Does Bank Concentration A ect Bank Performance?
Chung-Hua Shen
Department of Money and Banking
National Chengchi University
Mucha, Taipei, 116
Taiwan ROC
TEL.: (02) 29393091{81020
FAX: (02) 29398004
email: [email protected]
May 27, 2003
1 Introduction
The e ect of bank concentration on bank performance and bank lending has recently attracted a lot of attention. Bank concentration is important because it may a ect bank competition, in particular on local markets for retail banking services. 1 Also,
the soundness and stability of the nancial sector may in various ways be in uenced by the degree of competition and concentration.2 Increased concentration may result from
mergers and acquisitions, which again will change the banking landscape radically, then a ects the capacity.3
Opponents of bank concentration holds that increasing concentration in some seg-ments of the banking market may eventually result in an undesirable exercise of market power by banks. They claim that the competition could promote eciency in the banking sector. The bene ts of competition for allocative eciency include the selection e ect in favor of ecient rms and the reduction of slack, or X eciency. Increased concentra-tion fosters collusion and excess pro ts for the nancial instituconcentra-tions (Gual and Neven, 1992, Bikker and Groeneveld, 1998).
Advocates of bank concentration argue that banks are fragile and rather special in-stitutions. The tense competition in the nancial sector may increase the fragility of banking since it creates the cut-throat competition between many small banks. The previous rejected marginal customers could obtain loan in a competitive environment. Matutes and Vives (1996), for example, claim that more competition can increase the
1See Bikker and Haaf (2001, 2002), Hondroyiannis, et al. (1999), Koskela and Stenbacka (2000) for
the studies of the relationship between bank competition and concentration.
2See Demirgu c-Kunt and Levine (2001) for the study of concentratoin and nancial stability. Also,
see Cetorelli's (2001) discussion for the bad or good of bank competition.
probability of failure through runs. The rst implication re ects the wave of mergers in global nancial market. It is generally believed that bank concentration can High bank concentration is bene cial for bank performance since it creates franchise e ect, making bank more risk averse.
While there are opposing views about the concentration, higher concentration is typ-ically companied with high returns because of monopolistic power.4 For example, Jeon
and Miller (2001), using U.S. state-level data, nd the positive relationship between bank concentration and bank performance. Demirgu c-Kunt and Levine (2001), however, using the averaged data 1990-1997 of roughly 60 cross-country to show that the bank concentration has neither negative impacts on slower economic growth, less competi-tive industrial sectors and more poorly functioning nancial system, nor any impact on greater bank fragility. Levine (2000) also use cross-country data to show that the similar results. Though this is not what the authors' original meanings, but in fact both of these two papers demonstrate that the concentration has no signi cant impact on bank performance and nancial stability. Thus, whether or not the concentration can increase the bank performance is controversial.
This paper re-evaluate the issue of bank concentration on bank performance. the relations between bank concentration and bank performance. We use 52 countries from 1993-2000 to estimate the model. We are only concerned with the correlation between banking concentration and banking performance.
The issue is also important in other discipline. For example, the merger wave during end of 1999 is to increase the bank concentration. This may in turn a ect the competition
in retail banking services. Berger et al. (1995) documented that large banks tend to lend to medium and large business borrowers, while small banks tend to specialize in lending to small businesses. This evidence would seem to support the fear that ongoing consolidation process in the U.S. and in Europe will be detrimental to small business.
Bikker and Haaf (2002) examine competition measures using the Panzar-Rosse model. They nd competition is weaker in local markets and stronger in international markets. The estimation of the large panel data often su ers the heterogeneity problem in both intercept and slopes. The heterogeneity in the intercept is often removed by using xed or random e ects modeling approach. The removal of the heterogeneity in slope, however, is often ignored in the literature. Haque, Pesaran and Sharma (2000) showed that neglected heterogeneity in slope distort the estimation results. This paper adopts two approaches to overcome this problem. The rst approach explicitly models the coecient of concentration ratio is a function of other exogeneous variables. The second approach adopts Swamy's (1970) random coecient model (RCM), which speci es the all banks' slopes are generated from a normal distribution.
2 Concentration, Competitiveness and Bank
Perfor-mance
2.1 Concentration and Competitiveness
Researchers are interested in concentration partly because it a ects the competitiveness. Hence, most studies focus on the relationship between these two Cs.
While the two are highly correlated, it is well known that less concentration does not imply competition. Vives (2000), for example, claims that the entry policies{not
actual entry{are important. A competitive nancial system does not necessarily require a large number of nancial institutions; a concentrated system can be competitive if it is contestable (that is, if competition is open). The contestability theory stresses that a concentrated banking industry can behave competitively if the hurdles to be surmounted by new entrants to the market are low (Baumol, 1982).5 Financial systems in many
European countries are considered to be quite competitive even though they have a limited number of banks. Sha er (1993), for example, nds no evidence of monopolistic behavior even though the ve largest banks in the country account for more than 80 percent of all banking assets.
Banking systems are the same. In many developing countries, even though they are concentrated, they are also competitive since competition from other nancial institutions and other forms of nancial intermediation is strong. Banks face credit threat from new entrants that are permitted to o er rival services in the market.
Vives (1991)
Bikker and Haaf (2002) divides the measurement of competition into structural and non-structural approaches, where the former can be further divided into nonformal and formal approaches. The nonformal approach includes structure-conduct-performance (SCP) paradigm and eciency hypothesis, whereas the formal approach includes the measures of concentration ration indices. The non-structural approach contains the Iwata model, the Bresnahan model and the Panzar and Rosse (P-R) model
They use the P-R model to test the market competition condition rst.6 They provide
5This theory asserts that the threat of potential entry forces banks with large market shares to price
their products competitively under certain conditions. A contestable market has no entry barriers, neither economic nor legal.
strong evidence that the banking markets in the industrial world are characterized mostly by monopolistic competition, but perfect competition exists for a few countries. Then, they nd that the relationship between market structure (i.e., concentration) and conduct (i.e., competition) display strong links, especially when market structures are proxied by the k bank concentration indices.
2.2 Concentration and Banking Performance
The pro t-structure relationship has received considerable attention in the industrial organization and banking literature. Typically, a positive correlation emerges between pro tability and concentration or market shares. Jeon and Miller (2001) de ne the state-level in the U.S. as a region to study the relation between bank concentration and bank performance. They nd strong support for a positive correlation.
Demirgu c-Kunt and Levine (2001) ask a broad question. They study the relationship between bank concentration and the nancial sector policies (such as deposit insurance generosity, regulatory restrictions on bank activities, and the tax system), industrial competition (such as the degree of market domination, the e ectiveness of anti-trust laws), and the institutional environment (including measures of integrity, rule of law, tax compliance, the protection of minority shareholders, and the transparency and accuracy of the accounting system).7 They employ the averaged data over the 1990{1997 period of
bank concentration and the sample countries are around 70 depending on the explanatory variables used. In contrast to Jeon and Miller's (2001) ndings, their results are un-favorable to the above links.
7Cetorelli (2001) asks a similar but slightly di erent question. He evaluates and show that there is a
The direct test of the correlation between bank concentration and bank performance may be subject to Berger's (1995) criticism. Berger (1995) claims that most prior tests of the market-power theories produce suspect ndings, since they as a rule do not control for the eciency-structure theories.8 To implement this joint test, he uses individual bank
balance sheet and income statement data. His evidence does not support the structure-conduct-performance and scale-eciency hypothesis.
3 Regression Model
To develop a concentration measure, we need to determine the geographic spread of the relevant market. For example, the conventional wisdom argues that policy makers should consider whether bank consolidation leads to excessive market power in retail, rather than wholesale markets.
The regression model is
bankIncomeit = 0 + 1CRit+ 2Bank Control Variablesit 3MacroControlV ariableit+ 4Institutional Variablesi
(1) where CR is the concentration ratio, proxied by CR3, CR4, CR5 and HHI, respectively, Bank Control Variables include Equity, MAG, Liquid and Credit, where Equity is the equity/asset ratio, MAG is the proxy for management eciency and is de ned as the earning asset/asset ratio, Liquid is the liquid ratio, which is de ned as the liquidity/asset ratio, Credit is the loan to the private sector by banks. Macro Control Variables includ
8Berger (1995) presents two competition theories to explain the positive correlations between
concen-tration (structure) and bank performance (pro t){market power and ecient structure theories. The market-power theory includes two hypotheses{the traditional structure-conduct-performance (SCP) and the relative-market power (RMP) hypotheses. The SCP argues that more concentrated markets lead to higher loan rates and lower deposit rates because of lessened competition whereas the RMP hypothesis argues that only large banks with some brand identi cation can in uence pricing and raising pro ts.
The ecient-structure theory also includes two hypotheses{the X-eciency and scale-eciency hy-potheses.
GDPper, In a and Rate, where GDPper is the GDP per capita, In a is the in ation rate and Rate is the one-month time deposit rate. The Institutional Variables contains four dummy variables, which are Bank S, Bank I, Bank R and Bank N. These four dummy variables are taken from Barth et al. (1998), which are the restrictions of bank to engage in security, insurance, real estate and non-bank activities. See Table 1 for the operational de nition of variables and their sources.
We next assume that the e ect of the concentration on pro t is in uneced by the following eight institutional variables. These institutional variable could proxy the gov-ernance with higher number denotes the better govgov-ernance. We hypothesize that the better governance negatively a ect the concentration e ect.
1 = 0+ 1zt (2)
where
zt = LawHabitjt; LawRiskjt; Corruptjt; Con scationjt; Contractjt; Accountjt; GovE jt; BankSupejt
(3)
4 Data Sources and Basic Statistics
4.1 Sources of Data
Table 1 reports the source of data. Bank control variables are taken from BankScope and Macro control variables are taken from IFS. There are three sources of institutional vari-ables. Fisrt is from LLSV, including Good Government Index. The restrctions on bank activities, including Bank S Bank I Bank R and Bank N. The Banksupe and LawHaibt are from Barth et al. (2001).
4.2 Basic Statistics
Table 2 reports the average of concentration ratios and HHI based on the period 1993-2000. The CR3, CR4 and CR5 denote the concentration ratio of the third, fourth and fth largest banks. The rst four columns are calculated by their asset sizes and the later four columns are calculated based on the deposits, respectively. The rank of con-centration ratio of countries change little when we use di erent CRs. Employing CR3 as the criterion, the highest ratio falls on Jordan (85.82), followed by Romina (74.02) and Ecuado (64.67), and the lowest ratio is Japan (13.6), followed by Germany (15.04) and U.S. (16.45). Employing the HHI as the criterion, the highest ratio is Ergudo (1890), followed by South Africa (1863), Finland (1602), Hong Kong (1503) and China (1493). The lowest ratio is US (163.02), followed by Germany (191.65), Jordon (202.64), Japan (249.72), Italy (249.38). Based on the above nding, the concentration is unlikely to be a ected by the bank-based or market-based economies since the lowest concentration include both USA and Germany and Taiwan.
Table 3's rst three columns report the pro tablity of banking industry of 52 countries from 1993-2000. The highest ROA falls on Zimbabwe (3.009), followed by Turkey (2.266) and Nigeria (2.06). The lowest ROA is typically those su ered from the Asian nancial crisis, such as Indonesia ({6.143), Thailand ({0.794), South Korea ({0.484) and Japan ({0.713). The later four columns of Table 3 reports the restrictions on banks' activities. Table 4 reports the institutional factors, the z, in equation 2. The rst column presents the LawHabit, where Hong Kong, Netherlands, New Zealand, Norway, Sweden, Swiss, UK and USA get the full scores and the lowest score falls on Indonesia (2.0), Thailand (3.25) Turkey (4) and the Philippines (4.7). The second column is the LawRisk,
where Australia, Austria, Finland, Norway, Sweden, Swiss and USA get the full scores and the lowest countries are Sri Lanka (1.9), Colombia (2.08), Peru (2.5), the Philippine (2.73).
5 Regression Results
Table 5 is our benchmark model without considering zj. Also, the bank income variable
is proxied by ROE and the CR is attempted by CR3, CR4, CR5 and HHI calculated by the assets and deposits. Absolute t values are reported in parentheses based on White's heteroscedasticity consistent estimator. As reported in the table, the interested coecients of CR are overwhelmingly signi cantly positive regardless of the proxies. The estimated coecients are around 0.15~0.21 when CR3, CR4 and CR5 are employed and are 0.001 when HHI is taken. Accordingly, the higher concentration ratio is indeed positively related to the higher pro t.
Bank and macro control variables also show th expected sign. With regarding to the bank control variables, Equity are also found to be signi cantly di erent from zero but MAG and Liquid are insigni cant. Turning to the case of Macro control variables, GDP per captia, the in ation rate and interest rate all show the expected signs and are signi cant to some extend. That is, banks' pro ts are higher in the richer countries but smaller in the poorer countries. The pro ts are positively related to the interest rate but negatively related to the in ation rate.
Estimated results on restrictions on bank's activities on securities, insurances, real estate and non-bank industries are also interesting. First, diversi cation e ect may exist if banks are allowed to engage in insurance and non-bank activities. This is because
the estimated coecients are negative on Bank I and Bank N suggesting that the tighter restrictions on the insurance and non-bank activities hurt bank's pro t. Next, restrictions on real estate increases banks pro ts. We conjecture this is because the asset price of real estate uctuates recently and thus the restrictions on this activity insulate bank from losing money. Last, the diversi cation e ect does not exist if banks are engaged in security activity as the coecient on Bank S is insigni cant.
Table 6 reports the estimated results using ROE and NIM as the dependent variable. Also, for brevity, the CR is calculated only based on deposits. Results change slightly when ROE is used. First, coecients of bank control variables, such as MAG, Liquid and Credit, become signi cant. Next, coecients on Macro control variables, such as GDP per capita, In a and Rate become insigni cant. Similar results are obtained when ROE is replaced by NIM.
Table 7 reports the estimated results when the interactive term, CR z is added in alternatively, where z is de ned as in (3). The signs of CR z are overwhelmingly nega-tive, suggesting that the better governance will decrease the the impact of concentration ratio on the pro t. Among them, the LawRisk, Corruption, Con scation and Accounting Standard are signi cantly di erent from zero. Thus, a country with a low law risk, low corruption, low con scation and a consistent accounting standard, its concentration will have less e ect on pro t.
Table 8 takes z into account in a di erent speci cation. First, our z includes four restrictions on banking activity, while at the same time they are removed in the bench-mark equation. Results show that restrictions on bank to engage the activity in security and non-banking activity will decrease the concentration e ect but restriction on
bank-ing to engage the activity in real estate will increase the concentration e ect. Estimated coecients become mostly insigni cantly di erent from zero when the same variables are also considered in the benchmark equation. The last column considers all z variable at the same time. It is found that the impact of corruption on the concentration e ect is robust to this new speci cation.
6 Temporary Conclusion
This paper nds that the higher the concentration, the higher the pro t is. Also, this concentration e ect is negatively a ected by Law Risk, Corruption, Con scation and accounting standard.
References
Acharya, V. V., I. Hasan and A. Saunders (2001), Should Banks be Diversi ed? Ev-idence from Individual Bank Loan Portfolio, in Financial Market Behavior and Apporpriate Regulation over the Business Cycle, 309-334, The 38th Annual Con-ference on Bank Structure and Competition, Federal Reserve Bank of Chicago. Bikker, J. A. and J. M. Groeneveld (1998), Competition and concentration in the EU
banking industry, working paper, De Netherlandish Bank, NV.
Bikker, J. A. and K. Haaf (2001), Measures of competition and concentration in the banking industry: a review of the literature, working paper, De Netherlandish Bank, NV
Bikker, J. A. and K. Haaf (2002), Competition, concentration and their relationship: an empirical analysis of the baning industry, Journal of Banking and Finance, 2002, forthcoming
Cetorelli, N. (2001a) Does Bank Concentration Lead to Concentration in Industrial Sectors? paper presented in 38th Annual Conference on Bank Structure and Com-petition, Fed of Chicago.
Cetorelli, N. (2001b) Competition among banks: good or bad? Economic Perspective, second quarter, Fed of Chicago.
Demirgu c-Kunt, A. and R. Levine (2001), Bank concentration: cross-country evidence, world bank, working paper
Jeon, Y. and S. M. Miller (2001), Bank concentration and performance, working paper, Central Michigan University
Haque, N. U., M. H. Pesaran and S. Sharma (2000), Neglected Heterogeneity and Dy-namics in Cross-Country Savings Regressions, ed. by J. Krishnakumar and E. Ronchetti of Panel Data Econometrics: Future Direction{Papers in honour of pro-fessor Pietro Balestra, 53-82
Hondroyiannis, G, S. Lolos and E. Papapetrou (1999), Assessing competitive conditions in the Greek banking system, Journal of International Financial Markets Institu-tions and Money, 9, 377-391
Levine, R. (2000), Bank concentration: Chile and international comparisons, Working paper 62, Central Bank of Chile
Sha er, S. (1993), Market conduct and excess capacity in banking: a cross-country cemparison, Working paper: 93-28, Fed of Philadelphia, November
Stiroh, K and J. P. Poole, (2000), Explaining the rising concentration of banking assets in the 1990s, Current Issues, Fed of New York, August,
Table 1: Variable De nition and Sources Bank Return Variable and CR
ROAA Return on average pro t/averaged asset Bankscope ROAE Return on equity pro t/averaged equity Bankscope NIM Net interest margin (interest revenue-interest Bankscope
payment)/earning asset
CR concentration ratio sum of ve largest banks' Bankscope assets/Total banks' assets
Bank Control Variables
MG Macro Governance 0-50, sum of e . of judicial system, rule of law, corru-ption, risk of expropriation risk of contract repudiation
Equity Equity Ratio Equity/Assets Bankscope Mage Management Eciency Earning Assets/Total Assets Bankscope Liquidity Liquidity Ratio Liquidity/Total Assets Bankscope Credit Bank Credit claim to the private sector IFS
by banks
Bank S Restriction on Bank's 1: prohibit 2: restricted Barth et al. (2000) investment activities 3: limited, 4: permitted
Bank I Restriction on Bank's same as above Barth et al. (2000) insurance activities
Bank R Restriction on Bank's same as above Barth et al. (2000) real estate activities
Bank N Restriction on Bank's same as above Barth et al. (2000) non-bank activities
Macro Variables
GDP per GDP per Capita Real GDP/Population IFS In ation In ation rate (CPIt CPIt 1)/CPIt 1 IFS
Rate nominal interest rate paid to the demand deposit IFS rate
Other Institutional Variables
GoodGov Good Government 0-50, sum of e . of judicial LLSV (1998) Index system, rule of law,
corru-ption, risk of expropriation risk of contract repudiation
Super Ocial supervision 0-16, sum of prompt correc- Barth et al. (2001) power tive action, restructuring
power and declaring insolvency power
LawHabit Ocial supervision 0-16, sum of prompt correc- Barth et al. (2001) LawRisk Ocial supervision 0-16, sum of prompt correc- Barth et al. (2001) Corrupt Ocial supervision 0-16, sum of prompt correc- Barth et al. (2001) Con s Ocial supervision 0-16, sum of prompt correc- Barth et al. (2001)14
Table 2: Concentration Ratio: 52 countries, 1993-2000 Asset
Concentration Ratio: Asset Concentration Ratio: Deposit
Country CR3 CR4 CR5 HHI CR3 CR4 CR5 HHI
Argentina 29.92 35.70 40.76 566.65 33.17 39.02 44.06 638.34 Australia 33.80 42.59 50.83 731.72 36.78 46.20 55.02 836.80 Austria 33.03 40.36 45.82 596.86 34.49 42.43 48.55 644.88 Belgium 37.51 46.05 53.85 896.66 39.19 47.86 55.91 939.59 Brazil 32.10 38.00 43.32 583.92 38.23 44.95 49.65 658.03 Canada 50.61 64.74 76.66 1286.42 50.77 64.99 76.63 1297.50 Chile 27.26 34.28 40.99 590.57 26.76 33.63 40.30 592.31 China 56.65 69.94 79.62 1493.31 59.05 72.50 82.30 1551.65 Colombia 29.62 37.37 44.37 615.53 30.32 38.20 45.27 624.51 Denmark 42.82 53.04 61.76 898.88 54.70 65.16 71.88 1258.71 Ecuador 64.76 72.26 77.07 1890.05 63.94 71.35 76.18 1943.54 Egypt 52.69 60.60 65.23 1158.31 54.31 62.42 66.75 1149.89 Finland 60.05 70.88 79.21 1602.67 64.17 76.08 83.45 1750.63 France 20.28 25.94 31.30 326.66 21.63 27.48 32.82 346.74 Germany 15.04 18.62 22.02 191.65 15.75 19.21 22.35 187.29 Greece 47.02 54.84 61.79 1101.19 48.04 55.72 62.61 1130.39 Hong Kong 58.85 66.63 69.13 1503.28 57.73 66.49 69.03 1445.83 India 33.43 38.45 43.31 703.34 33.90 39.70 45.25 718.33 Indonesia 38.44 47.05 53.83 808.61 38.60 47.13 53.90 797.97 Ireland 51.09 63.41 68.57 1158.38 51.66 64.79 68.85 1174.54 Israel 48.48 61.06 69.69 1128.38 47.34 60.28 69.11 1106.62 Italy 17.31 21.44 25.28 249.72 18.15 22.54 26.54 270.76 Japan 13.60 17.31 20.87 202.64 15.09 19.21 22.80 225.40 Jordan 85.82 88.56 90.95 3375.21 85.81 88.62 91.21 3403.48 Kenya 36.04 46.65 53.43 727.06 36.54 47.02 54.58 759.07 South Korea 17.41 22.59 27.56 358.34 17.02 21.96 26.75 352.27 Rousanberg 16.66 21.93 27.00 301.13 16.66 21.70 26.59 297.64 Malaysia 34.29 40.51 45.96 653.26 34.30 40.68 46.19 652.37 Mexico 38.24 47.70 55.65 809.93 40.19 50.54 56.86 853.40 Netherlands 62.43 73.27 79.40 1616.32 65.35 76.39 80.64 1744.39 New Zealand 38.64 49.82 60.58 1031.90 40.17 51.11 61.13 1044.59 Nigeria 45.75 55.51 63.96 1101.91 46.55 56.55 64.98 1127.39 Norway 35.77 44.84 52.64 673.67 37.02 46.23 55.05 744.15 Pakistan 60.01 70.82 79.94 1461.98 60.60 71.52 80.83 1489.60 Panama 22.53 27.90 32.71 374.77 23.43 28.77 33.35 387.67 Peru 51.18 62.25 69.13 1213.30 51.77 62.06 69.40 1245.27 Philippines 32.80 40.73 47.22 646.76 35.84 44.19 50.55 714.37 Poland 41.30 48.13 53.09 785.69 42.79 50.13 54.78 832.43 Portugal 28.16 35.45 41.44 546.01 28.49 35.64 41.87 559.08 Romania 74.02 81.97 86.43 2354.47 75.34 82.30 86.57 2455.45 Singapore 41.07 50.82 59.82 969.23 40.54 50.48 59.72 971.73 South Africa 67.16 83.36 88.95 1863.02 71.61 88.24 93.97 2113.46 Spain 28.68 34.79 39.11 435.16 27.53 33.54 38.10 413.81 Sri Lanka 54.49 69.43 75.55 1383.44 54.17 70.95 77.64 1421.35 Sweden 30.64 39.26 47.02 671.50 35.52 45.68 55.35 909.58 Swiss 52.38 60.83 67.91 1237.18 52.08 62.01 68.55 1233.03 Taiwan 31.49 38.21 43.53 581.66 32.71 39.60 45.03 614.36 Thailand 41.69 50.64 58.31 910.82 41.69 50.90 58.73 915.35 Turkey 31.82 38.21 44.15 15642.30 31.77 38.48 44.43 659.95
Table 3: Concentration Ratio: 52 countries, 1993-2000 Asset
Country ROA ROE NIM Restriction on Bank
Secu- Insur- Real
Non-rity ance Estate Finac
Argentina 0.675 5.026 5.12 3 2 2 3 Australia 0.989 14.05 2.88 1 2 3 2 Austria 1.291 7.01 1.6 1 2 1 1 Belgium 0.404 12.44 1.48 2 2 3 3 Brazil 0.575 5.93 7.42 2 2 3 3 Canada 0.665 13.06 2.71 2 2 2 3 Chile 1.066 13.81 5.32 3 2 3 3 China 0.281 6.76 1.94 4 3 4 3 Colombia 0.49 2.04 8.14 2 2 2 4 Denmark 0.638 11.59 1.996 1 2 2 2 Ecuador -0.041 10.27 5.09 2 4 4 na Egypt 0.711 12.03 1.37 2 2 3 3 Finland 0.17 3.07 1.99 1 3 2 1 France 0.264 6.15 1.44 2 2 2 2 Germany 0.236 6.78 1.37 1 3 2 1 Greece 0.994 17.1 2.21 2 3 3 1 Hong Kong 1.666 19.26 1.596 1 2 2 3 India 0.56 8.79 3.07 2 4 4 2 Indonesia -6.143 8.22 0.85 2 4 4 4 Ireland 0.809 13.06 1.84 1 4 1 1 Israel 0.544 9.15 2.51 1 1 1 1 Italy 0.318 4.99 2.78 1 2 3 3 Japan -0.173 -5.04 1.26 3 4 3 3 Jordan 0.979 12.9 2.74 2 4 3 2 Kenya 1.468 14.47 4.49 2 4 3 1 South Korea -0.484 9.36 1.86 2 2 2 3 Rousanberg 0.52 15.55 0.79 1 2 1 2 Malaysia 0.923 12.9 3.11 2 2 3 2 Mexico 0.566 8.12 4.99 3 4 3 3 Netherlands 0.601 11.47 1.56 1 2 2 1 New Zealand 0.96 22.2 2.72 1 1 1 2 Nigeria 2.06 15.24 7.21 1 2 2 2 Norway 1.04 16.63 2.92 2 2 2 2 Pakistan -0.26 -10.65 2.84 2 4 3 1 Panama 0.933 13.41 1.86 1 2 3 2 Peru 1.021 11.19 6.89 2 2 2 2 Philippines 1.391 10.9 4.37 1 2 2 3 Poland 1.454 20.13 5.45 2 3 3 2 Portugal 0.714 12.2 2.89 1 2 3 2 Romania -0.328 -24.29 7.97 2 4 4 3 Singapore 1.049 9.46 2.11 2 2 2 3 South Africa 1.13 14.66 4.15 2 2 1 1 Spain 0.67 11.31 2.89 1 2 3 1 Sri Lanka 0.954 14.64 4.74 2 2 2 2 Sweden 0.549 13.08 1.88 4 2 3 3 Swiss 0.403 8.11 1.12 1 1 1 3 Taiwan 0.671 9.82 2.2 1 4 4 3 Thailand -0.794 -16.44 2.798 2 2 2 3 Turkey 2.266 26.26 11.38 3 2 4 3 UK 0.749 14.72 1.65 1 2 1 1 16
Table 4: Appendix: Governament Eciency
Law Law Corrup- Con s- Cont- Accou- Gove't Ocial
country Habit Risk tion cation tract nting Stand- Supervisor
ard Power Argentina 6 5.35 6.02 5.91 4.91 45 5.448 12 Australia 10 10 8.52 9.27 8.71 75 9.000 14 Austria 9.5 10 8.57 9.69 9.6 54 8.793 12 Belgium 9.5 10 8.82 9.63 9.48 61 8.922 11 Brazil 5.75 6.32 6.32 7.62 6.3 54 6.285 14 Canada 9.25 10 10 9.67 8.96 74 9.213 10 Chile 7.25 7.02 5.3 7.5 6.8 52 6.512 13 China na na na na na na na 8 Colombia 7.25 2.08 5.00 6.95 7.02 50 5.550 15 Denmark 10.0 10.0 10.0 9.67 9.31 62 9.197 9 Ecuador 6.25 6.67 5.18 6.57 5.18 na 5.970 7 Egypt 6.50 4.17 3.87 6.30 6.05 24 4.882 13 Finland 10.0 10.0 10.0 9.67 9.15 77 9.420 11 France 8.00 8.98 9.05 9.65 9.19 69 8.628 na Germany 9.00 9.23 8.93 9.90 9.77 62 8.838 10 Greece 7.00 6.18 7.27 7.12 6.62 55 6.615 na Hong Kong 10.0 8.22 8.52 8.29 8.82 69 8.458 na India 8 4.17 4.58 7.75 6.11 57 6.052 13 Indonesia 2.5 3.98 2.15 7.16 6.09 na 4.376 9 Ireland na na na na na na na 14 Israel 10 4.82 8.33 8.25 7.54 64 7.557 na Italy 66.7 8.33 6.13 9.35 9.17 62 7.655 12 Japan 10 8.98 8.52 9.67 9.69 65 8.893 13 Jordan 8.66 4.35 5.48 6.07 4.86 na 5.884 12 Kenya 5.75 5.42 4.82 5.98 5.66 na 5.526 9 South Korea 6 5.35 5.3 8.31 8.59 62 6.625 13 Rousanberg na na na na na na na 8 Malaysia 9 6.78 7.38 7.95 7.43 76 7.690 9 Mexico 6 5.35 4.77 7.29 6.55 60 5.993 11 Netherlands 10 10 10 9.98 9.35 64 9.288 3 New Zealand 10 10 10 9.69 9.29 70 9.330 10 Nigeria 7.25 2.73 3.03 5.33 4.36 59 4.767 4 Norway 10.0 10.0 10.0 9.88 9.71 74 9.498 na Pakistan 5.00 3.03 2.98 5.62 4.87 na 4.300 10 Panama na na na na na na na 9 Peru 6.75 2.5 4.7 5.54 4.68 38 4.662 11 Philippines 4.75 2.73 2.92 5.22 4.8 65 4.487 14 Poland na na na na na na na 6 Portugal 5.5 8.68 7.38 8.9 8.57 36 7.105 11 Romania na na na na na na na 13 Singapore 10 8.57 8.22 9.3 8.86 78 8.792 8 South Africa 6 4.42 8.92 6.88 7.27 70 6.748 na Spain 6.25 7.8 7.38 9.52 8.4 64 7.625 6 Sri Lanka 7 1.9 5 6.05 5.25 na 5.040 11 Sweden 10 10 10 9.4 9.58 83 9.547 13 Swiss 10 10 10 9.98 9.98 68 9.460 11 Taiwan na na na na na na na 8 Thailand 3.25 6.25 5.18 7.42 7.57 64 6.012 12 Turkey 4 5.18 5.18 7 5.95 51 5.402 15 UK 10 8.57 9.1 9.71 9.63 78 9.135 14 17
Table 5: Concetration Ratio and Bank Returns: Benchmark Model
Dependent Variable: ROE
Asset Deposit CR3 CR4 CR5 HHI CR3 CR4 CR5 HHI const. -6.577 -6.700 -6.675 -6.299 -6.171 -6.345 -6.366 -5.963 (2.34) (2.39) (2.38) (2.28) (2.20) (2.26) (2.26) (2.14) CR 0.021 0.019 0.019 0.001 0.016 0.015 0.015 0.001 (4.16) (4.30) (4.33) (4.33) (3.51) (3.70) (3.82) (3.91) Equity 0.473 0.474 0.476 0.475 0.472 0.474 0.475 0.474 (4.27) (4.30) (4.32) (4.36) (4.20) (4.23) (4.25) (4.26) MAG 0.033 0.033 0.031 0.034 0.031 0.031 0.030 0.032 (1.55) (1.52) (1.44) (1.60) (1.41) (1.42) (1.37) (1.45) Liquid -0.012 -0.010 -0.010 -0.017 -0.008 -0.007 -0.008 -0.013 (1.06) (0.95) (0.97) (1.48) (0.76) (0.69) (0.73) (1.13) Credit -0.002 -0.002 -0.002 -0.003 -0.001 -0.001 -0.001 -0.002 (0.89) (0.78) (0.66) (1.24) (0.61) (0.56) (0.49) (0.90) GDP Per 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 (3.23) (3.28) (3.29) (3.23) (3.07) (3.11) (3.14) (3.07) In a -0.002 -0.002 -0.002 -0.002 -0.002 -0.002 -0.002 -0.002 (1.93) (1.78) (1.85) (1.89) (1.69) (1.55) (1.61) (1.58) rate 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 (1.69) (1.55) (1.63) (1.66) (1.54) (1.40) (1.47) (1.46) Bank S 0.082 0.070 0.064 0.087 0.074 0.065 0.056 0.070 (0.74) (0.65) (0.60) (0.78) (0.66) (0.59) (0.51) (0.63) Bank I -0.241 -0.239 -0.223 -0.267 -0.229 -0.229 -0.213 -0.250 (2.48) (2.47) (2.34) (2.66) (2.33) (2.35) (2.20) (2.50) Bank R 0.710 0.733 0.739 0.724 0.683 0.703 0.709 0.697 (4.15) (4.19) (4.21) (4.20) (4.02) (4.06) (4.08) (4.08) Bank N -0.587 -0.586 -0.598 -0.598 -0.596 -0.595 -0.601 -0.595 (3.42) (3.43) (3.48) (3.53) (3.38) (3.40) (3.43) (3.43) R2 0.685 0.687 0.688 0.629 0.670 0.678 0.625 0.682
Table 6: Concetration Ratio and Bank Returns: Benchmark Model II
Concentration Ratio is based on Deposit
Dep. Var. = ROE Dep. Var. = NIM
CR3 CR4 CR5 HHI CR3 CR4 CR5 HHI const. 43.637 42.621 42.849 45.267 16.092 16.059 15.953 15.821 (2.38) (2.31) (2.33) (2.54) (6.22) (6.17) (6.09) (6.16) CR 0.181 0.160 0.152 0.005 -0.017 -0.013 -0.011 -0.000 (3.17) (3.08) (3.09) (3.30) (2.92) (2.48) (2.21) (2.97) Equity -0.062 -0.043 -0.023 0.003 0.251 0.251 0.251 0.250 (0.13) (0.09) (0.05) (0.01) (15.06) (15.00) (14.91) (15.68) MAG -0.341 -0.346 -0.359 -0.324 -0.133 -0.133 -0.132 -0.134 (2.06) (2.09) (2.18) (1.94) (4.60) (4.55) (4.49) (4.62) Liquid -0.289 -0.270 -0.266 -0.322 -0.013 -0.016 -0.018 -0.010 (3.15) (2.99) (2.95) (3.36) (1.10) (1.40) (1.52) (0.80) Credit -0.083 -0.080 -0.077 -0.089 -0.016 -0.016 -0.016 -0.015 (2.58) (2.52) (2.46) (2.65) (7.50) (7.53) (7.52) (7.31) GDP Per 0.000 0.000 0.000 0.000 -0.000 -0.000 -0.000 -0.000 (0.48) (0.52) (0.51) (0.46) (3.85) (3.78) (3.66) (3.72) In a -0.005 -0.003 -0.004 -0.004 0.003 0.003 0.003 0.003 (0.69) (0.46) (0.57) (0.59) (1.48) (1.40) (1.42) (1.45) rate 0.001 0.000 0.001 0.001 -0.000 -0.000 -0.000 -0.000 (0.38) (0.16) (0.29) (0.32) (0.21) (0.14) (0.19) (0.18) Bank S -0.544 -0.644 -0.687 -0.494 0.397 0.407 0.410 0.395 (0.69) (0.83) (0.88) (0.61) (3.36) (3.43) (3.45) (3.35) Bank I -1.567 -1.540 -1.409 -1.769 -0.273 -0.281 -0.294 -0.257 (1.96) (1.94) (1.78) (2.16) (2.20) (2.25) (2.35) (2.06) Bank R 2.956 3.122 3.150 3.044 0.117 0.118 0.127 0.116 (1.90) (1.94) (1.96) (1.95) (0.72) (0.72) (0.78) (0.71) Bank N -1.325 -1.340 -1.453 -1.501 -0.171 -0.169 -0.159 -0.161 (1.03) (1.04) (1.11) (1.14) (1.41) (1.39) (1.31) (1.32) R2 0.122 0.123 0.122 0.124 0.718 0.717 0.715 0.719
Table 7: Concetration Ratio and Bank Returns: Interative Modle I
Dep. Var. = ROE z is:
LawHab LawRisk Corrupt Con s Contract Account GoodGov Supervision
const. -8.927 -7.034 -7.381 -7.533 -7.192 2.237 -8.509 -6.942 (2.96) (2.42) (2.55) (2.58) (2.47) (1.47) (2.69) (2.43) CR 0.024 0.035 0.037 0.053 0.044 0.022 0.015 0.022 (3.93) (2.93) (3.47) (2.82) (2.37) (4.15) (2.58) (2.85) CR z -0.000 -0.002 -0.002 -0.004 -0.003 -0.000 0.000 0.000 (0.93) (1.64) (2.00) (1.95) (1.44) (2.40) (1.30) (0.04) Equity 0.517 0.481 0.483 0.477 0.477 0.143 0.485 0.476 (5.27) (4.46) (4.49) (4.47) (4.41) (4.40) (4.51) (4.34) MAG 0.046 0.037 0.040 0.044 0.041 -0.031 0.052 0.037 (1.87) (1.56) (1.72) (1.84) (1.74) (1.96) (1.94) (1.67) Liquid -0.008 -0.019 -0.012 -0.020 -0.017 -0.008 -0.001 -0.010 (0.61) (1.16) (0.95) (1.23) (1.04) (0.83) (0.04) (0.78) Credit -0.001 -0.001 -0.001 -0.001 -0.001 -0.003 -0.002 -0.003 (0.32) (0.37) (0.55) (0.28) (0.34) (1.38) (0.58) (1.03) GDP Per 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 (3.60) (3.34) (3.41) (3.58) (3.40) (1.65) (3.28) (3.16) In a 0.003 -0.003 -0.001 -0.001 -0.002 0.000 0.003 -0.002 (1.52) (2.28) (1.18) (1.06) (2.10) (0.17) (1.86) (1.54) rate -0.001 0.001 0.000 0.000 0.001 -0.000 -0.001 0.001 (1.80) (2.11) (0.93) (0.87) (1.86) (0.51) (2.10) (1.39) Bank S -0.023 0.034 0.041 0.073 0.031 0.086 0.037 0.095 (0.18) (0.31) (0.38) (0.66) (0.29) (1.31) (0.35) (0.82) Bank I -0.232 -0.369 -0.384 -0.394 -0.391 -0.238 -0.308 -0.258 (2.01) (2.97) (2.92) (3.05) (3.00) (2.81) (2.54) (2.40) Bank R 0.981 0.835 0.822 0.812 0.793 0.416 0.826 0.722 (4.78) (4.35) (4.43) (4.49) (4.41) (3.81) (4.65) (4.07) Bank N -0.599 -0.604 -0.604 -0.643 -0.583 -0.341 -0.559 -0.603 (3.72) (3.72) (3.68) (3.83) (3.68) (2.92) (3.59) (3.36) R2 0.736 0.698 0.699 0.700 0.698 0.277 0.273 0.703
Table 8: Concetration Ratio and Bank Returns: Interative Modle II
Dep. Var. = ROE
const. -6.846 -4.830 const. 2.683 (2.39) (2.09) (1.69) CR 0.020 -0.026 CR -0.014 (2.84) (1.76) (0.38) CR BD S -0.000 -0.006 CR LawHabit -0.000 (0.06) (1.07) (1.92) CR BD I -0.006 0.001 CR LawRisk -0.007 (2.81) (0.37) (1.38) CR BD R 0.019 0.006 CR Corrupt -0.010 (3.93) (0.75) (2.37) CR BD N -0.012 0.018 CR Con s 0.010 (2.93) (2.94) (0.87) Equity 0.476 0.476 CR Contract -0.005 (4.19) (4.38) (0.52) MAG 0.035 0.037 CR Account 0.000 (1.62) (1.81) (0.02) Liquid -0.017 -0.013 CR Gove 0.015 (1.38) (1.25) (0.94) Credit -0.002 -0.003 CR Supervision 0.000 (0.77) (0.99) (0.27) GDP Per 0.000 0.000 Equity 0.112 (3.11) (2.96) (3.04) In a -0.002 -0.002 MAG -0.031 (2.18) (2.14) (1.84) rate 0.001 0.001 Liquid -0.021 (1.89) (1.87) (1.37) Bank S 0.396 Credit -0.006 (1.84) (2.83) Bank I -0.302 GDP Per 0.000 (1.48) (2.71) Bank R 0.411 In a -0.001 (1.40) (0.79) Bank N -1.320 rate 0.000 (4.48) (0.80) Bank S 0.249 (2.41) Bank I -0.435 (3.82) Bank R 0.448 (3.27) Bank N -0.461 (2.61) R2 0.677 0.702 0.321
