎ ′ ! ",. (3.2). The Z-score is a measure of distance from insolvency (Roy, 1952) and is calculated based on this equation (Beck, de Jonghe and Schepens, 2013):. $ − =. , +. ,. '()*+,,. (3.3). Where ROAi,t represents return on assets for bank i in year t, E/Ai,t is the ratio of equity to total assets and ' denotes the variance of return on assets. The Herfindahl-Hirschman Index (HHI) will be computed using the following formula:. 11.

(14) 0. --

(15) = . / 12. (3.4). Where si is the market share of bank i. A HHI index greater than 0,25 shows a low level of competition and a high level of concentration.. Another measure of concentration that I will use is the four bank concentration ratio – the total market share of the four largest banks in an economy. It is calculated as the sum of the market shares of the largest four banks in the economy.. By including several control variables to isolate factors that are directly related to bank performance, the accuracy of this research and its results will increase. To control for omitted variable bias, I will make use of bank-specific characteristics, especially ratios like Fixed Assets to Total Assets, Capital Ratio and Liquidity Ratio. I will also use country macro-economic characteristics like GDP/capita Growth and Interest Rate. Table 1 in the Appendix provides detailed information on the definition of variables used in this thesis.. For my empirical model I will run the following regression: 4, = 56 + 52 Χ, + 7 + 8 + 9,. (3.5). Where υ:,; is a vector consisting of several measures of bank performance (profitability and stability) and Xi,t is a vector consisting of several measures of bank competition (Herfindahl-Hirschman Index, four-bank concentration ratio). η is a vector of country-specific characteristics and ω a vector consisting of bank-specific characteristics. β1 is the regression coefficient of interest and shows the relation between the level of concentration and performance of banks. It shows the extent to which an increase in the level of concentration affects bank’s profitability and stability. An important assumption of this 12.

(16) thesis is the adherence to the Efficient-Structure Hypothesis. Athanasoglou et al (2008) cannot find any evidence in favor of the Structure-Conduct-Performance hypothesis.. I use clustered standard errors at country-year level. OLS standard errors are assumed to be unbiased only when there is no correlation between residuals across the observations in the data set. In other words, they are unbiased when the independence assumption is not violated. Since using panel data often violates this assumption, other method of estimating standard errors must be used. Petersen (2009) finds clustered standard errors to be unbiased when the independence assumption is violated in panel data. The standard errors which I use are clustered at the country-year level, since other clustering methods like clustering at bankyear level yield smaller standard errors.. I use several robustness tests to enforce my results. First, I use other measures of profitability than Return on Assets. Instead of ROA, I use Return on Equity and Net Interest Margin. Second, instead of the Herfindahl index, I use the four bank concentration ratio. Third, I restrict my sample to banks for which data is available for all the years during the period between 2001 and 2012. Fourth, I reduce my sample by dropping the four and eight largest banks from each country. Fifth, I use country fixed effects and year fixed effects. My main results withstand all these tests.. Hypotheses As already mentioned, there is no consent in the literature regarding the relation between profitability and stability on one side and concentration on the other side. As a consequence, further research in this matter is necessary. However, extant literature may provide some insights in order to formulate some hypotheses with respect to the relation that. 13.

(17) someone would expect to find between the previously reminded variables. The main hypotheses of my thesis could be formulated as following:. 1. An increase in the level of market concentration will increase the profits of banks. This would be consistent with the findings of Short (1979). 2. There will be important variation between countries with different legal origins. 3. An increase in competition will positively affect bank stability, but this will depend very much on the level of bank regulation, level of financial development etc. A country with a low level of development could face an increase in fragility, rather than stability, when competition increases. This would be consistent with the research of Beck, de Jonghe and Schepens (2013).. IV.. Data and Descriptive Statistics My sample covers the period between 2001 and 2012 for 21,240 banks from 101. countries. In order to find whether competition between banks has an impact on performance and stability, I combine several databases. To obtain data related to banks’ financial situations, I use Bankscope, a database compiled by Bureau Van Dijk. From this database I use data on commercial banks, cooperative banks, savings banks, and bank holding and holding companies, which represent 63.69%, 12.40%, 11.10% and 12.81% of the sample, respectively. Whether a bank is foreign or domestic may influence my results, so I take into account the type of ownership (foreign or domestic). To find whether a bank is foreign or not, I use a database compiled by Claessens and Van Horen (2013) who offer information on the ownership of several banks for the period between 1995 and 2009.. 14.

(18) For data related to country-specific characteristics, I mostly use the World Development Indicators database compiled by the World Bank to get data for variables like GDP per capita growth, inflation and income group. The values for the interest rates in each country from the sample are taken from the International Financial Statistics compiled by IMF. I manually complete this database where certain data is missing using central banks’ websites. Data related to economic freedom index for each year is collected by the World Heritage Foundation. I also use the Systemic Banking Crises Database of Laeven and Valencia (2008, 2010 and 2013) to take into account the impact of systemic banking crises on profitability and stability.. Besides that, in order to gain insights into how the country legal origin affects the performance of banks, I use a database created by La Porta, Lopez-de-Silanes and Shleifer (1999) who divide the legal origins of countries in five categories: Common Law, French Civil Law, German Civil Law, Scandinavian Law and Socialist Law. However, this database was compiled before 2000 and some things related to Law have changed. Communism fell in most countries, so the Law in each of them has suffered significant changes when compared to the communist era or the last decade of the 20th century. So I do not take any country in my database as having Socialist Law, since most of them revised their constitutions and other legal aspects. So, for the countries which were seen as having socialist legal origin by La Porta et al., I use the CIA Factbook which among others, provide relevant information regarding the concerned topic.. After merging all these databases, I apply several selection criteria. First, I drop observations with missing data on most relevant variables. Second, I take care to eliminate all duplicates. In case a single bank reported data at both consolidated and unconsolidated level, I drop the unconsolidated values to avoid further duplicates. Third, I drop the banks with data. 15.

(19) available for less than three years and countries with less than 100 bank-year observations, in order to have a reasonable cross-sectional set of data. Finally, I control for outliers by winsorizing at 1% and 99% levels.. After applying all these filters, I obtain a final sample of 21,240 banks from 101 countries, totaling 194,408 observations. All data is expressed in millions of US dollars. The summary statistics for the variables used in my analysis are presented in Table 2 from the Appendix. A correlation matrix is provided in Table 3. Although most of the correlation coefficients are significant, their values are quite small, so I can assume that my estimators are not biased from this point of view.. As it can be seen from Table 2, ROA has an average value of 0.658%, with a range between -5,176% and 5.469%. ROE has a larger average value of 6.284%, since equity is much smaller than total assets. It ranges from -52.63% to 34.77%. ROA has a smaller variation than ROE, as shown by the smaller values for standard deviation and standard error of mean. Z-score, the measure of stability that I use in this thesis is on average 24.07, with the minimum value of -33.93 and the maximum value of 202.2.. Regarding the measures of concentration used in this thesis, it can be observed from Table 2 that the average value of the Herfindahl Index is 0.0555 (the Herfindahl Index can have values between 0 and 1; a value of 1 would represent a monopoly). The minimum value for my sample is 0.0111, while the maximum is 0.291. The latter value shows a high level of concentration. The Four-Bank Concentration Ratio measures the total market share of the largest four banks in the economy and has an average value of 32.72%. The maximum value is 88.69%.. 16.

(20) As it can be seen from Table 2, the average values for the concentration measures are not very high for the entire sample. However, the situation changes when taking into account smaller samples. For example, Scandinavian Law countries (Denmark, Norway, Sweden, Finland and Iceland) have much higher levels of concentration than the average country in the sample, with an average Herfindahl Index of 0.246 and an average Four-Bank Concentration Ratio of 80.68%. On the opposite side, the Common Law countries have an average Herfindahl of only 0.024 and an average Four-Bank Concentration Ratio of 21.54%. French Civil Law and German Civil Law countries are between these two extremes. The distribution of concentration measures by legal origin of countries taken into account in this analysis is presented in Figure 1 of the Appendix.. Not only that there are important differences by legal origin, but the income groups as established by the World Bank present their own characteristics. High income countries have the lowest level of concentration measured by both indices used in this thesis, while lowermiddle income countries have the highest one. Figure 2 illustrates the levels of concentration across income groups.. There is also a significant cross-country variation in concentration levels. Denmark, Sweden and Netherlands present a high level of concentration as measured by Herfindahl Index, while countries like USA and Japan have average values below 0.05. Also, Denmark Sweden and Netherlands, as well as Norway and Switzerland have a big value for the other measure of concentration (Four-Bank Concentration Ratio). Figure 3 exhibits a more detailed overview regarding the concentration measures across several countries from the sample.. 17.

(21) V.. Results In this section, I provide my main results regarding the relation between competition. and profitability on one hand and, on the other hand, between competition and stability measured by z score. Detailed results concerning the empirical analysis run in this thesis are presented in the Appendix at the end of this paper.. Profitability and Concentration Table 4 from the Appendix provides information related to the determinants of banks’ profitability. Contrary to my first hypothesis, I find a negative relation between profitability and concentration when including bank- and country-specific characteristics in my analysis. An increase in the level of the Herfindahl Index will lead to a decrease in banks’ profits measured by Return on Assets. The results are statistically significant at 1% level. As the Efficient Structure theory asserts, market structure is the result and not the cause of superior performance of banks. More efficient banks gain market share at the expense of less efficient ones, so the increase in concentration can be related to the higher level of competition between banks. More efficient banks can remain on the market, while less efficient ones leave it, maybe as a result of their poor strategy. The banks remaining in the market will face their more powerful competitors, since the weak ones left the market.. This finding of my thesis contradicts Short (1979). The reason for the differences between my thesis and Short’s paper could be related to the sample used in my analysis. Short uses only 60 banks from Canada, Japan and ten countries from Western Europe, while I use 21,240 banks from 101 countries all over the world. Moreover, Short has an average of five banks per country, while I have at least nine banks per country.. 18.

(22) Regarding the sources of cross-country variation, I first find that banks from countries with more economic freedom, as measured by the World Heritage Foundation’s Index, have a lower level of profitability. This relationship also stands when I control for the form of ownership (foreign/domestic and public/private). Second, countries with higher interest rates will see the profitability of their banks increasing, but this depends on the legal origin as well. More details about this will be provided in the next subsection regarding the impact of legal law on the profitability of banks. A positive relation between interest rates and banks’ profitability is also found by Molyneux and Thornton (1992). Third, systemic banking crises are negatively related to profitability of banks. When a country is confronted with a crisis, banks from its system will react negatively, as shown by the reduction in the level of profitability.. Legal Origin Concerning my second aim, that is to investigate whether legal origin matters for the performance of banks, I find for Scandinavian Law countries that a positive relation exist between profitability measured by ROA and concentration measured by Herfindahl Index, while the negative relation stays the same for Common Law countries just like in the case of the whole sample. This could have something to do with the fact that the Scandinavian Law countries already have a highly concentrated market, while Common Law countries have low levels of concentration (as shown in the Data and Descriptive Statistics Section). When the level of concentration is already large like in Scandinavian Law countries, banks face lower competition from opponents, so in this case competition does not erode profits. As a consequence, the profitability of banks from a highly concentrated market increases when concentration increases. The results regarding the relation between concentration and profitability when legal origin is taken into account are summarized in Table 5.. 19.

(23) As reminded in the previous subsection, there is an important cross-country variation when legal origin is accounted for. Although countries with Common and Civil (French or German) Law exhibit a positive relation between the level of interest rate and profitability of the banking sector, Scandinavian Law countries present an opposite relationship. The higher the interest rate is, the lower the profitability of Scandinavian Law countries.. The situation is the same for the Credit Depth of Information Index. As shown in the Table 1 of the Appendix, higher values of this index indicate that more credit information is available through a private bureau or public credit registries in order to facilitate lending decisions. In Scandinavian Law countries, when this index goes up, the profitability of the banks also goes up, which is not the case for countries with other legal origin. A possible explanation for this situation is that Scandinavian Law countries have a lower average level of Credit Depth of Information Index than the other countries (on average, it has a value of 4.03). Countries like United States of America and United Kingdom (countries with Common Law system) have a value of six for the entire period 2001-2012.. Luoto et al (2007) find that higher availability of credit information promotes transparency in lending. It mitigates adverse selection and lowers the proportion of borrowers who default on their loans. So the profitability of banks should go up if fewer borrowers defaulted. Since on average less information is available in the Scandinavian Law countries than in Common Law ones (as shown by the average levels of Credit Depth of Information Index), banks from the former countries may face higher rates of defaults. So an increase in the level of credit information availability will help them obtain greater profits. To explain the negative relation found for other countries which already have a high level of credit information availability, it can be inferred that obtaining more information than already. 20.

(24) available will raise the costs for banks, and the decrease in the proportion of defaults thanks to better availability of information will not compensate for the increase in the costs level.. Types of Banks Table 6 extends the results from column (4) from Table 4 to subsamples of banks by their type. As previously reminded, I use data for commercial banks, savings banks, cooperative banks and bank holding and holding companies. Bank holding and holding companies seem to be positively related to concentration, as shown by column (4) in Table 6, while savings banks exhibit a different pattern, just like in the case of the whole sample. Using dummies for the different types of banks (unreported results) yields the same results. The profitability of commercial banks, savings banks and cooperative banks is negatively related to concentration, while an opposite trend can be observed for bank holding companies. Bank holding companies have the possibility to better diversify risk and have more options than commercial banks for raising capital. An interesting fact is that unreported results in this thesis show that the positive relation between concentration and profitability holds only for non-USA bank holding and holding companies (BHCs), while for USA BHCs, the relation is found to be negative. The results are significant at 5% and 1% levels, respectively. Explanations for this fact might be related to variations in omitted factor, but for the moment this is not the object of this thesis and can be further investigated in future research.. Type of Ownership Another aim of this thesis is to find whether foreign and public banks behave differently than domestic and private banks, respectively, when concentration increases. In. 21.

(25) this subsection, I investigate the variation in profitability between foreign and domestic banks on one hand and, on the other hand, between publicly and privately held banks.. First, foreign banks are more affected than domestic ones when concentration increases. Their profitability will go down when compared to domestic banks. This relationship stands even when Return on Equity and Net Interest Margin are used instead of Return on Assets. Second, public banks have lower profitability than private ones when concentration increases. Farabullini and Hester (2005) also found for a sample of Italian Banks that banks performance increased after privatization. Hau and Thum (2009) also found that after the subprime crisis, banks which were already publicly held faced the largest losses. So my findings regarding the worse performance of public banks is in line with the extant literature.. Table 7 of the Appendix exhibits detailed results concerning the relation between competition and profitability when the form of ownership is accounted for.. Mergers and Acquisitions As threshold for mergers and acquisitions (M&A), I use the 50% level for the growth in total assets. If a given bank in a given year had a growth in total assets higher than 50%, then I take it as an M&A. My initial finding that an increase in concentration negatively impacts the profitability of banks stands when the growth in total assets is below this threshold. However, banks with a growth above this level saw their profits increasing when concentration went up. The coefficient for the Herfindahl Index is statistically significant at 10% level. Details are provided in Table 8 of the Appendix.. 22.

(26) Concentration and Stability Regarding the third objective of my thesis, I find that when the overall sample is taken into account, there is no significant relation between competition and stability. However, when I split my sample in two parts, one for high-income countries and one for the others, I find two interesting facts. First, in high-income countries, when the level of concentration measured by the four-bank concentration ratio is higher, stability is also higher. Second, there is an opposite relation in countries with lower income (upper-middle income, lower-middle income and low income), where an increase in concentration destabilizes the banking system. The results are exhibited in Table 9 of the Appendix. This is consistent with my third hypothesis and with the findings of Honohan and Stiglitz (2001) and of Beck, de Jonghe and Schepens (2013). So there is indeed a high level of heterogeneity across countries in the relation between competition and stability. Developed countries present a positive relation between stability in the banking sector and concentration, while increasing competition may act as a threat in less developed ones.. I get the same results when I control for foreign ownership. Moreover, foreign banks still seem to be more negatively affected than their domestic counterparts when concentration increases. Extensions of the relation between stability and competition are presented in Table 10.. VI.. Robustness Check I use several robustness tests in this thesis. First, I use Return on Equity and Net. Interest Margin instead of Return on Assets as proxies for bank profitability. Second, I. 23.

(27) replace the Herfindahl index with another measure of concentration: four bank concentration ratio. Third, in order to see whether my results are affected or not by the sample I use in this thesis, I restrict the sample to banks for which data is available for all the years during the period between 2001 and 2012. Fourth, I drop the four and then the eight largest banks in the sample to see whether the relation initially found stands. Fifth, I use country- and time-fixed effects to better control for omitted variable bias.. Other measures of profitability In Section V of this thesis, in order to test for the relation between competition and profitability of banks, I mainly used ROA. To see whether my results were influenced or not by the variables I chose to use in my analysis, instead of using Return on Assets as a proxy for profitability, I use Return on Equity and Net Interest Margin. This will enforce my main findings in Section V. I still find a negative relation between concentration and profitability, even if I use different measures for the latter. Columns (1) and (2) of Table 11a in the Appendix illustrate the impact of the Herfindahl Index on ROE and Net Interest Margin, respectively.. The Four-Bank Concentration Ratio Instead of running regressions of a measure of profitability against the Herfindahl Index, I use the Four-Bank Concentration Ratio. Even when using a different measure of competition, my empirical analysis yields the same results. Indeed, profitability is negatively related to concentration, despite using the total market share of the first four banks in each country for a given year. Columns (3) and (4) from Table11a show the results described in this subsection.. 24.

(28) Smaller Samples The results found in Section V could be biased due to the sample I used. To check this, I make use of three different samples. First, I keep only the banks for which data is available for all the period 2001-2012. So banks with available data on 11 years or less are dropped from the sample. Second, I drop the first four largest banks in each country and third, I do the same thing for the first eight banks in each country. Table 11b in the Appendix shows the relation between Return on Assets and Herfindahl Index for various subsamples. All columns yield significant results at 1% level. The profitability of banks remains negatively related to concentration in all three cases, which enforce the main findings of this thesis.. Fixed Effects There are several authors who emphasize that using fixed effects provides significant evidence in favor of the model used in an econometrical analysis and that it reduces omitted variable bias. Table 11c provides a comparison between a model without fixed effects (more exactly the model used in column (4) of Table 4) and two models which use year- and country-fixed effects in order to test for the relation between profitability, measured by Return on Assets and concentration quantified through the Herfindahl Index. Even in this case, I find a negative relation between the two and the coefficient of interest is statistically significant at 1% level for all the cases.. 25.

(29) VII. Conclusions As reminded before, competition is an important factor for banks and regulatory institutions, as well. In this thesis I examine whether there is a significant relation between profitability and stability of banks on one side and two measures of concentration on the other side.. I find that concentration negatively impacts profitability of banks. This is consistent with the Efficient-Structure Hypothesis. However, this relationship does not hold for Scandinavian Law countries, where a more concentrated banking system than in other countries exists.. Regarding the relation between stability and competition, in line with Beck, de Jonghe and Schepens (2013), I find important heterogeneity across countries. Competition provides stability in high-income countries, while it causes instability in less developed ones. This has important policy implications for regulatory institutions in developed countries, where there is a strong debate related to the impact of competition on stability of the banking system. But as this thesis show, it is reasonable to assume that competition will provide more stable banking systems.. 26.

(30) References Athanasoglou, P., Brissimis, S., Delis, 2006, Bank-specific, industry-specific and macroeconomic determinants of bank profitability, Journal of International Financial Markets, Institutions and Money, Vol.18(2), pp.121-136. Beck, T., Demirgüç-Kunt, A., Levine, R., 2006, Bank concentration, competition, and crises: First results, Journal of Banking & Finance, Vol.30(5), pp. 1581-1603. Berger et al, 1999, The Consolidation of the Financial Services Industry: Causes, Consequences, and Implications for the Future, Journal of Banking and Finance, Vol.23, pp. 135-194. Berger et al, 2004, Bank Concentration and Competition: An Evolution in the Making, Journal of Money, Credit, and Banking, Vol. 36, No. 3 (Part 2), pp. 433-451. Beck, T., 2008, Bank competition and financial stability: friends or foes ?, World Bank, Policy Research Working Paper Series No.4656. Beck, T., de Jonghe, O., Schepens, G., 2013, Bank Competition and Stability: Cross-country Heterogeneity, Journal of Financial Intermediation, Vol.22(2), pp.218-244. Bikker, J.A., Groeneveld, J.M., 1998, Competition and Concentration in the EU Banking Industry, DNB Research Series Supervision no. 8. Boyd, J., de Nicolo, G., 2005, The Theory of Bank Risk Taking and Competition Revisited, The Journal of Finance, Vol LX, No.3. Caminal, R. and Matutes, C., 2002, Market power and bank failures, International Journal of Industrial Organisation, Vol.20 (9), pp.1341-1361 27.

(31) Cerasi, C., 2010, The impact of mergers on the degree of competition in the banking industry, FEEM Working Paper No. 95. Claessens, S., Laeven, L., 2012, What Drives Bank Competition? Some International Evidence, Journal of Money, Credit and Banking, Vol.36, No.3, pp.563-583. Degryse, H., Ongena, S., 2004, The Impact of Competition on Bank Orientation and Specialization, CESifo Working Paper Series 1119, CESifo Group Munich. Demsetz, H., 1973, Industry structure, market rivalry, and public policy, Journal of Law and Economics, vol. 16, pp.1–9. Evanoff, D.D., Fortier, D.I., 1988, Reevaluation of the structure-conduct performance paradigm in banking, Journal of Financial Services Research, Vol.1(3), pp.277-294. Farabullini, F., Hester, D., 2005, The performance of some privatized Italian banks, Rivista di Politica Economica, Vol.95(7–8), pp. 235–272. Ferraris, L., Minetti, R., 2013, Foreign Banks and the Dual Effect of Financial Liberalization, Journal of Money, Credit and Banking, Vol.45(7), pp. 1301-1333. Gilbert, A., 1984, Bank market structure and competition: a survey, Journal of Money, Credit and Banking, Vol. 16, No. 4, Part 2: Bank Market Studies, 617-645. Goddard, J., Molyneux, P., Wilson, J.O., 2004, The Profitability of European Banks: A Cross-Sectional and Dynamic Panel Analysis, The Manchester School, Vol. 72, No.3. Hameeteman, D. and Scholtens, B., 2000, Size, growth, and variance among the world's largest non-merged banks, International Journal of the Economics of Business, no. 7, pp. 313-323 28.

(32) Hau, H., Thum, M., 2009, Subprime crisis and board (in-) competence: private versus public banks in Germany, Economic Policy, Vol.24(60), pp.701-752. Honohan, P., Stiglitz, J.E., 2001, Robust Financial Restraint, in Financial Liberalization: How Far, How Fast?, ed. By Caprio, G., Honohan, P. and Stiglitz, J.E., Cambridge University Press, Massachussets, pp. 26-47. Keeley, M.C., 1990, Deposit Insurance, Risk, and Market Power in Banking, American Economic Review, Vol.80(5), pp. 1183-1200.. La Porta, R., Lopez-de-Silanes F., Shleifer, A., 1999, Corporate Ownership around the World, The Journal of Finance. Vol. 54, No. 2, pp. 471-517. Laeven, L., Valencia, F., 2010. Resolution of banking crises: The good, the bad, and the ugly. IMF Working Paper 10/146.. Laeven, L., Valencia, F., 2012. Systemic Banking Crises Database: An Update. IMF Working Paper 12/163. Luoto, J., McIntosh, C., Wydick, B., 2007, Credit Information Systems in Less Developed Countries: A Test with Microfinance in Guatemala, Economic Development and Cultural Change, Vol.55, No.2. Molyneux, P., Thornton, J., Lloyd-Williams, D.M., 1996, Competition and market contestability in Japanese commercial banking, Journal of Economics and Business, vol. 48, pp. 33–45. Molyneux, P., Thornton, J., 1992, Determinants of European bank profitability: A note, Journal of Banking and Finance, No.16, pp.1173-1178. 29.

(33) Petersen, M., 2009, Estimating Standard Errors in Finance Panel Data Sets: Comparing Approaches, The Review of Financial Studies, 22(1), pp.435-480. Roy, A.D., 1952, Safety first and the holding of assets, Econometrica, Vol. 20(3) (1952), pp.431–449. Samad, A., 2008, Market structure, conduct and performance: Evidence from the Bangladesh banking industry, Journal of Asian Economics, Vol. 19, Issue 2, pp.181-193. Scaeck, K., Cihak, M., Wolfe, S., 2009, Are More Competitive Banking Systems More Stable?, Journal of Money, Credit and Banking, Vol. 41(4), pp.711-734. Scholtes, B., 2000, Competition, Growth, and Performance in the Banking Industry, Wharton School Center for Financial Institutions, University of Pennsylvania , Center for Financial Institutions Working Paper No. 00-18. Shehzad, C.T., de Haan, J., Scholtens, B., 2013, The relationship between size, growth and profitability of commercial banks, Applied Economics, Vol.45(13), pp.1751-1765. Short, B.K., 1979, The relation between commercial bank profit rates and banking concentration in Canada, Western Europe and Japan, Journal of Banking and Finance, no.3, pp.209-219. Sufian, F. and Chong, R., 2008, Determinants of Bank Profitability in a Developing Economy: Empirical Evidence from the Philipinnes, Asian Academy of Management Journal of Accounting and Finance, vol. 4, issue 2, pp.91-112. Weill, L., 2013, Bank competition in the EU: How has it evolved?, Journal of International Financial Markets, Institutions and Money, Vol.26, pp.100-112. 30.

(34) Appendix FIGURE 1 Concentration Measures by Legal Origin. Herfindahl Index 0.3 0.25 0.2 0.15 0.1 0.05 0 Common Law. French Civil Law. German Civil Law. Scandinavian Law. Four Bank Concentration Ratio 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% Common Law. French Civil Law. German Civil Law. Scandinavian Law. 31.

(35) FIGURE 2 Measures of Concentration across Income Groups. Four-Bank Concentration Ratio 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% High Income. Upper-middle Income Lower-middle Income. Low Income. Herfindahl Index 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 High Income. Upper-middle Income Lower-middle Income. Low Income. 32.

(36) FIGURE 3 Measures of Concentration across Countries. Herfindahl Index 0.3 0.25 0.2 0.15 0.1 0.05 0. Four-Bank Concentration Ratio 100.00% 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00%. 33.

(37) TABLE 1 Definition of variables. This table provides information related to variables used in the empirical model described in the methodology section. The first column illustrates the name of the variable, while the second and the third show the source and the definition for each variable. Panel A reports information related to bank-specific characteristics, while panel B reports information concerning country-specific characteristics. Bank specific characteristics are mainly taken or computed based on the data available in Bankscope, a database compiled by Bureau van Dijk containing financial information related to banks all over the world.. Country specific characteristics are mainly downloaded from the World Development Indicators database. Data related to interest rates in each country are taken from the IMF International Financial Statistics database, while the data related to the occurrence of crises is taken from another database compiled by Laeven and Valencia (2008, 2010, 2013). Information regarding the legal origin of each country is provided by La Porta, Lopez-deSilanes and Shleifer (1999). Income groups are specified for each country based on the classification made by the World Bank.. 34.

(38) Variable name. Source. Definition. Return on Assets (ROA). Own calculations based on Bankscope data. Net Income/Total Assets. Return on Equity (ROE). Own calculations based on Bankscope data. Net Income/Shareholders Equity. Net Interest Margin. Bankscope. Net Interest Margin from Bankscope database *. Z Score. Own calculations based on Bankscope data. (ROA+Equity/Total assets)/volatility of ROA. Herfindahl Index. Own calculations based on Bankscope data. ∑/ where s is the deposits market share of bank i in a given country and in a single year. Four Bank Concentration Ratio. Own calculations based on Bankscope data. Total Market Share of the four largest banks in a country. Ln(Total Assets). Own calculations based on Bankscope data. Natural logarithm of Total Assets. Capital Ratio. Own calculations based on Bankscope data. Equity/Total Assets * 100. Cost to Income Ratio. Own calculations based on Bankscope data. Overheads / (Net Interest Revenue + Other Operating Income) * 100. Liquidity Ratio. Bankscope. Liquid Assets / Deposits and short term funding * 100. Loan Loss Provisions Ratio. Own calculations based on Bankscope data. Loan Loss Provisions / Net Interest Revenue * 100. Fixed Assets to Total Assets Ratio. Own calculations based on Bankscope data. Fixed Assets/Total Assets * 100. Net Loans to Total Assets Ratio. Bankscope. Net Loans/Total Assets *100. Public. Bankscope. Dummy variable equal to 1 if the bank is publicly owned. Foreign. Database on Bank Ownership (Claessens and van Horen, 2013). Dummy variable equal to 1 if the bank is owned by a foreign bank. A.Bank-specific characteristics. 35.

(39) B.Country-specific characteristics GDP/capita growth. World Development Indicators. Growth of GDP per capita (%). Inflation Rate. World Development Indicators. Rate of inflation, GDP Deflator (%). Ln(Market Capitalization). World Development Indicators. Natural Logarithm of Market Capitalization of Listed Companies. Ln(Stocks Traded). World Development Indicators. Natural Logarithm of Stocks Traded. Credit Depth of Information Index. World Development Indicators. 0=low; 6=high. Interest Rate. International Financial Statistics. Interest rate in each country in the sample. Crisis. Systemic Banking Crises Database (Laeven and Valencia). Dummy variable equal to 1 in case of systemic banking crises. 36.

(40) TABLE 2 Summary Statistics. This table presents the summary statistics for the variables used in this thesis. Column (1) shows the number of observations for each variable, while column (2) illustrates the mean. The standard error of mean is presented in column (4). Columns (5) and (6) show the minimum and the maximum, respectively, for each variable from my database. Columns (7), (8) and (9) present the 5th, 50th (median) and 95th percentiles.. As it can be seen, ROA has an average value of 0.658%, with a range between 5,176% and 5.469%. ROE has a larger average value of 6.284%. It ranges from -52.63% to 34.77%. ROA has a smaller variation than ROE, as shown by the smaller values for standard deviation (1.295 for ROA and 11.790 for ROE). Z-score is on average 24.07, with the minimum value of -33.93 and the maximum value of 202.2. The average value of the Herfindahl Index is 0.0555. The minimum value for my sample is 0.0111, while the maximum is 0.291. The four-bank concentration ratio has an average value of 32.72%. Its maximum value is 88.69%.. 37.

(41) Variables. (1) Obs.. (2) Mean. (3) Standard Deviation. (4) Standard Error. (5) Min. (6) Max. (7) p5. (8) p50. (9) p95. A. Bank-specific characteristics Return on Assets. 194,408. 0.658. 1.295. 0.003. -5.176. 5.469. -1.031. 0.637. 2.439. Return on Equity. 194,408. 6.284. 11.790. 0.027. -52.630. 34.770. -10.000. 6.667. 22.420. Net Interest Margin. 194,408. 3.822. 1.918. 0.004. 0.400. 13.400. 1.410. 3.680. 6.880. Z Score. 180,659. 24.070. 35.150. 0.082. -33.930. 202.200. -7.506. 14.670. 88.390. Herfindahl Index. 194,408. 0.0555. 0.062. 0.0001. 0.011. 0.291. 0.0111. 0.025. 0.200. Four Bank Concentration Ratio. 194,408. 32.720. 19.850. 0.045. 14.570. 88.690. 14.570. 23.700. 75.840. Ln(Total Assets). 194,407. 5.808. 1.892. 0.004. 2.079. 11.690. 3.178. 5.537. 9.516. Capital Ratio. 194,408. 11.330. 8.830. 0.020. 2.086. 66.670. 4.167. 9.375. 23.740. Cost to Income Ratio. 191,292. 71.530. 25.950. 0.059. 7.002. 200.000. 40.000. 66.670. 100.000. Liquidity Ratio. 193,171. 19.060. 22.330. 0.051. 1.340. 131.300. 2.540. 11.540. 64.930. Loan Loss Provisions Ratio. 184,569. 12.460. 22.380. 0.052. -20.000. 130.700. 0.000. 3.125. 50.000. Fixed Assets to Total Assets. 193,674. 1.775. 1.668. 0.004. 0.000. 9.780. 0.000. 1.459. 4.762. Net Loans to Total Assets. 194,233. 60.150. 18.620. 0.042. 0.400. 91.67. 23.530. 63.000. 85.680. B. Country-specific characteristics GDP/capita Growth. 193,536. 1.262. 2.431. 0.005. -6.060. 8.840. -3.650. 1.660. 4.950. Inflation Rate. 193,871. 2.739. 3.278. 0.007. -1.710. 18.830. 0.100. 2.000. 9.690. Ln(Market Capitalization). 191,581. 28.320. 1.706. 0.004. 17.260. 29.240. 24.61. 29.240. 29.240. Ln(Stocks Traded). 191,624. 28.510. 2.541. 0.006. 13.500. 29.820. 22.760. 29.820. 29.820. 38.

(42) Variables. (1) Obs.. (2) Mean. (3) Standard Deviation. (4) Standard Error. (5) Min. (6) Max. (7) p5. (8) p50. (9) p95. Interest Rate. 192,832. 3.150. 3.873. 0.009. 0.250. 48.000. 0.500. 2.000. 7.500. Economic Freedom Index. 193,693. 74.130. 8.516. 0.019. 36.600. 90.000. 52.500. 78.200. 81.200. Credit Depth of Information Index. 146,012. 5.574. 1.047. 0.002. 0.000. 6.000. 4.000. 6.000. 6.000. 39.

(43) TABLE 3 Correlation Matrix. Correlation Matrix. ROA. ROA. Z-score. Herf.. 4Bank Ln(TA) Capital Cost to Conc Ratio Income. Liq. Ratio. Loan Loss Prov. Ratio. Fixed Net GDP Inflation Ln(Mkt Ln(Stocks Credit Interest Assets Loans Growth Cap) Traded) Depth Rate to TA to TA of Info. 1. Z-score. 0.521 (0). 1. Herfindahl. 0.0665 (0). 0.0036 (0.131). 1. 4Bank Conc. 0.0268 (0). -0.024 (0). 0.956 (0). 1. Ln(TA). 0.0236 (0). -0.037 (0). 0.200 (0). 0.293 (0). 1. Capital Ratio. 0.153 (0). 0.244 (0). 0.106 (0). 0.0545 (0). -0.324 (0). 1. Cost to Income. -0.528 (0). -0.333 (0). -0.083 (0). -0.083 (0). -0.262 (0). 0.0603 (0). 1. Liquidity Ratio. 0.0379 -0.0041 (0) (0.0825). 0.415 (0). 0.416 (0). -0.0253 (0). 0.435 (0). 0.0665 (0). Loan Loss Provisions. -0.422 (0). -0.244 (0). 0.147 (0). 0.179 (0). 0.221 (0). -0.077 (0). 0.0885 0.0526 (0) (0). Fixed Assets to TA. -0.037 (0). -0.127 (0). 0.109 (0). 0.167 (0). 0.0586 (0). 0.0489 (0). 1. Net Loans to TA. -0.03 (0). -0.0452 0.0843 0.0463 (0) (0) (0) -0.0651 -0.115 0.0971 (0) (0) (0). 0.0316 (0). -0.304 (0). -0.059 (0). -0.547 (0). 0.0707 (0). 0.0393 (0). 1 1. 1. EFI. Crisis.

(44) ROA. Z-score. Herf.. 4Bank Ln(TA) Capital Cost to Ratio Income Conc. Liq. Ratio. 0.175 (0). 0.0893 (0). 0.222 (0). 0.183 (0). 0.00930 0.0778 (0) (0). -0.112 (0). 0.160 (0). 0.199 (0) Ln(Market 0.0563 Capitalization) (0) Ln(Stocks -0.066 Traded) (0). 0.0923 (0). 0.427 (0). 0.340 (0). -0.0916 (0). 0.252 (0). -0.046 (0). 0.352 -0.0059 0.216 (0) (0.0108) (0). 0.0136 (0). -0.703 (0). -0.745 (0). -0.207 (0). -0.065 (0). 0.0883 -0.359 (0) (0). -0.164 (0). 0.01 (0). -0.661 (0). -0.704 (0). -0.202 (0). -0.072 (0). 0.0886 -0.371 (0) (0). Credit Depth of Info. -0.139 (0). -0.064 (0). -0.638 (0). -0.588 (0). -0.109 (0). -0.123 (0). Interest Rate. 0.141 (0). 0.0386 (0). 0.279 (0). 0.284 (0). 0.0240 (0). EFI. -0.092 (0). -0.016 (0). -0.632 (0). -0.665 (0). -0.194 (0). Crisis. -0.151 (0). -0.092 (0). -0.002 0.0468 -0.0287 0.0340 (0.265) (0) (0) (0). GDP/capita Growth Inflation. Loan Loss Prov. Ratio -0.130 (0). Fixed Net GDP Inflation Ln(Mkt Ln(Stocks Credit Interest Assets Loans Growth Cap) Traded) Depth Rate to TA to TA of Info 0.0774 -0.092 1 (0) (0). EFI. -0.117 (0). 0.434 (0). 1. -0.082 (0). 0.157 (0). -0.192 (0). -0.404 (0). 1. -0.153 (0). -0.087 (0). 0.180 (0). -0.200 (0). -0.409 (0). 0.978 (0). 1. 0.119 (0). -0.355 -0.0672 (0) (0). -0.089 (0). 0.111 (0). -0.279 (0). -0.542 (0). 0.559 (0). 0.588 (0). 1. 0.142 (0). -0.052 (0). 0.227 0.00680 0.125 (0) (0.0038) (0). -0.127 (0). 0.230 (0). 0.537 (0). -0.483 (0). -0.490 (0). -0.27 (0). 1. -0.114 (0). 0.0633 -0.453 (0) (0). -0.155 (0). -0.147 (0). 0.186 (0). -0.334 (0). -0.572 (0). 0.640 (0). 0.613 (0). 0.597 (0). -0.388 (0). 0.133 (0). 0.137 (0). 0.0123 0.0383 -0.362 (0) (0) (0). -0.100 (0). 0.132 (0). 0.146 (0). 0.174 (0). -0.261 0.112 (0) (0). -0.016 (0). Crisis. 1 1. 41.

(45) TABLE 4 Determinants of Banks’ Profitability Measured by ROA. The table provides information on the factors driving banks’ profitability. Column (1) presents a basic regression of ROA against the Herfindahl index. Column (2) controls for bank characteristics, while column (3) controls for country-specific characteristics. Column (4) uses both bank- and country-specific characteristics. Using available data described in the data and descriptive statistics section, I estimate the following model: , = 56 + 52Her?indahl, + 7 + 8 + 9, Where ROAi,t is the level of Return on Assets in percents. The Herfindahl Index measures concentration and is defined in Table 1 of the Appendix. β1 is the regression coefficient of interest and shows the relation between the level of competition and performance of banks. It shows the extent to which an increase in the level of concentration affects bank’s profitability measured by ROA. η is a vector of country-specific characteristics and ω a vector consisting of bank-specific characteristics. I use clustered standard errors at country-year level..

(46) VARIABLES Constant Herfindahl Index. (1) Basic. (2) Bank Characteristics. (3) Country Characteristics. (4) Bank and Country Characteristics. 0.6150*** (0.1870). 1.8850*** (0.115). -2.0860*** (0.7570). 2.4860*** (0.5930). 0.6180 (0.9630). 0.181 (0.530). 0.8110 (0.7260). -1.7700*** (0.5260). Ln(Total Assets). 0.00217 (0.01350). 0.0226* (0.0121). Capital Ratio. 0.0264*** (0.0041). 0.0228*** (0.0050). Cost to Income Ratio. -0.0214*** (0.0006). -0.0209*** (0.0005). Liquidity Ratio. 0.00137 (0.0008). -0.0007 (0.0009). Loan Loss Provisions Ratio. -0.0212*** (0.00121). -0.0212*** (0.0012). Fixed Assets to Total Assets Ratio. -0.0148** (0.0063). -0.0227*** (0.0068). Net Loans to Total Assets Ratio. 0.0043*** (0.0008). 0.0037*** (0.0008) 0.0411*** (0.0136). 0.00315 (0.00581). 0.0231* (0.0129). 0.0293*** (0.0073). 0.1680*** (0.0586). 0.0515 (0.0527). Ln(Stocks Traded). -0.0566* (0.0341). -0.0391 (0.0342). Credit Depth of Information Index. -0.0312 (0.0224). -0.0681*** (0.0180). 0.0622*** (0.0142). 0.0167*** (0.0049). Economic Freedom Index. -0.0079 (0.0074). -0.0084** (0.0033). Crisis. -0.129* (0.0697). -0.0520* (0.0304). GDP/capita growth Inflation rate Ln(Market Capitalization). Interest Rate. Observations. 194,408. 183,643. 140,698. 133,125. Number of Banks. 21,240. 20,713. 19,805. 19,323. 43.

(47) TABLE 5 Determinants of Banks’ Profitability: Extensions (Legal Origin). The table provides information on the factors driving banks’ profitability by different legal origins of the countries used in the sample. Column (1) presents the determinants of profitability measured by ROA for countries with common law. The second column describes the same thing for Scandinavian Law countries, while columns (3) and (4) extend the results for French Civil Law and German Civil Law countries, respectively. Using available data described in the data and descriptive statistics section, I estimate the following model: , = 56 + 52Her?indahl, + 7 + 8 + 9, Where ROAi,t is the level of return of assets in percents. The Herfindahl Index measures concentration and is defined in Table 1 of the Appendix. β1 is the regression coefficient of interest and shows the relation between the level of competition and performance of banks. It shows the extent to which an increase in the level of concentration affects bank’s profitability measured by ROA. η is a vector of country-specific characteristics and ω a vector consisting of bank-specific characteristics. I use clustered standard errors at country-year level.. 44.

(48) (1) English Common Law. (2) Scandinavian Law. (3) French Civil Law. (4) German Civil Law. Constant. 5.0160*** (1.0640). -23.3500*** (6.5320). 1.7880** (0.7030). 2.4650*** (0.9500). Herfindahl Index. -3.1390** (1.3220). 4.460** (2.150). -0.676 (0.597). -0.5870 (0.6310). Ln(Total Assets). 0.0703*** (0.0114). 0.0586* (0.0306). -0.0206 (0.0200). 0.0289*** (0.0112). 0.0127* (0.0069). 0.0628*** (0.0113). 0.0234*** (0.0029). 0.0411*** (0.0067). Cost to Income Ratio. -0.0198*** (0.0004). -0.0161*** (0.0027). -0.0272*** (0.0018). -0.0118*** (0.0012). Liquidity Ratio. -0.0053*** (0.0011). -0.0029 (0.0021). 0.0018** (0.0009). 0.0023* (0.0013). Loans Loss Provisions Ratio. -0.0254*** (0.0007). -0.0230*** (0.0023). -0.0187*** (0.0009). -0.0116*** (0.0014). Fixed Assets to Total Assets. -0.0435*** (0.0072). 0.0964*** (0.0349). -0.0139 (0.0113). 0.0299** (0.0144). Net Loans to Total Assets. 0.0030** (0.0012). -0.0032 (0.0022). 0.0053*** (0.0013). 0.0007 (0.0013). GDP/capita Growth. 0.0013 (0.0058). -0.0367*** (0.0117). 0.0200*** (0.0069). 0.0006 (0.0096). Inflation. -0.0050 (0.0135). 0.0150 (0.0168). 0.0277*** (0.0065). 0.0251 (0.0187). Ln(Market Capitalization). -0.0262 (0.0709). 0.2020 (0.1910). 0.0737* (0.0410). -0.0501 (0.0638). Ln(Stocks Traded). -0.0524 (0.0463). 0.1170 (0.1970). -0.0488* (0.0250). 0.0393 (0.0412). Credit Information Depth Index. -0.0361 (0.0361). 3.6650*** (0.7940). -0.0268* (0.0143). -0.1780*** (0.0467). 0.0148*** (0.0057). -0.1550** (0.0654). 0.0276*** (0.0064). 0.0173 (0.0256). Economic Freedom Index. -0.0082 (0.0066). 0.0127 (0.0191). -0.0022 (0.0045). -0.0060 (0.0057). Crisis. -0.0267 (0.0207). -0.7070*** (0.1460). 0.0181 (0.0432). -0.1280** (0.0526). Observations. 86,267. 2,560. 19,113. 25,185. Number of Banks. 12,322. 376. 3,309. 3,316. VARIABLES. Capital Ratio. Interest Rate. 45.

(49) TABLE 6 Determinants of Banks’ Profitability: Extensions (Types of Banks). The following table presents the relation between profitability of banks (measured by ROA) and concentration (measured by Herfindahl Index) in the banking sector for four subsamples of banks. The criterion for dividing the main sample into four smaller subsamples is related to the type of banks. As reminded before, the banks used in this analysis are commercial banks, savings banks, cooperative banks and bank holding and holding companies. Each column provides the same results as in column (4) of table 4, but for the subsamples reminded before. The coefficients for the regression between competition and profitability are shown for commercial banks, savings banks, cooperative banks and bank holding and holding companies in column (1), (2), (3) and (4), respectively.. Using available data described in the data and descriptive statistics section, I estimate the following model: , = 56 + 52Her?indahl, + 7 + 8 + 9, Where ROAi,t is the level of return of assets in percents. The Herfindahl Index measures concentration and is defined in Table 1 of the Appendix. β1 is the regression coefficient of interest and shows the relation between the level of competition and performance of banks. It shows the extent to which an increase in the level of concentration affects bank’s profitability measured by ROA. η is a vector of country-specific characteristics and ω a vector consisting of bank-specific characteristics. I use clustered standard errors at country-year level.. 46.

(50) (1) Commercial Banks. (2) Cooperative Banks. (3) Savings Banks. (4) Bank Holding & Holding Companies. 2.1710*** (0.5990). 0.0524 (1.253). 4.0950*** (1.2100). 1.6030 (1.2930). Herfindahl Index. 0.1840 (0.5070). -0.1340 (0.9470). -2.4880*** (0.8090). 1.3340* (0.7330). Ln(Total Assets). 0.0192 (0.0163). 0.0391*** (0.0102). 0.0459* (0.0261). 0.0073 (0.0091). Capital Ratio. 0.0169*** (0.0051). 0.0443*** (0.0083). 0.0416*** (0.0068). 0.0458*** (0.0066). Cost to Income Ratio. -0.0215*** (0.0009). -0.00882*** (0.00103). -0.0166*** (0.0011). -0.0261*** (0.0012). Liquidity Ratio. -0.0008 (0.0009). -0.0002 (0.0010). -0.0049** (0.0021). 0.0012 (0.0015). Loan Loss Provisions Ratio. -0.0231*** (0.0009). -0.0113*** (0.0013). -0.0178*** (0.0019). -0.0234*** (0.0011). Fixed Assets to Total Assets Ratio. -0.0328*** (0.0066). 0.0076 (0.0115). -0.0144 (0.0163). 0.0147 (0.0171). Net Loans to Total Assets Ratio. 0.0050*** (0.0010). 0.0027*** (0.0010). 0.0020* (0.0012). 0.0010 (0.0010). GDP/capita Growth. 0.0096 (0.0073). -8.79e-05 (0.0077). -0.00853 (0.0103). -0.0089 (0.0072). 0.0267*** (0.0064). 0.0161 (0.0189). -0.0224 (0.0144). 0.0107 (0.0142). Ln(Market Capitalization). 0.0337 (0.0481). 0.0256 (0.0605). 0.0861 (0.1220). 0.1240 (0.0955). Ln(Stocks Traded). -0.0210 (0.0304). 0.0065 (0.0407). -0.1400* (0.0850). -0.0854 (0.0551). Credit Information Depth Index. -0.0468*** (0.0166). 0.0576 (0.0463). -0.1480** (0.0693). 0.0216 (0.0513). Interest Rate. 0.0162*** (0.00437). 0.0436** (0.0179). 0.00873 (0.0202). 0.0158** (0.0073). Economic Freedom Index. -0.0056 (0.0036). -0.0146** (0.0065). -0.0044 (0.0074). -0.0082 (0.0070). Crisis. -0.0010 (0.0213). -0.0949** (0.0480). -0.2160*** (0.0709). -0.0378 (0.0243). Observations. 84,319. 18,976. 16,457. 13,373. Number of Banks. 12,055. 2,478. 2,233. 2,557. VARIABLES Constant. Inflation. 47.

(51) TABLE 7 Determinants of Banks’ Profitability: Extensions (Foreign and Public Banks) The table provides information on the factors driving banks’ profitability taking into account the type of ownership (foreign/domestic, public/private). Columns (1), (2) and (3) show the relation between profitability and concentration for foreign and domestic banks when ROA, ROE and Net Interest Margin, respectively, are used as proxies for profitability. Column (4) exhibits the influence of public ownership when profitability of banks is measured as Net Interest Margin. Using available data described in the data and descriptive statistics section, I estimate the following model: 4, = 56 + 52 FourBankConcRatio, + 7 + 8 + 9, Where 4, is a vector consisting of several measures of proftability The Four Bank Concentration Ratio measures concentration and is defined in Table 1 of the Appendix. β1 is the regression coefficient of interest and shows the relation between the level of concentration and performance of banks. It shows the extent to which an increase in the level of concentration affects bank’s profitability. η is a vector of country-specific characteristics and ω a vector consisting of bank-specific characteristics. I use clustered standard errors at country-year level.. 48.

(52) VARIABLES. (1) (2) (3) Foreign vs Domestic Foreign vs Domestic Foreign vs Domestic Banks - ROA Banks - ROE Banks - Net Interest Margin. (4) Private vs Public Banks - Net Interest Margin. Constant. 2.6720*** (0.5010). 24.4100*** (4.3980). 8.8220*** (1.1090). 8.9780*** (1.3230). Four Bank Concentration Ratio. -0.0039** (0.0016). -0.0374*** (0.0124). -0.0119*** (0.0034). -0.0100*** (0.0039). Foreign. -0.1690*** (0.0302). -1.4430*** (0.2470). -0.0522 (0.0503) -0.5050*** (0.1940). Public 0.0114 (0.0118). 0.3200*** (0.1050). -0.2640*** (0.0294). -0.2260*** (0.0325). Capital Ratio. 0.0347*** (0.0025). -0.0162 (0.0242). 0.0379*** (0.00480). 0.0400*** (0.0064). Cost to Income Ratio. -0.0245*** (0.0009). -0.2230*** (0.00865). -0.0109*** (0.0009). -0.0108*** (0.0010). Liquidity Ratio. 0.0001 (0.0008). 0.0112 (0.0075). -0.0008 (0.0012). -0.0006 (0.0016). Loan Loss Provisions Ratio. -0.0217*** (0.0008). -0.2180*** (0.0111). -0.0015*** (0.0005). -0.0015*** (0.0006). Fixed Assets to Total Assets Ratio. -0.0226* (0.0127). -0.0524 (0.0801). 0.1070*** (0.0155). 0.1060*** (0.0175). Net Loans to Total Assets Ratio. 0.0024** (0.0011). 0.0209** (0.0098). 0.0186*** (0.0020). 0.0212*** (0.0021). GDP/capita Growth. 0.0177*** (0.0059). 0.1560** (0.0625). -0.0032 (0.0117). -0.0085 (0.0136). Inflation Rate. 0.0148*** (0.0056). 0.0751 (0.0463). 0.0323*** (0.0088). 0.0343*** (0.011). Ln(Market Capitalization). 0.0467 (0.0368). 0.4990 (0.3090). 0.1490** (0.0675). 0.1300 (0.0819). Ln(Stocks Traded). -0.0494** (0.0242). -0.4700** (0.2010). -0.1520*** (0.0402). -0.1360*** (0.0493). Credit Depth of Information Index. -0.0372*** (0.0139). -0.3740*** (0.1430). 0.0014 (0.0348). -0.0004 (0.0426). Interest Rate. 0.0186*** (0.0051). 0.2030*** (0.0409). 0.0535*** (0.0142). 0.0426*** (0.0153). Economic Freedom Index. -0.0025 (0.0033). -0.0010 (0.0253). -0.0548*** (0.0084). -0.0607*** (0.0106). Crisis. -0.0204 (0.0279). -1.2440*** (0.3060). 0.1960** (0.0836). 0.1660* (0.0899). Observations. 33,133. 33,133. 33,133. 24,820. Number of Banks. 4,615. 4,615. 4,615. 3,296. Ln(Total Assets). 49.

(53) TABLE 8 Determinants of Banks’ Profitability: Extensions (Growth in Total Assets larger than 50%) The table provides information on the factors driving banks’ profitability. Column (1) extends the results of column (4) in Table 4 for the subsample consisting of bank-year observations with a growth in total assets larger than 50%, while column (2) applies the same criterion for bank-year observations with a growth in total assets smaller than 50%. Using available data described in the data and descriptive statistics section, I estimate the following model: , = 56 + 52Her?indahl, 7 + 8 + 9, Where ROAi,t is the level of Return on Assets in percents. The Herfindahl Index measures concentration and is defined in Table 1 of the Appendix. β1 is the regression coefficient of interest and shows the relation between the level of competition and performance of banks. It shows the extent to which an increase in the level of concentration affects bank’s profitability measured by ROA. η is a vector of country-specific characteristics and ω a vector consisting of bank-specific characteristics. I use clustered standard errors at country-year level.. 50.

(54) (1) Growth in Total Assets larger than 50%. (2) Growth in Total Assets smaller than 50%. Constant. 1.5080** (0.6900). 2.2820*** (0.6270). Herfindahl Index. 0.9140* (0.5120). -2.1140*** (0.5370). Ln(Total Assets). -0.0222 (0.0160). 0.0416*** (0.0110). Capital Ratio. 0.0047 (0.0053). 0.0396*** (0.0030). Cost to Income Ratio. -0.0274*** (0.000973). -0.0190*** (0.0006). Liquidity Ratio. 0.0010 (0.0015). -0.0005 (0.0010). Loan Loss Provisions Ratio. -0.0121*** (0.0010). -0.0221*** (0.0012). Fixed Assets to Total Assets. -0.0014 (0.0088). -0.0182*** (0.0064). Net Loans to Total Assets. 0.0020 (0.0016). 0.0038*** (0.0010). GDP/capita Growth. -0.0007 (0.0136). 0.0025 (0.0054). Inflation. 0.0305*** (0.0110). 0.0219*** (0.0074). Ln(Market Capitalization). 0.1670*** (0.0547). 0.0273 (0.0552). Ln(Stocks Traded) Credit Depth of Information Index. -0.1140*** (0.0343). -0.0259 (0.0360). 0.0112 (0.0196). -0.0916*** (0.0181). Interest Rate. 0.0252*** (0.0095). 0.0189*** (0.0053). Economic Freedom Index. -0.0090** (0.0045). -0.0049 (0.0032). 0.0401 (0.0849). -0.0703** (0.0297). Observations. 9,769. 123,356. Number of Banks. 6,506. 19,161. VARIABLES. Crisis. 51.

No results found