The impact of bank capital on profitability and risk in Asian banking

Journal of International Money and Finance 32 (2013) 251–281
Contents lists available at SciVerse ScienceDirect
Journal of International Money

and Finance
journal homepage: www.elsevier.com/locate/jimf
The impact of bank capital on profitability and risk in Asian

banking
Chien-Chiang Lee a, *, Meng-Fen Hsieh b
a
Department of Finance, National Sun Yat-sen University, Kaohsiung, Taiwan
b
Department of Finance, National Taichung University of Science and Technology, Taichung, Taiwan
a b s t r a c t
JEL classification: This article applies the Generalized Method of Moments technique
G21 for dynamic panels using bank-level data for 42 Asian countries
C23 over the period 1994 to 2008 to investigate the impacts of bank
E52
capital on profitability and risk. Ignoring influence factors, the
Keywords: extant literature presents an ambiguous impact of bank capital on
Bank capital profitability (risk), however, when the effects from the influencing
Profitability factors are taken into consideration, three conclusions are reached.
Risk First, along with the change in the categories of banks, investment
Asia banks have the lowest and positive capital effect on profitability,
Dynamic panel
whereas commercial banks reveal the highest reverse capital effect
on risk. Second, banks in low-income countries have a higher
capital effect on profitability; banks in lower-middle income
countries have the highest reverse capital effect on risk, while
banks in high-income countries have the lowest values. Third,
banks in Middle Eastern countries own the highest and positive
capital effect on profitability. Far East & Central Asian banks have
the largest reverse capital effect on risk, while the lowest value
occurs in Middle Eastern countries’ banks. Finally, our results also
reveal that persistence of profit is greatly affected by different
profitability variables, and all risk variables show persistence from
one year to the next.
Ó 2012 Elsevier Ltd. All rights reserved.
* Corresponding author. Tel.: þ886 7 5252000x4825; fax: þ886 7 5254899.

E-mail address: cclee@cm.nsysu.edu.tw (C.-C. Lee).
0261-5606/$ – see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jimonfin.2012.04.013
252 C.-C. Lee, M.-F. Hsieh / Journal of International Money and Finance 32 (2013) 251–281
1. Introduction
How does bank capital affect risk? In response to the recent global financial crisis, the Basel
Committee on Banking Supervision (BCBS) set forth to update the guidelines for capital and banking
regulations. Basel III proposes many new capital, leverage, and liquidity standards to strengthen
regulation, supervision, and risk management in the banking sector. The capital standards and new
capital buffers will require banks to hold more capital and a higher quality of capital than under current
Basel II rules.
Regulators in most jurisdictions around the world are still planning to implement the new accord
(Basel II), but have widely varying timelines and are restricting the use of varying methodologies. For
instance, in 2008 the Federal Reserve Board of the U.S. proposed a rule for public comment that would
institute certain less-complex approaches for calculating risk-based capital requirements. The
proposal, known as the standardized framework, would be available for banks, bank holding compa-
nies, and savings associations not subject to the advanced approaches of Basel II.1 More importantly,
the recent credit crisis has emphasized the need to further understand the determinants of bank risk in
an environment of lower bank capital (Festic et al., 2011).
The European Union has already implemented the Basel II accord via the EU Capital Requirements
Directives, and many European banks have already reported their capital adequacy ratios according to
the new system. All credit institutions in the EU adopted Basel II at the beginning of 2008. The
Monetary Authority of Singapore (MAS) also implemented Basel II for all Singapore-incorporated banks
on January 01, 2008.2 The Reserve Bank of India (RBI) implemented the Basel II standardized norms on
31 March 2009 and is moving to internal credit ratings for operational risks in banks.
It is thus no surprise that the relationship between bank capital and risk (profitability) has recently
become a cause for concern, especially as the level of capital may give rise to both beneficial and
adverse effects on bank profitability. Despite this, empirical work on the topic is either scant or mixed,
particularly in Asia. For example, Barth et al. (2008) find notably that some Asian countries, such as
Philippines, Singapore, and Indonesia, are strengthening capital requirements, while some are easing
their capital requirements, such as South Korea and Japan, in the aftermath of their crises. This differs
from Argentina, which made the same move, but in advance of the crisis. At the same time, South
Korea, Malaysia, Singapore, and Thailand provided supervisors with more explicit power. Unfortu-
nately, an increase in supervisors’ powers was found to be not helpful in bank performance and
stability (Barth et al., 2006). Worse, according to Barth et al. (2008) and their simulation results, some
Asian countries have increased banking-system fragility by intensifying regulatory restrictions.
The Asian banking industry therefore provides an interesting laboratory for investigation. Against
the current background of change in Asian banking, we aim to shed some crucial light on the deter-
minants of bank risk-taking and analyze its relationship with capital and profitability. This is, to our
knowledge, the first study of bank capital’s impacts on profitability and risk for different income levels,
specializations, and geographic regions in the Asia banking industry.
Previous studies focusing on the relationship between capital and risk have mixed results (Aggarwal
and Jacques, 1998). Some studies find a positive relationship between capital and risk, meaning
regulators encourage banks to increase their capital commensurably with the amount of risk taken,
which refers to the ‘regulatory hypothesis’ (Pettway, 1976; Shrieves and Dahl, 1992; Berger, 1995;
Demirgüç-Kunt and Huizinga, 2000; Iannotta et al., 2007; etc.). Nevertheless, opposite results are
found in some studies. A negative relationship between capital and risk may refer to the ‘moral hazard
hypothesis’ whereby banks have incentives to exploit existing flat deposit insurance schemes
(Demirgüç-Kunt and Kane, 2002). For instance, Jahankhani and Lynge (1980), Brewer and Lee (1986),
Karels et al. (1989), Jacques and Nigro (1997), and Agusman et al. (2008) show that equity-to-total-
assets is negatively related to risk. As also indicated by Kahane (1977), Koehn and Santomero (1980),
and Kim and Santomero (1988), banks could respond to regulatory actions, forcing them to increase
their capital by increasing asset risk (Altunbas et al., 2007). However, the puzzle between capital and
1
http://www.federalreserve.gov/newsevents/press/bcreg/20080626b.htm.
2
http://www.mas.gov.sg/fin_development/banking/Basel_II.html.
C.-C. Lee, M.-F. Hsieh / Journal of International Money and Finance 32 (2013) 251–281 253
risk, as suggested by Hughes and Moon (1995), Hughes and Mester (1998), and Altunbas et al. (2007), is
that capital and risk are likely to be influenced by the level of profitability in the banking sector.
The relationship between capital and profitability usually focuses on the macro prospective, or the
‘structure-conduct-performance (SCP hereafter) hypothesis’. Some studies show that operating perfor-
mance is significantly related to market structure (measured by concentration indices, such as CR4 (the
four-bank concentration ratio) or HHI (Herfindahl-Hirschman index)). Market structure, which refers
to the degree of market concentration within an industry, represents the degree of competition within
a specific industry.3 Claessens and Laeven (2004) relate competitiveness to indicators of countries’
banking system structures and regulatory regimes. Importantly, and consistent with some other
studies, they find no evidence that their competitiveness measure negatively relates to banking system
concentration or the number of banks in the market. A subsequent increase in the level of capital
should lead to an increase in profitability due to lower insurance expenses on uninsured debt. Alter-
natively, a bank that holds a relatively high proportion of capital is unlikely to earn high profits; yet is
less exposed to risk (Goddard et al., 2004).
Previous studies on the relationship between capital and risk (profitability) mainly focus on the U.S.
or European area, with much less discussion and insight on the banking industry in Asian economies.
Thus, in order to allow for changes in the macroeconomic environment that might systematically
impact observed risk and capital in any given year, binary variables are usually added to the specifi-
cation; otherwise, macroeconomic factors are treated as control variables, implying a linear relation-
ship between capital and risk.4 The question now arises as to whether the U.S. or European results are
valid for Asian countries. Research on Asian banks is sparse, but very important, because they are the
predominant source of finance for businesses in the Asian private sector (Deesomsak et al., 2004).
Firm-specific risk is more important than systematic risk, and the results are robust even though
differences exist across Asia in each country’s banking activities, capital adequacy requirements, and
deposit insurance protection. Furthermore, banks in this region have experienced a banking crisis, and
bank restructuring programs still continue in several Asian countries (Agusman et al., 2008). This
interesting analysis on Asian banks can potentially help in understanding banking in other countries as
well.
We contribute to existing empirical analyses in several ways. First, the existing literature has drawn
a lot of attention on U.S. or European cases (Fernández de Guevara et al., 2007; Carbo et al., 2009).
Although Asia has become an important economic area, the Asian topic has not earned enough
discussions. Thus, the purpose of this paper is to examine Asian banks with the latest and a wider range
of panel data that cover 2,276 banks from 1994 to 2008 in 42 countries.
Second, most studies focus mainly on the relationship between capital and risk, yet seldom on the
relationship between capital and profitability. This study discusses capital, risk, and profitability
together.
Third, using panel data produces another problem in which different countries are treated together
as an entity and not as a separate unit. Therefore, one cannot identify the differences in the rela-
tionships among capital, profit, and risk. This study thus classifies the panel data into different sub-
panels and considers the effect of different conditions on the relationship between capital and risk
(or profitability), including income levels, specializations, geographic regions (Far East & Central Asia,
Middle East - also called Western Asia), and the Asian crisis (Non-Crisis and Post-Crisis). Indeed, if
related conditions play an important role in influencing the effects of capital on profit (or risk), then
one can expect countries with the same level of capital to have very different outcomes in terms of
3
In the field of banking competition measurement, Maudos and Solís (2011) examine the evolution of the degree of
competition in the Mexican banking system by applying two indicators from the theory of industrial organization: the Panzar
and Rosse H-statistic and the Lerner index of market power.
4
Instead, Claessens and Klingebiel (2001) analyze the costs and benefits of different degrees of financial sector competition
and permissible activities, as well as the related implications for regulation and supervision. According to country experiences
and theory, they suggest that competitiveness does not necessarily require having many financial institutions, because
a concentrated system can be competitive if it is contestable (i.e., open to competition). Claessens and Laeven (2003) relate
a competition measure to industrial growth for 29 banking systems, finding that the effects of competition on access to
financing (and growth) depends on level of development in the financial systems.
profitability (or risk). An investigation on the impact of these conditions has profound implications as
far as policy-making goes, because it may reveal that a country should not just blindly follow the
general consensus.
Fourth, previous research usually samples commercial banks, and it is very rare to consider other
specializations. We examine the overall banking system, including commercial banks, cooperative
banks, investment banks, and others.5 Banks of different ownership characteristics differ in their
attitudes to managing capital, profitability, and risks. Further classification enables us to answer
whether different types of banks will influence the relationship between banking capital and profit-
ability (risk).
Fifth, we adopt four proxies for profitability - i.e. return on assets (ROA), return on equities (ROE), net
interest margin (NIM), and net interest revenue against average assets (NR) - and three for risk - i.e.
variance of ROA (VROA), variance of ROE (VROE), and loan loss reserves (LLR). Doing so enables this study
to provide robust tests and find out the proxy for profitability and risk that is suitable for Asian banks.
Sixth and finally, dynamic panel techniques are adopted to analyze the panel data, which are
designed to check the persistence of profit (or risk). We thus investigate the persistence of profit when
entry and exit are sufficiently free to eliminate any abnormal profit, and that the profit rates of all banks
tend to converge to the same long-run average value. If abnormal profit tends to persist from year to
year, then differences in average profit rates may be sustained indefinitely.
The remainder of this paper is organized as follows. Section 2 reviews the relevant literature.
Section 3 describes the econometric model and data sources employed herein. Section 4 discusses the
empirical results, and Section 5 presents our conclusions.
2. Literature review
2.1. Relationship between capital and risk
Does a higher level of capital structure refer to a lower risk for banks? This topic earns a round of
discussions, including from the perspective of supervisors. The relationship between a bank’s capi-
talization and its risk-taking behaviors is one of central topics in banking studies, because of the
potential implications for regulatory policies. Table 1 provides a comprehensive literature survey. As
can be seen, from the prospective of data selection most studies either focus on U.S. or European banks.
From the methodology viewpoint, most studies utilize a static model. Thus, the contribution of this
paper to the literature is by increasing significantly the focus on Asian banks and adopting the dynamic
the Generalized Method of Moments (GMM) technique.
We specifically employ the recent two-step dynamic panel data approach to assess the relationships
among banking capital, profitability, and risk. Capital, profitability, and risk should be considered
simultaneously when examining the regulatory, SCP, and moral hazard hypotheses, because a change
in capital gives rise to a change in banks’ profitability and risk. Therefore, the level of capital should be
treated as an endogenous variable when we investigate the regulatory and moral hazard hypotheses.
The GMM model resolves the possible simultaneity between the degree of capital and profitability
(risk) and takes into account the causal effect of the exogenous component.6
This study differs from an influential paper, Altunbas et al. (2007), who use a static model to analyze
the relationships among capital, risk, and efficiency for European banks between 1992 and 2000.
However, our study augments Altunbas et al. (2007) and their European analysis to Asian countries for
the period 1994 to 2008, which is much longer than the existing studies. The data covered in our
research in the extended period after 2000 could account for the fact that more financial deregulation
took place. Second, this study adopts the dynamic model, which is able to provide the persistence of
profitability (risk) estimation. Third, more profitability (risk) variables are considered, as compared to
Altunbas et al. (2007), who use a single profitability variable (return on assets) and risk (loan loss
5
The set of “other” banks is defined as all remaining banks.
6
The dynamic GMM estimator is employed as instruments lagged values of the dependent variable in levels and in
differences, and lagged values of other regressors that could potentially suffer from endogeneity.
Table 1
Literature on the relationships among capital, profitability, and risk.
Authors (year) Period Countries Methodologies Empirical results

Aggarwal and 1990–1993 2,552 U.S. FDIC-insured 2SLS Mixed results are found in assessing
Jacques (1998) commercial banks the relationship between changes in
capital ratios and changes in risk.
Agusman et al. (2008) 1998–2003 Asian banks Panel data model Equity-to-total-assets are negatively
(46 listed banks) related to risk, but do not reach any
significance.
Altunbas et al. (2007) 1992–2000 15 European banks Panel data model Inefficient European banks appear to
hold more capital and take on less
risk. Empirical evidence shows
a negative relationship between risk
and the level of capital.
Berger (1995) 1983–1989 U.S. commercial banks Granger-causality There is a strong positive
relationship between capital and
earnings (ROE), meaning well
capitalized firms face lower
expected bankruptcy costs, which in
turn reduce their cost of funding and
increase their profitability.
Brewer and 1978–1984 44 U.S. bank holding Panel data model Capital is associated with negative
Lee (1986) companies risk.
Demirgüç-Kunt and 1990–1997 44 countries: Panel data model Profit is positively related to the
Huizinga (2000) Developed and lagged equity variable. GNP per
Developing capita is set as one of the
macroeconomic independent
variables and is also positively
related to bank profit.
Goddard et al. (2004) 1992–1998 665 banks from six Dynamic panel The relationship between the
European countries model capital-assets ratio and profitability
is positive.
Iannotta et al. (2007) 1999–2004 15 European countries Panel data model Capital is associated with positive
(181 large banks) profitability (the ratio of operating
profit to total earning assets) and
risk, which can be attributed to the
fact that a different asset risk can be
compensated by a different level of
capitalization. The coefficients of the
GDP growth rate on profit and risk
are significantly positive (Table 4).
Jacques and 1990–1991 2,570 U.S. FDIC-insured 3SLS to A negative (positive) association is
Nigro (1997) commercial banks estimate the found between changes in risk
simultaneous (profitability) and capital.
equation model
Jahankhani and 1972–1976 95 U.S. commercial Panel data model Capital is negatively associated
Lynge (1980) banks and bank with risk.
holding companies
Karels et al. (1989) 1977–1984 24 U.S. banks CAPM and The coefficients of correlation
correlation between the primary capital
adequacy ratio and systematic risk
are negative in each case with levels
of significance.
Pettway (1976) 1971–1974 U.S banks and bank Regression Equity-to-total-assets are positively
holding companies and significantly related to risk
(77 issues of capital (premium) (Table 1, p. 868).
notes)
Shim (2010) 1993–2004 U.S. property liability 3SLS to Undercapitalized insurers increase
insurance industry estimate the capital to avoid regulatory costs and
simultaneous take on more risk to generate higher
equation model returns, meaning capital has
a positive relationship between
profitability (ROA) and risk.
(continued on next page)
Table 1 (continued )
Authors (year) Period Countries Methodologies Empirical results

Shrieves and 1983–1987 1,800 FDIC-insured 3SLS to A positive association is found
Dahl (1992) commercial banks estimate the between changes in risk and capital.
simultaneous
equation model
Rime (2001) 1989–1995 Swiss banks 3SLS to estimate A positive association is found
the simultaneous between changes in risk and capital.
equation model Current earnings have a significant
and positive impact on capital,
indicating that profitable banks
improve their capitalization through
retained earnings.
reserve). Finally, Altunbas et al. (2007) ignore the possible effects of income levels and geographic
regions, as well as financial market regulations and institutional developments.
A pioneer research by Pettway (1976) explores the relationship between capital structure and risk
for U.S. banks and bank holding companies over the period of 1971 and 1974, surprisingly finding
a positive relationship between equity-to-total-assets and risk. Shrieves and Dahl (1992) also adopt U.S.
data and reach the same positive result. Similar results are reached by applying Europe data, such as in
Rime (2001) and Iannotta et al. (2007).
Some works find opposite results. In a utility maximizing and mean-variance framework, banks
with relatively low risk aversion will choose relatively high leverage (low capital) and relatively high
asset risk (Kim and Santomero, 1988). Equity-to-total-assets are found to be negatively related to risk
(Jahankhani and Lynge, 1980; Brewer and Lee, 1986; Karels et al., 1989; Jacques and Nigro, 1997; and
Agusman et al., 2008). More specifically, Jahankhani and Lynge (1980) adopt data from 95 U.S.
commercial banks over 1972–1976. Brewer and Lee (1986) analyze 44 U.S. bank holding companies
during 1978–1984. Karels et al. (1989) explore 24 U.S. banks between 1977 and 1984. Jacques and Nigro
(1997) conduct a study on 2,570 U.S. FDIC-insured commercial banks during 1990 and 1991, offering
evidence of a negative association between changes in risk and capital. Agoraki et al. (2011) find that
capital requirements reduce risk in general, but for banks with market power this effect significantly
weakens or can even be reversed.
The negative relationship between capital and risk may refer to the ‘moral hazard hypothesis’ that
undercapitalized banks take on excessive risk to exploit existing flat deposit insurance schemes
(Demirgüç-Kunt and Kane, 2002). Kwan and Eisenbeis (1997) explore findings that show a positive
effect of inefficiency on risk-taking, which supports the moral hazard hypothesis that poor performers
are more vulnerable to risk-taking than high performance banks. As for a different country sample, the
same result is still found, such as in Altunbas et al. (2007) and Agusman et al. (2008).
As noted by Shrieves and Dahl (1992), a positive correlation between capital and risk may result
from regulatory costs, the unintended impact of minimum capital requirements, bankruptcy cost
avoidance, or risk aversion by bank managers, while a negative correlation may result from the mis-
pricing of deposit insurance.7 Altunbas et al. (2007) further refer to a positive relationship between
capital and risk as the ‘regulatory hypothesis’, meaning regulators encourage banks to increase their
capital commensurably with the amount of risk taken, while a negative relationship may refer to the
‘moral hazard hypothesis’ that banks have incentives to exploit existing flat deposit insurance schemes.
2.2. Relationship between capital and profitability
For the puzzle between capital and risk demonstrated in the previous section, as suggested by
Hughes and Moon (1995), Hughes and Mester (1998), and Altunbas et al. (2007), capital and risk are
7
Sanyal and Shankar (2011) demonstrate that liberalization has had some positive performances: a fall in the share of non-
performing loans in banks’ portfolios; increased entry of new private sector banks; branch expansion; and the achievement of
the capital adequacy ratio by 90% of domestic banks.
also likely to be influenced by banking firms’ level of profitability. However, the extant literature on
capital structure and risk seldom takes profitability into consideration - see Table 1 for details. Capital is
found to be associated with positive profitability (Berger, 1995; Jacques and Nigro, 1997; Demirgüç-
Kunt and Huizinga, 2000; Rime, 2001; and Iannotta et al., 2007).
Differing from the positive relationship between capital and profitability, Altunbas et al. (2007) find
that inefficient European banks appear to hold more capital. Goddard et al. (2004) demonstrate that
the relationship between the capital-assets ratio and profitability is positive in six major European
banking sectors for the period 1992–1998. However, Goddard et al. (2010) explore that a negative
relationship between the capital ratio and profitability reflects the standardized risk-return payoff for
eight European Union member countries between 1992 and 2007.
Another literature stream concentrating on the relationship between capital and profitability
focuses on the macro prospective, structure-conduct-performance hypothesis. The results of such
research show that operating performance is significantly related to market structure. Market struc-
ture, which refers to the degree of market concentration within an industry, represents the degree of
competition within the specific industry. For example, Heggestad (1977), Short (1979), and Akhavein
et al. (1997) find that, within a financial system characterized by less competition, firms tend to
have larger scales of operation, and this in turn leads to a higher degree of market concentration and
profits (Hannan and Berger, 1991; Neumark and Sharpe, 1992; Demirgüç-Kunt and Huizinga, 1999).8
3. Methodological issues and data sources
3.1. Methodology
This study applies the two-step dynamic panel data approach suggested by Arellano and Bover
(1995) and Blundell and Bond (2000) and also uses dynamic panel GMM technique to address
potential endogeneity, heteroskedasticity, and autocorrelation problems in the data (Doytch and
Uctum, 2011). The system estimator provides for a more flexible variance-covariance structure
under the moment conditions. The GMM approach is superior to traditional OLS in examining financial
variable movements. For instance, Driffill et al. (1998) indicate that a conventional OLS analysis of the
actual change in the short rate on the relevant lagged term spread yields coefficients with some wrong
signs and wrong size.
There are two different estimators for the dynamic panel models: (i) the difference panel estimator
eliminates a potential source of omitted variable bias in the estimation, and (ii) the system panel model
estimator combines the regression difference with the regression in levels in order to reduce the
potential biases and imprecision associated with the difference estimator (Arellano and Bover, 1995).
Linear GMM estimators have one- and two-step variants. The two-step estimator that we use is
generally more efficient than the one-step estimator, especially for the system GMM. We employ
Windmeijer’s (2005) finite-sample correction to report standard errors of the two-step estimation,
without which those standard errors tend to be severely downward biased.
The dynamic panel model technique - the GMM model - is particularly well-suited to handling short
macro panels with endogenous variables and is also helpful in amending the bias induced by omitted
variables in cross-sectional estimates and the inconsistency caused by endogeneity. It is rather
convenient that the dynamic GMM technique at the same time allows us to control for the endogeneity
bias induced by reverse causality running from profit (or risk) to banking capital and other explanatory
variables.
Our study adopts the dynamic panel data approach and GMM to estimate the parameters. Even
though there is correlation or heteroskedasticity among the equations, the estimated standard devi-
ation still appears to be robust. Therefore, the independent variable with lagged periods is included in
Eqs. (1) and (2), as shown below. Beyond the dynamic panel data, the model that establishes the
relationship between bank capital and profitability (risk) is based on the earlier literature. According to
8
Please refer to Hsieh and Lee (2010) for more discussion.
the earlier literature discussion and this study’ purpose of research, we modify the works of Altunbas
et al. (2007), Casu and Girardone (2006) and Goddard et al. (2004) to establish the relationship
between bank capital and profitability (risk).
This paper mainly investigates the relationships among capital, profitability, and risk for Asian
banks with the latest and a wider range of panel data that cover 2,276 banks from 1994 to 2008. The
relationship between bank capital and profitability (risk) can be specified as follows:
pit ¼ a0 þ a1 pit1 þ a2 CPit þ a0Fit þ li þ hit ; ci; t: (1)
Vit ¼ b0 þ b1 Vit1 þ b2 CPit þ b0Fit þ mi þ vit ; ci; t: (2)

Here, t and i denote time period and banks, respectively, li (mi ) is an unobserved bank-specific effect,
and hit (vit ) is the idiosyncratic error term.
Eqs. (1) and (2) are designed to examine the impact of bank capital on bank profitability and risk,
respectively. Term CPit is the level of bank capital, proxied by the equity-to-assets ratio; pit refers to the
i th bank’s profitability in year t, proxied by four profitability variables: return on assets (ROA), return
on equities (ROE), net interest margin (NIM), and net interest revenue against average assets (NR). Here,
Vit denotes the i th bank’s risk in year t, proxied by three risk variables (Lepetit et al., 2008): variance of
ROA (VROA), variance of ROE (VROE), and loan loss reserves (LLR). Term Fit includes the set of
explanatory variables, while a1 and b1 are the estimated persistence coefficient for profitability and
risk, respectively. A significant a1 implies that there is abnormal profitability or that risk will last from
one year to the next (Goddard et al., 2004, 2010). Banks are always accompanied by the feature of
profitability persistence, difficulty in entry-and-exit, a monopoly on resources, and a special ability for
management resource allocation. Thus, it is crucial to consider the persistence of profitability through
the dynamic panel model.
As for the related internal control variables, according to Casu and Girardone (2006), Short (1979),
and Smirlock (1985), they include loan loss reserves to gross loans (LLGL), net loans to total assets
(NITA), and liquid assets to customer and short-term deposits (LADSF). The coefficients of LLGL and NITA
are expected to be positive with profitability (risk). A higher level of loans implies a higher profit (risk)
will be generated. On the contrary, LADSF is expected to be negative with profitability (risk), since
keeping more liquid assets is usually accompanied with lower return (risk).
Four macro control variables are set as the related external control variables: inflation (INFL), GDP
growth rate (GW), domestic credit to private sector (DCPS), and real interest rate (RIR). The coefficients
of INFL and RIR are uncertain. In high-inflation countries, banks may charge customers more, yet at the
same time they face due loans that are shrinking. A higher GW may imply that banks can generate more
profitability and less risk, but DCPS may go another way. Since a greater DCPS refers to a financial
environment that is more competitive, then DCPS should be negative (positive) to profitability (risk).
Four variables on financial market regulations and four institutional developments are finally
included. The four financial market regulations followed from Agoraki et al. (2011) and Delis et al.
(2011) are: Capital requirements (CAPR), Supervisory power (SPR), Market discipline and private
monitoring (MDPM), and Activity restrictions (ACTR). CAPR is an index of capital requirements that
accounts for both initial and overall capital stringency. It is calculated by considering the sources of
funds used as capital and by taking into account various issues that emerge during the calculation of
the capital-to-assets ratio. The index takes values between 0 and 8, with higher values indicating
greater capital stringency. SPR measures the power of supervisory agencies, indicating the extent to
which these authorities can take specific actions against bank management and directors, share-
holders, and bank auditors. This index takes values between 0 and 14 with higher values indicating
more SPR. MDPM is an indicator of market discipline and shows the degree to which banks are forced to
disclose accurate information to the public and whether there are incentives to increase market
discipline. This index ranges between 0 and 8 with higher values indicating greater MDPM. Finally,
ACTR is a proxy for the level of restrictions on banks’ activities in each country. It is determined by
considering whether participation in securities, insurance, and real estate activities and ownership of
non-financial firms are unrestricted, permitted, restricted, or prohibited. ACTR takes values between 4
and 16, with higher values indicating higher restrictions.
The four institutional developments are: Deposit insurance explicit (DEP), Shareholder protection
(GSP), Creditor protection (GCP), and Legal efficiency (GLE). According to Demirgüç-Kunt and Kane
(2002), under the explicit deposit insurance schemes banks have more incentives for risk-taking.
The last three variables are proxies for corporate governance. GSP and GCP range from 0 to 5, with
higher scores for higher protection, while GLE is the multiple value for the rule of law and the efficiency
of the judicial system.
To specify whether the instruments are valid, we adopt the specification test suggested by Blundell
and Bond (2000) and use the Sargan test of over-identifying restrictions, which examines the validity of
the instruments.9 If the null hypothesis of the Sargan test for validly over-identifying restrictions
cannot be rejected, then the instrumental variables are valid. On the contrary, if we reject the null
hypothesis, then the instrumental variables are inappropriate. Under the null of joint validity of the full
instrument set, the Sargan test statistics are asymptotically c2. The second test examines the
hypothesis that the error term is not serially correlated. In both the difference regression and the
system difference-level regression, we test whether the differenced error term is second-order serially
correlated.
For operational and analytical purposes, this study further considers the effect from different
economic levels, whereby the effects are not linear, but rather bell type. In order to distinguish the
effect from different economic levels, this study also looks at different regulatory and institutional
developments. According to the World Bank, the main criterion for classifying economies is 2008 gross
national income (GNI) per capita, calculated using the World Bank Atlas method. Based on its GNI per
capita, every economy is classified as: low income, $975 or less; lower-middle income, $976–$3,855;
upper-middle income, $3,856–$11,905; and high income, $11,906 or more (Claessens et al., 2001; Shen
and Lee, 2006; Aggarwal and Goodell, 2009b).10
This study considers the other two different definitions of economic levels by the IMF and OECD as
well. According to the International Monetary Fund, advanced economies comprise 65.8% of global
nominal GDP and 52.1% of global GDP (PPP) in 2010 (or 54.7% in 2008). Thus, six countries among our
sample are classified as advanced countries: Hong Kong, Israel, Japan, Singapore, South Korea, and
Taiwan. The others are grouped as developing countries.11
Based on OECD standards, there are 31 high-income OECD members. Europe encompasses 24
countries, 2 countries are in North America, 2 countries are in Oceania, and 3 countries are in Asia
(Israel, Japan, and South Korea).12
Other analytical groups taken from different specializations (such as commercial banks, cooperative
banks, investment banks, and others), geographic regions (Far East & Central Asia, Western Asia -
Middle East), and sample period (Post-Asian Crisis period, and Non-crisis period) are also considered.
3.2. Descriptions and sources of data
This study analyzes a panel dataset comprising 42 Asian countries over the period 1994–2008. A
bank is deleted from the sample if it does not have 3 continuous years of data. As shown in the fourth
column of Table 2, of the total number of banks (2,276) in the sample countries, Japan (850), Indonesia
9
Xtabond2 is a command of Stata. It also reports a test of over-identifying restrictions - i.e., whether or not the instruments,
as a group, appear exogenous.
10
Aggarwal and Goodell (2009a) include a measure of economic inequality using the Gini coefficient. They mention that
nations with greater economic inequality might favor market financing as wealthy families may play a significant role.
Furthermore, GDP per capita or the GDP growth rate is usually used to measure economic development (level), such as in
Hassan et al. (2011). The GNI per capita calculated by the World Bank, however, provides a clear cutoff point in classifying
countries and is widely adopted by the extant literature, such as Claessens et al. (2001), Shen and Lee (2006), Aggarwal and
Goodell (2009b), etc. For example, Aggarwal and Goodell (2009b) use nations classified according to wealth from Beck et al.
(2000): the World Bank classification of countries according to their income levels (World Development Indicators 1998).
Four country groups are distinguishable: high-income countries with a GNP per capita in 1997 higher than $9,656; upper-
middle income countries with a GNP per capita between $3,126 and $9,655; lower-middle income countries with a GNP per
capita between $786 and $3,125; and low-income countries with a GNP per capita of less than $786.
11
Please refer to World Economic Outlook Database published by the International Monetary Fund.
12
Please refer to http://data.worldbank.org/about/country-classifications/country-and-lending-groups#OECD_members.
Table 2
Numbers of banks and income levels.
No. Code Country Bank number Income level Deposit insurance

explicit ¼ 1
implicit ¼ 0
Far East and Central Asia
1 AM Armenia 16 lower- middle 0
2 AZ Azerbaijan 22 lower- middle 0
3 BD Bangladesh 38 low 1
4 BT Bhutan 2 lower- middle 0
5 CN China 96 lower- middle 0
6 GE Georgia 13 lower- middle 0
7 HK Hong Kong 96 high 0
8 ID Indonesia 112 lower- middle 1
9 IN India 102 lower- middle 1
10 JP Japan 850 high 1
11 KG Kygyzstan 6 low 1
12 KH Cambodia 12 low 0
13 KR South Korea 72 high 1
14 KZ Kazakhstan 29 upper- middle 1
15 LA Laos 3 low 0
16 LK Sri Lanka 20 lower- middle 1
17 MN Mongolia 10 lower- middle 0
18 MO Macau 10 high ?
19 MY Malaysia 103 upper- middle 1
20 NP Nepal 17 low 0
21 PH Philippines 55 lower- middle 1
22 PK Pakistan 56 low 0
23 SG Singapore 63 high 0
24 TH Thailand 51 lower- middle 1
25 TJ Tajikistan 2 low 0
26 TW Taiwan 97 high 1
27 UZ Uzbekistan 14 low 0
28 VN Vietnam 32 low 1
Middle East
29 AE United Arab Emirates 31 high 0
30 AF Afghanistan 2 low 0
32 BH Bahrain 39 high 1
32 IL Israel 20 high 0
33 IQ Iraq 9 lower- middle 0
34 IR Iran 15 lower- middle 0
35 JO Jordan 15 lower- middle 1
36 KW Kuwait 27 high 0
37 LB Lebanon 63 upper- middle 1
38 OM Oman 14 upper- middle 1
39 QA Qatar 9 high 0
40 SA Saudi Arabia 15 upper- middle 0
41 SY Syria 6 lower- middle 0
42 YE Yemen 12 low 0
Notes: Economies are divided according to 2008 GNI per capita, calculated using the World Bank Atlas method. If GNI per capita
is not available, then GDP per capita is substituted. The groups are: low income, $975 or less; lower-middle income, $976–
$3,855; upper-middle income, $3,856–$11,905; and high income, $11,906 or more. The deposit insurance data are taken from
Demirgüç-Kunt et al. (2005).
(112), and Malaysia (103) have the largest number. The original data source is Bankscope using the 2009
version. Income level is presented in the fifth column of Table 2. Economies are divided according to
2008 GNI per capita, calculated using the World Bank Atlas method. If GNI per capita is not available,
then GDP per capita is substituted. The groups are: low income (11 countries); lower-middle income
(16 countries); upper-middle income (4 countries); and high income (11 countries). Economies are also
divided according to the IMF definition of advanced (6 countries) and developing (36 countries). Based
on OECD standards, there are 3 countries in our sample filed as high-income members, while the rest
are non-high-income ones.
The final column of Table 2 is whether a country has explicit or implicit deposit insurance.
According to Demirgüç-Kunt et al. (2005), 2 countries (Bangladesh and Vietnam) have implemented
explicit deposit insurance, occupying 18% of the low-income sample; 50% of lower-middle income
countries have set up explicit deposit insurance; 75% for upper-middle countries; and 45% for high-
income countries. Thus, middle-income countries have a higher proportion using explicit deposit
insurance.13
Table 3 gives definitions of the variables and sources of the data. The financial statements are ob-
tained from the Bureau Van Dijk BankScope database. The macroeconomic data are from World
Development Indicator. Figs. 1 and 2 show the trends of capital, profitability, and risk in different
classifications. Indeed, CP, ROA, and VROA perform differently among groups. For example, investment
banks have a higher level of capital versus the others and are more aggressive after the Asian crisis.
Furthermore, there is a decline in bank profitability and an increase in bank risk after the Asian crisis.
The most significant case is the lower-middle income group classified by the World Bank, and it is
equally the same for developing countries defined by the IMF, or non-high-income countries by the
OECD, as their profitability largely drops and risk increases during the post-crisis.
Table 4 gives the variable means over the sample period for different types of banks in terms of
specialization, income level, and geographic region. For the sample averages of the independent
variable (Capital), CP, the sample mean is 12.08%. The highest values for the different types of banks are
for investment banks (27.93%), middle-low income countries (14.75%), and Middle East countries
(19.05%). As for the dependent variable (bank profitability), the average ROE is 5.90%, and the highest
values for the different types of banks are for commercial banks (8.36%), middle-high income countries
(12.01%), and Middle East countries (13.89%). Regarding the other dependent variable (risk), the
average variance of ROE (VROE) is 1553.55%, and the high value mainly results from middle-high
income countries (3092.14%).
Table 5 provides the matrix of Pearson correlation coefficients. The correlation coefficients are
usually very small (less than 0.3), indicating that the correlation between variables has weak
association.14
4. Empirical analysis
4.1. Benchmark results
Our testable hypotheses are based on the SCP, Moral Hazard, or Regulatory hypotheses. Specifically,
we first examine all Asian banks together and then check the different sub-panels separately. In our
sub-panel studies, we investigate the impacts of country-specific and banking industry characteristics
on the linkage among capital, profitability, and risk.
This study econometrically adopts the two-step GMM dynamic system panel estimator developed
by Arellano and Bover (1995) and Bond and Blundell (2000). Blundell and Bond (2000) show that
a system panel estimator that simultaneously uses difference panel data and data based on the original
levels of specification gives rise to dramatic increases in both consistency and efficiency (Beck and
Levine, 2004; Doytch and Uctum, 2011).
Table 6 reports the empirical results of the full sample when Eq. (1) is considered, which focuses on
when the two-step system GMM and dynamic panel data approach are adopted, as well as the esti-
mation results of capital and profitability. We find significantly positive relationships between capital
and profitability (proxied by ROA, NIM, and NR) for overall Asian banking, indicating that SCP holds, and
this is consistent with the findings of Goddard et al. (2004), Iannotta et al. (2007), and Shim (2010), but
contravene the findings of Berger (1995) and Demirgüç-Kunt and Huizinga (2000). The capital
coefficient is around 0.107–0.285, implying a 1% increase in capital raises profitability by 0.11–0.28%.
13
Thanks to the reviewer’s great suggestion, this study will conduct further analysis on explicit deposit insurance in the next
section.
14
Kennedy (2008) indicates that multicollinearity is a critical problem when the correlation is above 0.80, which is not the
case here.
Table 3
Summary of variables, descriptions, and data sources.
Classification Variable Descriptions Sources

Dependent variables
Profitability ROA Return on assets Bankscope
ROE Return on equities Bankscope
NIM Net interest margin Bankscope
NR Net interest revenue against average assets Bankscope
Risk VROA Variance of ROA is calculated using the overlapping Calculated
ROA data averaged every three years by authors
VROE Variance of ROE is calculated using the overlapping Calculated
ROE data averaged every three years by authors
LLR Loan loss reserves Bankscope
Capital CP Equity-to-total-assets Bankscope
Bank control LLGL Loan loss reserve to gross loans Bankscope
variables NITA Net loans to total assets Bankscope
LADSF Liquid assets to customer and short-term deposits Bankscope
Macro control INFL Inflation WDI
variables GW GDP growth rate WDI
DCPS Domestic credit to private sector WDI
RIR Real interest rate WDI
Financial market Capital This variable is determined by adding 1 if the Bank Regulation
regulations requirements answer is yes to questions 1–6 and 0 otherwise, and and Supervision
(CAPR) the opposite occurs for questions 7 and 8 (i.e., Database,
yes ¼ 0, no ¼ 1). The questions are: (1) Is the World Bank;
minimum required capital asset ratio (risk Barth et al., 2001,
weighted) in line with Basel guidelines? (2) Does 2006, 2008)
the ratio vary with market risk? (3–5) Before
determining minimum capital adequacy, are any of
the following deducted from the book value of
capital? (a) Market value of loan losses not realized
on the financial statements, (b) unrealized losses on
securities portfolios, and (c) unrealized foreign
exchange losses. (6) Have regulatory/supervisory
authorities verified the sources of funds to be used
as capital? (7) Can assets other than cash or
government securities provide the initial or
subsequent injections of capital? (8) Can borrowed
funds provide the initial disbursement of capital?
Thus, CAPR is an index of capital requirements that
accounts for both initial and overall capital
stringency. CAPR takes values between 0 and 8, with
higher values indicating greater capital stringency.
Supervisory This variable is determined by adding 1 if the Bank Regulation
power (SPR) answer is yes and 0 otherwise, for each of the and Supervision
following 14 questions: (1) Does the supervisory Database, World
agency have the right to meet with external Bank; Barth et al.,
auditors to discuss their report without the 2001, 2006, 2008)
approval of the bank? (2) Are auditors legally
required to communicate directly to the supervisory
agency any presumed involvement of bank
directors or senior managers in illicit activities,
fraud, or insider abuse? (3) Can supervisors take
legal action against external auditors for
negligence? (4) Can the supervisory authorities
force a bank to change its internal organizational
structure? (5) Does the institution disclose off-
balance-sheet items to supervisors? (6) Can the
supervisory agency order the bank’s directors or
management to constitute provisions to cover
actual or potential losses? (7) Can the supervisory
agency suspend directors’ decisions to distribute
dividends? (8) Can the supervisory agency suspend
directors’ decisions to distribute bonuses? (9) Can

the supervisory agency suspend directors’ decisions
to distribute management fees? (10) Can the
supervisory agency supersede bank shareholder
rights and declare the bank insolvent? (11) Does
banking law allow a supervisory agency or any
other government agency (other than a court) to
suspend some or all ownership rights at a problem
bank? (12) Regarding bank restructuring and
reorganization, can the supervisory agency or any
other government agency (other than a court)
supersede shareholder rights? (13) Regarding bank
restructuring and reorganization, can the
supervisory agency or any other government
agency (other than a court) remove and replace
management? (14) Regarding bank restructuring
and reorganization, can the supervisory agency or
any other government agency (other than a court)
remove and replace directors? Thus, SPR is
a measure of the power of supervisory agencies
indicating the extent to which these authorities can
take specific actions against bank management and
directors, shareholders, and bank auditors. This
index takes values between 0 and 14 with higher
values indicating more SPR.
Market discipline This variable is determined by adding 1 if the Bank Regulation
and private answer is yes to questions 1–7 and 0 otherwise, and and Supervision
monitoring the opposite occurs for questions 8 and 9 (i.e., Database, World
(MDPM) yes ¼ 0, no ¼ 1). (1) Is subordinated debt allowed (or Bank; Barth et al.,
required) as capital? (2) Are financial institutions 2001, 2006, 2008)
required to produce consolidated accounts covering
all bank and any non-bank financial subsidiaries?
(3) Are off-balance-sheet items disclosed to the
public? (4) Must banks disclose their risk-
management procedures? (5) Are directors legally
liable for erroneous/misleading information? (6) Do
regulations require credit ratings for commercial
banks? (7) Is an external audit by a certified/
licensed auditor mandatory for banks? (8) Does
accrued, unpaid interest/principal on non-
performing loans appear on the income statement?
(9) Is there an explicit deposit-insurance protection
system? Thus, MDPM is an indicator of market
discipline and shows the degree to which banks are
forced to disclose accurate information to the public
and whether there are incentives to increase market
discipline. This index ranges between 0 and 8 with
higher values indicating greater MDPM.
Activity The score for this variable is determined on the basis Bank Regulation
restrictions of the level of regulatory restrictiveness for bank and Supervision
(ACTR) participation in: (1) securities activities, (2) Database, World
insurance activities, (3) real estate activities, and (4) Bank; Barth et al.,
bank ownership of non-financial firms. These 2001, 2006, 2008)
activities can be unrestricted, permitted, restricted,
or prohibited and receive values of 1, 2, 3, or 4,
respectively. We create an overall index by
calculating the summation value of the four
categories. ACTR ranges from 4 to 16, with higher
values indicating higher restrictions.
(continued on next page)

Institutional DEP Deposit Insurance explicit ¼ 1; otherwise 0. Demirgüç-Kunt
development et al. (2005)
GSP Shareholder protection; index scales range from La Porta et al.
0 to 5, with higher scores for higher protection (1998)
GCP Creditor protection; index scales range from 0 to 5, La Porta et al.
with higher scores for higher protection (1998)
GLE Legal efficiency is the multiple value for the rule of La Porta et al.
law and the efficiency of the judicial system, with (1998)
higher scores for higher legal efficiency.
Notes: Bankscope, version 2009; WDI: World Development Indicator, version 2009.
The Sargan and the serial-correlation tests do not reject the null hypothesis of correct specification,
which means that we have valid instruments and no serial correlation.
The findings show that different profitability variables have different results on persistence of profit.
The coefficients of ROA and ROE with one period lag are both negative at 5% significance, exhibiting that
variables such as ROA and ROE do not show persistence of profit, while NIM and NR do have persistence
of profit. Their related coefficients are significantly positive at 0.358 and 0.353. Other control variables
also perform differently. For example, the coefficients of the ratio of loan loss reserves to gross loans
(LLGL) and net loans to total assets (NITA) are significantly negative on ROA and ROE, but the same
coefficients are positive on NIM and NR. Liquid assets to customer and short-term deposits (LADSF)
have the opposite pattern. The coefficients of the GDP growth rate (GW) on different profitability
variables are all significantly positive, meaning economic development will boost banks’ profit.
Among the other control variables, the effects from LLGL and NITA on bank profit (in terms of NIM and
NR) are significantly positive, matching our expectations (but not true for ROA and ROE). The same situ-
ation goes for LADSF, which is expected to be negative with profitability (risk) earlier. The coefficients of
GW and DCPS consistently match our expectations, which are positive and negative, respectively.
Table 7 provides the empirical results when Eq. (2) is considered for the full sample, as well as the
estimation results of capital and risk. The significantly negative relationship between capital and risk is
consistently found for the whole Asian banking, which matches the findings of Jahankhani and Lynge
(1980), Brewer and Lee (1986), Karels et al. (1989), Kwan and Eisenbeis (1997), Demirgüç-Kunt and
Kane (2002), and Agusman et al. (2008), but nevertheless violates Pettway (1976), Shrieves and Dahl
(1992), Jacques and Nigro (1997), Iannotta et al. (2007), and Shim (2010). This implies that a 1%
increase in capital decreases risk by 1.44–8.45%. The Sargan and the serial-correlation tests do not
reject the null hypothesis of correct specification, which means that we have valid instruments and no
serial correlation. All of the three risk variables also present the persistence of risk. Their coefficients
are 0.561 for VROA, 0.529 for VROE, and 0.181 for LLR. The lowest persistence of risk falls on LLR,
indicating that if banks are only concerned about the problem of loan loss reserves as only one source
of risk, then it may underestimate the risk level, since LLR has the lowest persistence of risk.
As for the other control variables, LLGR is consistently positive throughout three risk proxies. The
same goes for GW, but instead with a negative relation. The other control variables have opposite
results when different risk proxies are adopted. For example, the coefficient of DCPS is significantly
negative on VROA, but positive on VROE and LLR.
In summary, for the full sample the empirical results reveal that increasing Asian banking capital
will improve banks’ profitability and decrease their risk. In terms of persistence, banks’ profit (NIM and
NR) and risk are both significantly increasing, showing that the previous period of profit and risk will be
enhanced in the next period.
Tables 8–10 demonstrate the empirical results for samples that are further classified according to
their specializations, income levels, and geographic regions, respectively. To save space we do not
report the results of other control variables. Table 8 repeats the exercise of Tables 6 and 7, but groups
banks by specialization: commercial, cooperative, investment, and other banks. Although most results
are similar to those in Tables 6 and 7, there are some exceptions. On the left-hand side of Table 8, when
the sample falls into cooperative or investment banks, the coefficient of CP on ROE changes from
CP (%) by Specialization ROA (%) by Specialization VROA by Specialization
Full Sample Commercial banks Full data Commercial banks Full data Commercial banks
Cooperation banks Investment banks Cooperation banks Investment banks Cooperation banks Investment banks
C.-C. Lee, M.-F. Hsieh / Journal of International Money and Finance 32 (2013) 251–281
Other banks Other banks Other banks
45 5.000 70.000
40 60.000
4.000
35
3.000 50.000
30
25 2.000 40.000
20 30.000
1.000
15
20.000
10 0.000
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
5 10.000
-1.000
0 0.000
-2.000
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
VROA by GNI per capita
CP (%) by GNI per capita ROA (%) by GNI per capita
Full data High- income
Full Sample High- income Full data High- income Middle- high income Middle -low income
Middle- high income Middle -low income Middle- high income Middle -low income Low-income
Low-income Low-income
12000.000
20 3.000
18 2.000 10000.000
16
14 1.000 8000.000
12 0.000
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
10 6000.000
-1.000
8
-2.000 4000.000
6
4 -3.000 2000.000
2 -4.000
0 0.000
-5.000
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
Fig. 1. Capital, profitability, and risk trends in classification of specialization and GNI per capita (1994–2008).
265
266
CP (%) by IMF Income Classification ROA (%) by IMF Income Classification VROA by IMF Income Classification
Full Sample Advanced Countries Developing Countries Full data Advanced Countries Developing Countries Full data Advanced Countries Developing Countries
20 3.000 60.000
18 2.500
16 50.000
2.000
14 40.000
1.500
12
10 1.000 30.000
8 0.500
6 20.000
0.000
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
4 -0.500 10.000
2
-1.000
0 0.000
-1.500
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
CP (%) by OECD Income Classification ROA (%) by OECD Income Classification VROA by OECD Income Classification
Full Sample High Income Non- High Income Full data High Income Non- High Income Full data High Income By OECD Non- High Income By OECD
20 2.500 50.000
18 2.000 45.000
16 40.000
14 1.500 35.000
12 1.000 30.000
10 0.500 25.000
8 20.000
0.000
6 15.000
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
4 -0.500 10.000
2 -1.000 5.000
0 0.000
-1.500
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
Fig. 2. Capital, profitability, and risk trends in classification of IMF and OECD income (1994–2008).
Table 4
Variable means over the sample period for different types of banks (1994–2008).
Variable Commer- Cooperation Investment Other High- Middle- Middle- Low- Middle Far Advanced Develop- High- Non- high- Total
cial banks banks banks banks income high- low- income East East & countries ing income income
market income income market Central by IMF countries by OECD by OECD
market market Asian by IMF
CP (%) 10.764 5.462 27.934 22.095 11.136 14.365 14.749 10.973 19.050 10.957 9.323 15.104 6.694 16.016 12.075
ROA (%) 0.791 0.029 1.857 1.837 0.496 1.316 1.041 1.235 2.317 0.529 0.160 1.455 0.018 1.357 0.774
ROE (%) 8.363 0.122 6.243 7.454 1.696 12.001 10.879 13.018 13.886 4.627 0.374 12.026 0.852 10.871 5.898
NIM (%) 3.415 2.081 3.247 4.554 2.363 3.923 4.703 3.918 3.810 3.043 2.226 4.182 2.185 3.865 3.149
NR (%) 2.934 1.981 1.861 3.051 2.017 2.870 3.930 3.052 3.150 2.499 1.899 3.361 1.889 3.109 2.589
LLR(th $) 103541.3 1961.6 29853.45 181322.2 59891.53 128040.8 99167.7 60702.0 101717.2 70905.3 54717.5 97575.4 37225.3 103662.1 75156.83
VROA 12.259 1.072 21.265 15.740 4.749 9.067 35.417 3.725 5.555 11.152 4.159 13.378 2.105 16.640 10.364
VROE 2005.59 600.842 708.551 2367.704 1138.178 1017.182 3092.143 1755.79 384.903 1745.637 1228.65 1922.11 1425.98 1650.87 1553.552
LLGL (%) 5.427 3.387 9.564 7.183 4.004 9.166 6.194 6.356 9.901 4.504 3.329 7.322 3.159 6.746 5.264
NITA (%) 53.691 52.750 33.830 52.777 53.034 43.056 52.929 49.111 42.948 52.918 54.067 48.781 55.031 48.961 51.552
LADSF(%) 9.944 0.449 51.761 33.022 17.453 17.035 18.402 13.717 28.302 15.275 13.434 20.781 8.848 23.039 17.480
INFL (%) 16.438 0.252 4.929 8.027 0.822 2.738 36.505 7.878 5.431 9.774 0.732 20.510 0.484 16.480 9.439
GW (%) 4.931 1.381 4.277 4.609 2.427 4.859 5.981 6.011 4.792 3.729 2.152 5.775 1.545 5.500 3.847
DCPS (%) 82.225 192.826 105.195 83.992 154.963 113.091 57.456 28.774 57.218 123.207 163.555 59.988 185.322 64.673 115.760
RIR (%) 5.460 3.162 4.527 4.694 3.679 6.368 4.999 7.534 7.118 4.338 3.614 5.919 3.313 5.649 4.590
CAPR 4.872 5.583 4.922 4.798 5.320 4.591 4.205 5.619 5.798 4.979 5.289 4.804 5.465 4.772 5.067
SPR 10.839 11.945 10.640 11.135 11.447 11.293 10.710 10.279 11.660 11.087 11.333 10.931 11.780 10.681 11.149
MDPM 5.335 5.792 5.805 5.815 5.972 5.225 5.016 4.745 5.829 5.540 5.905 5.176 5.813 5.392 5.571
ACTR 11.178 12.192 10.239 10.295 11.203 10.574 11.537 11.480 10.105 11.385 11.421 11.041 12.061 10.642 11.247
DEP 0.656 0.990 0.656 0.658 0.805 0.923 0.617 0.573 0.473 0.784 0.850 0.630 0.979 0.582 0.746
GSP 3.485 4.002 3.570 3.623 3.862 4.000 2.172 5.000 2.143 3.735 3.862 3.260 3.826 3.532 3.701
GCP 2.967 2.026 3.069 2.855 2.359 4.000 2.757 4.000 4.000 2.603 2.358 3.359 2.119 3.315 2.620
GLE 53.246 88.950 62.241 58.117 82.199 61.02 14.715 33.36 43.688 68.745 82.198 29.946 84.507 46.088 68.209
Bank no. 1,136 612 288 240 1,314 195 467 300 277 1,999 1,198 1,078 942 1,334 2,276
Notes: All values are sample means. CP: Equity-to-total-assets; ROA: Return on assets; ROE: Return on equities; NIM: Net interest margin; NR: Net interest revenue against average assets;
VROA (VROE): Variance of ROA (ROE) is calculated using the overlapping ROA (ROE) data averaged every three years; LLR: Loan loss reserves; LLGL: loan loss reserves to gross loans; NITA: Net
loans to total assets; LADSF: Liquid assets to customers and short-term deposits; INFL: Inflation; GW: GDP growth rate; DCPS: Domestic credit to private sector; RIR: Real interest rate. CAPR:
Capital requirements; SPR: Supervisory power; MDPM: Market discipline and private monitoring; ACTR: Activity restrictions; DEP: Dummy variable, one for explicit deposit insurance; GSP:
Shareholder protection; GCP: Creditor protection; GLE: Legal efficiency.
267
268
Table 5
Pearson correlation coefficients.
CP ROA ROE NIM NR LLR VROA VROE LLGL NITA LADSF INFL GW DCPS RIR
CP 1
ROA 0.383** 1
(0.000)
ROE 0.000 0.120** 1
(0.894) (0.000)
NIM 0.127** 0.138** 0.031** 1
(0.000) (0.000) (0.000)
NR 0.201** 0.300** 0.080** 0.738** 1
(0.000) (0.000) (0.000) (0.000)
LLR 0.087** 0.101** 0.010 0.043** 0.054** 1
(0.000) (0.000) (0.196) (0.000) (0.000)
VROA 0.120** 0.258** 0.092** 0.048** 0.121** 0.072** 1
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
VROE 0.081** 0.115** 0.043** 0.019* 0.044** 0.031** 0.119** 1
(0.000) (0.000) (0.000) (0.019) (0.000) (0.000) (0.000)
LLGL 0.085** 0.144** 0.006 0.030** 0.016 0.113** 0.213** 0.136** 1
(0.000) (0.000) (0.462) (0.000) (0.051) (0.000) (0.000) (0.000)
NITA 0.281** 0.090** 0.042** 0.020** 0.080** 0.069** 0.027** 0.025** 0.299** 1
(0.000) (0.000) (0.000) (0.005) (0.000) (0.000) (0.001) (0.002) (0.000)
LADSF 0.436** 0.148** 0.032** 0.022 0.032** 0.047** 0.000 0.024 0.131** 0.326** 1
(0.000) (0.000) (0.005) (0.051) (0.005) (0.000) (0.998) (0.060) (0.000) (0.000)
INFL 0.001 0.003 0.129** 0.013 0.041** 0.053** 0.179** 0.066** 0.142** 0.021** 0.025* 1
(0.931) (0.687) (0.000) (0.066) (0.000) (0.000) (0.000) (0.000) (0.000) (0.004) (0.031)
GW 0.167** 0.209** 0.104** 0.064** 0.155** 0.022** 0.048** 0.027** 0.027** 0.064** 0.074** 0.116** 1
(0.000) (0.000) (0.000) (0.000) (0.000) (0.006) (0.000) (0.001) (0.001) (0.000) (0.000) (0.000)
DCPS 0.221** 0.152** 0.119** 0.086** 0.238** 0.123** 0.053** 0.018* 0.164** 0.150** 0.166** 0.089** 0.427** 1
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.030) (0.000) (0.000) (0.000) (0.000) (0.000)
RIR 0.049** 0.088** 0.004 0.054** 0.137** 0.002 0.061** 0.001 0.078** 0.083** 0.041** 0.099** 0.110** 0.203** 1
(0.000) (0.000) (0.579) (0.000) (0.000) (0.821) (0.000) (0.883) (0.000) (0.000) (0.001) (0.000) (0.000) (0.000)
Notes: Numbers in parentheses are p-values. ** and * indicate the 1% and 5% significant levels, respectively. CP: Equity-to-total-assets; ROA: Return on assets; ROE: Return on equities; NIM:
Net interest margin; NR: Net interest revenue against average assets; VROA (VROE): Variance of ROA (ROE) is calculated using the overlapping ROA (ROE) data averaged every three years;
LLR: Loan loss reserves; LLGL: loan loss reserves to gross loans; NITA: Net loans to total assets; LADSF: Liquid assets to customers and short-term deposits; INFL: Inflation; GW: GDP growth
rate; DCPS: Domestic credit to private sector; RIR: Real interest rate.
Table 6
Full sample: Estimation results of capital and profitability (result of Eq. (1)).
(1) (2) (3) (4)

ROAit1 0.009** (0.000)
ROEit1 0.153** (0.000)
NIMit1 0.358** (0.000)
NRit1 0.353** (0.000)
CPit 0.285** (0.000) 0.582** (0.000) 0.135** (0.000) 0.107** (0.000)
LLGLit 0.153** (0.000) 0.209** (0.000) 0.011** (0.000) 0.001** (0.000)
NITAit 0.021** (0.000) 0.796** (0.000) 0.011** (0.000) 0.014** (0.000)
LADSFit 0.004** (0.000) 0.003** (0.000) 0.004** (0.000) 0.002** (0.000)
INFLit 0.186** (0.000) 1.802** (0.000) 0.024** (0.000) 0.014** (0.000)
GWit 0.108** (0.000) 1.950** (0.000) 0.020** (0.000) 0.011** (0.000)
DCPSit 0.008** (0.000) 0.001 (0.576) 0.003** (0.000) 0.001** (0.000)
RIRit 0.014** (0.000) 0.394** (0.000) 0.009** (0.000) 0.054** (0.000)
Sargan test (p-value) 0.946 0.958 0.067 0.059
Serial correlation test (p-value) 0.103 0.158 0.071 0.080
Number of banks 738 738 734 734
Observations 4031 4005 4017 4017
Notes: Dependent variable is profitability: ROA, ROE, NIM and NR, respectively. Estimation method is the two-step GMM dynamic
panel estimator. ** and * indicate significance at the 5% and 10% levels, respectively. The regressions include dummy variables for
the different time periods that are not reported. The null hypothesis of the Sargan test (or Hansen test) is that the instruments
used are not correlated with residuals (over-identifying restrictions). The null hypothesis of the serial correlation test is that the
errors exhibit no second-order serial correlation. P-values are in parentheses. ROA: Return on assets; ROE: Return on equities;
NIM: Net interest margin; NR: Net interest revenue against average assets; LLR: Loan loss reserves; CP: Equity-to-total-assets;
LLGL: loan loss reserves to gross loans; NITA: Net loans to total assets; LADSF: Liquid assets to customers and short-term
deposits; INFL: Inflation; GW: GDP growth rate; DCPS: Domestic credit to private sector; RIR: Real interest rate.
significantly negative to significantly positive when compared with the benchmark model (Table 6).
Investment banks have the lowest positive capital effect on NIM and NR at 0.040 and 0.019, respec-
tively. By contrast, the group of other banks has the highest and positive capital effect on ROE at 1.206.
In regards to the right-hand side of Table 8, commercial banks reveal the highest reverse capital effect
on VROA at 9.956, with the lowest one falling in the group of cooperative banks at 0.004. Hence, the
Table 7
Full sample: estimation results of capital and risk (result of Eq. (2)).
(1) (2) (3)

VROAit1 0.561** (0.000)
VROEit1 0.529** (0.000)
LLRit1 0.181** (0.000)
CPit 8.450** (0.000) 3.394** (0.000) 1.440** (0.000)
LLGLit 1.592** (0.000) 0.072** (0.000) 1.021** (0.000)
NITAit 0.337** (0.000) 0.679** (0.000) 2.490** (0.000)
LADSFit 0.347** (0.000) 0.090** (0.000) 0.009** (0.000)
INFLit 2.734** (0.000) 1.067** (0.000) 1.121** (0.000)
GWit 1.346** (0.000) 1.821** (0.000) 0.048** (0.000)
DCPSit 0.268** (0.000) 0.196** (0.000) 1.651** (0.000)
RIRit 1.697** (0.000) 1.456** (0.000) 0.063** (0.000)
Sargan test (p-value) 0.104 0.378 0.062
Serial correlation test (p-value) 0.533 0.602 0.135
Number of banks 598 596 699
Observations 3265 3229 3744
Notes: Dependent variable is risk: VROA, VROE, and LLR, respectively. The estimation method is the two-step GMM dynamic
panel estimator. ** and * indicate significance at the 5% and 10% levels, respectively. The regressions include dummy variables for
the different time periods that are not reported. The null hypothesis of the Sargan test is that the instruments used are not
correlated with residuals (over-identifying restrictions). The null hypothesis of the serial correlation test is that the errors exhibit
no second-order serial correlation. P-values are in parentheses. VROA (VROE): Variance of ROA (ROE) is calculated using the
overlapping ROA (ROE) data averaged every three years; LLR: Loan loss reserves; CP: Equity-to-total-assets; LLGL: loan loss
reserves to gross loans; NITA: Net loans to total assets; LADSF: Liquid assets to customers and short-term deposits; INFL:
Inflation; GW: GDP growth rate; DCPS: Domestic credit to private sector; RIR: Real interest rate.
Table 8
Different bank types: estimation results of capital, profitability, and risk.
Profitability Risk
(1) (2) (3) (4) (5) (6) (7)

Commercial banks
ROAit1 0.151**
(0.000)
ROEit1 0.168**
(0.000)
NIMit1 0.311**
(0.000)
NRit1 0.301**
(0.000)
VROAit1 0.573**
(0.000)
VROEit1 0.546**
(0.000)
LLRit1 0.149**
(0.000)
CPit 0.435** 1.209** 0.160** 0.137** 9.956** 3.195** 2.516**
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Number of banks 531 531 530 530 443 441 516
Observations 3111 3088 3106 3106 2605 2574 2927
Cooperative banks
ROAit1 0.308**
(0.000)
ROEit1 0.786**
(0.006)
NIMit1 0.343**
(0.000)
NRit1 0.238**
(0.000)
VROAit1 0.295**
(0.000)
VROEit1 0.175**
(0.000)
LLRit1 1.454**
(0.000)
CPit 0.186** 0.152** 0.116** 0.080** 0.004 2.776** 1.234**
(0.000) (0.027) (0.000) (0.000) (0.202) (0.000) (0.000)
Number of banks 34 34 34 34 33 33 34
Observations 236 236 236 236 220 200 205
Investment banks
ROAit1 0.003
(0.929)
ROEit1 0.051
(0.349)
NIMit1 0.479**
(0.000)
NRit1 0.456**
(0.000)
VROAit1 0.322**
(0.000)
VROEit1 0.870**
(0.000)
LLRit1 0.694**
(0.000)
CPit 0.050** 0.520** 0.040** 0.019** 0.241 2.055** 5.393
(0.013) (0.034) (0.000) (0.000) (0.239) (0.000) (0.257)
Number of banks 90 90 90 90 56 56 70
Observations 334 331 334 334 216 211 265
Profitability Risk
(1) (2) (3) (4) (5) (6) (7)

Other banks
ROAit1 0.053
(0.167)
ROEit1 0.242**
(0.000)
NIMit1 0.432**
(0.000)
NRit1 0.407**
(0.000)
VROAit1 0.461**
(0.000)
VROEit1 0.463**
(0.000)
LLRit1 0.709**
(0.000)
CPit 0.112** 1.206** 0.063** 0.045** 0.020 1.761 0.800
(0.000) (0.047) (0.005) (0.000) (0.941) (0.359) (0.603)
Number of banks 83 83 80 80 66 66 79
Observations 350 350 341 341 244 244 347
Notes: Dependent variable is profitability and risk, respectively. The estimation method is the two-step GMM dynamic panel
estimator. ** and * indicate significance at the 5% and 10% levels, respectively. The regressions include dummy variables for the
different time periods that are not reported. The null hypothesis of the Sargan test is that the instruments used are not correlated
with residuals (over-identifying restrictions). The null hypothesis of the serial correlation test is that the errors exhibit no
second-order serial correlation. P-values are in parentheses. To save space we do not report the results of other control variables.
ROA: Return on Assets; ROE: Return on equities; NIM: Net interest margin; NR: Net interest revenue against average assets; VROA
(VROE): Variance of ROA (ROE) is calculated using the overlapping ROA (ROE) data averaged every three years; LLR: Loan loss
reserves; CP: Equity-to-total-assets.
capital effect is impacted by different variables of profitability and risk, as well as different bank
categories.
Table 8 finds that investment banks have the highest profitability persistence, in terms of NIM and
NR, but have the lowest value in term of ROA. As for the persistence of risk, cooperative banks occupy
the highest value of LLR, whereas the lowest value falls into the group of commercial banks. Therefore,
the persistence of risk for LLR differs among different bank categories.
Table 9 classifies countries by income levels: high, upper-middle, lower-middle, and low income.
The study finds that, in the group of low-income countries, the relationship between CP and ROE
becomes significantly positive, partially confirming the finding of Hsieh and Lee (2010), who note that
countries with a low-income level may strengthen the positive relationship between concentration
and profit. Capital has the greatest positive impact on ROE for banks in upper-middle income countries,
but banks in high-income countries have the lowest positive capital effect on NR. Hence, in low-income
countries, banks’ capital has higher impacts on profitability.
For the persistence of profit, banks in high-income countries enjoy the highest value (in terms of
NR), while banks in lower-income countries have the lowest one (in terms of NIM). Referring to the
persistence of risk, banks in upper-middle income countries perform at the largest value (0.971 for
ROE), while the lowest value belongs to banks in lower-middle income countries (0.171 for LLR). Hence,
profit and persistence of risk do vary based on different income levels.
On the right-hand side of Table 9, banks in lower-middle income countries have the highest and
positive capital effect on risk (the coefficient of VROA is 0.577), but the lowest values, in terms of VROA
and VROE, are for banks in high-income countries. This shows evidence that banks in high-income
countries probably have a more sound financial supervisory and technology. Thus, their risk may be
lower than those for developing countries and emerging markets (Claessens, 2009). The coefficients of
capital on risk for lower-middle and upper-middle income countries are both negative and reach a 5%
significance level, but the same figure is significantly positive for low-income countries. This inter-
esting finding is probably due to (lower- and upper-) middle-income countries, compared with low-
income countries, having a much higher proportion for explicit deposit insurance, i.e. 50% and 75%
Table 9
Different income levels: estimation results of capital, profitability, and risk.
Profitability Risk
(1) (2) (3) (4) (5) (6) (7)

High-income countries
ROAit1 0.025
(0.202)
ROEit1 0.176**
(0.000)
NIMit1 0.241**
(0.000)
NRit1 0.487**
(0.000)
VROAit1 0.418**
(0.000)
VROEit1 0.489**
(0.000)
LLRit1 0.465**
(0.000)
CPit 0.110** 0.306 0.031** 0.012** 0.416** 0.452** 0.101
(0.000) (0.478) (0.000) (0.000) (0.000) (0.000) (0.347)
Number of banks 334 334 333 333 304 304 303
Observations 1990 1984 1986 1986 1747 1738 1745
Upper-middle income countries
ROAit1 0.048**
(0.000)
ROEit1 0.377**
(0.000)
NIMit1 0.362**
(0.000)
NRit1 0.315**
(0.000)
VROAit1 0.470**
(0.000)
VROEit1 0.971**
(0.000)
LLRit1 0.532**
(0.000)
CPit 0.071** 0.964** 0.061** 0.045** 0.002 1.423** 2.178**
(0.000) (0.000) (0.000) (0.000) (0.954) (0.000) (0.000)
Number of banks 81 81 78 78 59 59 81
Observations 405 405 396 396 315 315 402
Lower-middle income countries
ROAit1 0.167**
(0.000)
ROEit1 0.104**
(0.000)
NIMit1 0.288**
(0.000)
NRit1 0.286**
(0.000)
VROAit1 0.577**
(0.000)
VROEit1 0.558**
(0.000)
LLRit1 0.171**
(0.000)
CPit 0.455** 1.139** 0.185** 0.154** 9.616** 2.633** 3.547**
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Number of banks 277 277 277 277 199 197 273
Observations 1466 1466 1465 1465 1060 1033 1446
Profitability Risk
(1) (2) (3) (4) (5) (6) (7)

Lower-income countries
ROAit1 0.321**
(0.001)
ROEit1 0.288**
(0.000)
NIMit1 0.217**
(0.004)
NRit1 0.177
(0.107)
VROAit1 0.035
(0.569)
VROEit1 0.477**
(0.000)
LLRit1 0.906**
(0.000)
CPit 0.038* 0.035 0.058** 0.051** 0.017 3.077** 1.127**
(0.064) (0.948) (0.000) (0.000) (0.482) (0.010) (0.042)
Number of banks 46 46 46 46 36 36 42
Observations 170 170 170 170 143 143 151
ROA: Return on assets; ROE: Return on equities; NIM: Net interest margin; NR: Net interest revenue against average assets; VROA
vs. 18%, respectively. The empirical results indicate the dominance of incentives for exploiting deposit
insurance subsidies, which is consistent with the hypothesis of Shrieves and Dahl (1992). Hence, the
capital effects of profitability and risk do have different impacts by different country income levels.
Table 10 provides the empirical results for samples further classified based on their geographic
regions. It is found that banks in Middle Eastern countries own the highest and positive capital effect on
ROE (0.405), while Asian banks have the highest value for ROA (0.403). However, at the same time the
lowest value of the capital effect on profitability falls in Middle Eastern countries in terms of NIM and
NR, i.e. 0.038 and 0.032, respectively. Next to risk, on the right-hand side of Table 10, banks in Asia have
the largest reverse capital effect on VROE and LLR, while the lowest value of VROA is in Middle Eastern
countries. This offers evidence that banks in Middle Eastern countries are probably accompanied by the
lowest risk level due to the historical origin of Islam as the major religion there. The provisions of
Islamic precepts note that banks cannot receive interest for no reason, nor can they invest money for
the use of gambling and other activities. Thus, the level of risk-taking is limited by the principle of the
religion in the Middle East. Worth noting is that banks in the Middle East not only have the greatest
persistence of risk, but also show persistence of profit, meaning that banks in this area either enjoy
dominant competition or operate under information opacity (Berger et al., 2000). Finally, all tables pass
a test for second-order serial correlation and the Sargan test for over-identifying restrictions, con-
firming that our instruments are valid.
4.2. Robustness analysis
To check the robustness of the empirical results, we consider the following modifications of our
empirical model.15 First, we consider four variables on financial market regulations (CAPR, SPR, MDPM,
and ACTR) and three for institutional developments (GSP GCP, and GLE), including the existence of
15
Thanks to the reviewer and the editor for the great suggestions.
Table 10
Different geographical regions: estimation results of capital, profitability, and risk.
Profitability Risk
(1) (2) (3) (4) (5) (6) (7)

Far East and Central Asian countries
ROAit1 0.142**
(0.000)
ROEit1 0.133**
(0.000)
NIMit1 0.320**
(0.000)
NRit1 0.293**
(0.000)
VROAit1 0.544**
(0.000)
VROEit1 0.524**
(0.000)
LLRit1 0.199**
(0.000)
CPit 0.403** 1.040** 0.174** 0.144** 9.527** 3.315** 3.374**
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Number of banks 663 663 659 659 532 530 630
Observations 3499 3474 3485 3485 2844 2809 3233
Middle Eastern countries
ROAit1 0.253**
(0.000)
ROEit1 0.182**
(0.000)
NIMit1 0.523**
(0.000)
NRit1 0.524**
(0.000)
VROAit1 0.160**
(0.000)
VROEit1 0.895**
(0.000)
LLRit1 0.591**
(0.000)
CPit 0.093** 0.405** 0.038** 0.032** 0.038** 1.463** 5.351**
(0.000) (0.000) (0.000) (0.000) (0.002) (0.000) (0.000)
Number of banks 75 75 75 75 66 66 69
Observations 532 531 532 532 421 420 511
deposit insurance (DEP). Second, in addition to GNI per capita, this study also considers the other two
different definitions of economic levels by the IMF and OECD. Third, this study takes into account the
sub-period and the Asian crisis that occurred in 1998 by separating the sample into before the Asian
crisis (Non-Crisis period), and after the Asian crisis (Post-Crisis period).
Table 11 presents the estimation results after implementing the regulatory factors. Compared with
the benchmark results in Tables 6 and 7, the relationship between either capital and profitability or
between capital and risk presents the exact identical direction - that is, there is a significantly positive
relationship between capital and profitability and a negative one for capital and risk. Nevertheless,
taking regulatory variables into consideration will increase (decrease) the positive persistence of
Table 11
Full sample: involving factors of financial market regulations and institutional developments.
Profitability
(1) (2) (3) (4) (5) (6) (7) (8)

ROAit1 0.086** (0.000) 0.173** (0.000)
ROEit1 0.074** (0.000) 0.098 (0.584)
NIMit1 0.388** (0.000) 0.248** (0.000)
NRit1 0.443** (0.000) 0.363** (0.000)
CPit 0.129** (0.000) 0.236** (0.001) 0.178** (0.000) 0.076** (0.000) 0.171** (0.000) 0.557** (0.000) 0.077** (0.000) 0.045** (0.000)
CAPR 0.363** (0.021) 0.134** (0.000) 0.231 (0.141) 0.026 (0.778) 0.046 (0.623) 0.403** (0.022) 0.088 (0.127) 0.044 (0.315)
SPR 0.541** (0.018) 0.135** (0.014) 0.596** (0.000) 0.283** (0.001) 0.090 (0.447) 0.517** (0.004) 0.096 (0.168) 0.143** (0.008)
MDPM 0.104 (0.575) 0.121** (0.005) 0.247 (0.212) 0.130 (0.169) 0.091 (0.157) 0.354** (0.035) 0.040 (0.336) 0.002 (0.964)
ACTR 0.416** (0.007) 0.055** (0.017) 0.011 (0.933) 0.142* (0.066) 0.137** (0.045) 0.383** (0.007) 0.096** (0.017) 0.077** (0.015)
DEP 0.221** (0.000) 0.374** (0.000) 0.379 (0.299) 0.058 (0.842)
GSP 0.394 (0.151) 0.113** (0.030) 0.813** (0.000) 0.493** (0.000)
GCP 0.215 (0.137) 0.391** (0.049) 0.099 (0.230) 0.047 (0.472)
GLE 0.005 (0.704) 0.004* (0.058) 0.049** (0.000) 0.032** (0.000)
Risk
(1) (2) (3) (4) (5) (6)

VROAit1 0.266** (0.000) 0.417** (0.000)
VROEit1 0.528** (0.000) 0.438** (0.000)
LLRit1 0.090** (0.043) 0.095** (0.000)
CPit 1.471** (0.000) 1.813** (0.000) 1.305* (0.083) 0.144** (0.000) 2.169** (0.000) 0.639** (0.021)
CAPR 0.029 (0.802) 1.864** (0.000) 0.665** (0.005) 0.408 (0.339) 0.662* (0.095) 1.606* (0.051)
SPR 1.226** (0.002) 1.273** (0.006) 0.958** (0.014) 0.050 (0.910) 0.021 (0.961) 1.684** (0.011)
MDPM 0.841** (0.000) 2.588 (0.248) 0.436 (0.409) 0.294 (0.562) 0.376 (0.492) 2.973** (0.002)
ACTR 0.339 (0.136) 1.369** (0.000) 0.436 (0.150) 1.117** (0.001) 0.747** (0.032) 2.833 (0.693)
DEP 0.359 (0.866) 1.531 (0.501) 1.109 (0.783)
GSP 1.710 (0.241) 2.846** (0.035) 1.575 (0.507)
GCP 1.900** (0.002) 1.547** (0.006) 1.571 (0.154)
GLE 0.041 (0.531) 0.992* (0.083) 1.019 (0.243)
Notes: Dependent variable is profitability and risk, respectively. The estimation method is the two-step GMM dynamic panel estimator. ** and * indicate significance at the 5% and 10% levels,
respectively. The regressions include dummy variables for the different time periods that are not reported. The null hypothesis of the Sargan test is that the instruments used are not
correlated with residuals (over-identifying restrictions). The null hypothesis of the serial correlation test is that the errors exhibit no second-order serial correlation. P-values are in
parentheses. To save space we do not report the results of other control variables. ROA: Return on assets; ROE: Return on equities; NIM: Net interest margin; NR: Net interest revenue against
average assets; VROA (VROE): Variance of ROA (ROE) is calculated using the overlapping ROA (ROE) data averaged every three years; LLR: Loan loss reserves; CP: Equity-to-total-assets. CAPR:
Capital requirements; SPR: Supervisory power; MDPM: Market discipline and private monitoring; ACTR: Activity restrictions; DEP: Deposit insurance explicit; GSP: Shareholder protection;
GCP: Creditor protection; GLE: Legal efficiency.
275
Table 12
Different economic levels: estimation results of capital, profitability, and risk.
Profitability Risk
(1) (2) (3) (4) (5) (6) (7)

High-income countries
ROAit1 0.063**
(0.002)
ROEit1 0.065**
(0.001)
NIMit1 0.182**
(0.000)
NRit1 0.562**
(0.000)
VROAit1 0.367**
(0.000)
VROEit1 0.604**
(0.000)
LLRit1 0.223**
(0.000)
CPit 0.130** 0.422** 0.103** 0.016** 0.153** 0.367** 0.823**
(0.000) (0.018) (0.000) (0.000) (0.000) (0.000) (0.000)
Non-high-income countries
ROAit1 0.124**
(0.000)
ROEit1 0.032*
(0.065)
NIMit1 0.445**
(0.000)
NRit1 0.434**
(0.000)
VROAit1 0.441**
(0.000)
VROEit1 0.553**
(0.000)
LLRit1 0.140**
(0.000)
CPit 0.130** 0.202** 0.175** 0.116** 0.016** 1.181** 1.618**
(0.000) (0.011) (0.000) (0.000) (0.000) (0.000) (0.000)
Developing countries
ROAit1 0.118**
(0.000)
ROEit1 0.106**
(0.000)
NIMit1 0.457**
(0.000)
NRit1 0.435**
(0.000)
VROAit1 0.436**
(0.000)
VROEit1 0.554**
(0.000)
LLRit1 0.129**
(0.000)
CPit 0.122** 0.377 0.181** 0.131** 0.026** 1.217** 4.829**
(0.000) (0.209) (0.000) (0.000) (0.000) (0.000) (0.000)
profitability (risk). As shown, countries with greater capital stringency, or higher creditor protection, or
higher legal efficiency will decrease banks’ profitability and increase their risk. On the other hand,
banks’ profitability will improve for countries with higher supervisory power, a higher degree of
market discipline and private monitoring, or countries with greater regulatory restrictions on
commercial bank activities and higher shareholder protection. Particularly, countries with greater
regulatory restrictions and higher shareholder protection enhance banks’ profitability, but at the same
time reduce their risk. Noteworthy is that according to Demirgüç-Kunt and Kane (2002), under explicit
deposit insurance schemes, banks will have more incentives for risk-taking, which is potentially
supported in this study, but does not reach any statistical significance.
Table 12 indicates the empirical results using different economic levels that are defined by the OECD
(high-income and non-high-income) and the IMF (developing countries). Compared with the bench-
mark results shown on Table 9, it is found that the empirical results between capital and profitability
Table 13
Sub-period of crisis: estimation results of capital, profitability, and risk.
Profitability Risk
(1) (2) (3) (4) (5) (6) (7)

Post-Crisis period
ROAit1 0.103**
(0.001)
ROEit1 0.105**
(0.000)
NIMit1 0.385**
(0.000)
NRit1 0.413**
(0.000)
VROAit1 0.044*
(0.090)
VROEit1 0.532**
(0.000)
LLRit1 0.046*
(0.060)
CPit 0.188** 5.275** 0.193** 0.073** 0.107** 3.813** 3.137**
(0.000) (0.001) (0.000) (0.001) (0.023) (0.000) (0.016)
Non-Crisis period
ROAit1 0.346**
(0.045)
ROEit1 0.107**
(0.006)
NIMit1 0.928**
(0.000)
NRit1 0.827**
(0.000)
VROAit1 0.271**
(0.000)
VROEit1 0.201**
(0.582)
LLRit1 0.758**
(0.000)
CPit 0.071** 0.028 0.016** 0.015** 0.429* 0.338** 1.213**
(0.000) (0.878) (0.001) (0.000) (0.069) (0.033) (0.025)
(risk) in high-income countries defined by the OECD are mostly the same with that defined by the
World Bank - that is, except for ROE, there is a positive relationship between capital and profitability
and a negative one between capital and risk. Furthermore, the results of non-high-income countries by
the OECD and developing countries by the IMF are consistent with the grouping of upper-middle
income countries by the World Bank.
Table 13 presents the empirical results after considering the impacts from the Asian crisis. During
the Post-Crisis period, banks’ capital can generate an increase in bank profitability and a decline in bank
risk versus that during the Pre-Crisis period. Recall that investment banks have a higher level of capital
versus the others and are more aggressive after the Asian crisis (Fig. 1). Thus, not only is SCP supported,
but also the financial accelerator effect. The financial accelerator effect, first proposed by Bernanke
(1983), explains the behavior of bank credit and its relationship with the persistence and amplitude
of cyclical fluctuations in the economy. The link of the financial accelerator mechanism between the
real economy and financial markets stems from firms’ need for external finance in order to engage in
profitable investment opportunities. On the other hand, firms’ ability to borrow largely depends on the
market value of their financial and tangible assets. Therefore, banks with a higher level of capital
should be able to sustain themselves during the crisis period and engage in profitable investment
opportunities.
5. Conclusion
This article examines the relationship between bank capital and profit (risk), using bank-level data
for 42 Asian banks with the latest and a wider range of panel data that cover 2,276 banks over the
period from 1994 to 2008. We ask two critical questions: “Does the Asian dataset fit the SCP, Moral
Hazard, or Regulatory hypotheses?” and “Do different income levels, bank categories, or regions
influence the results of the SCP, Moral Hazard, or Regulatory hypotheses?”
To investigate the robustness of the empirical results, we also consider the following modifications
of our empirical model. First, we consider four variables on financial market regulations and four
institutional developments, including the existence of deposit insurance. Second, in addition to GNI per
capita, this study takes into account the other two different definitions of economic levels by the IMF
and OECD. Third, the sub-period and Asian crisis that occurred in 1998 are investigated herein by
separating the sample into before the Asian crisis (Non-Crisis period), and after the Asian crisis (Post-
Crisis period). The main results appear to differ according to the types of banks, the countries’
economic development, and across regions.
This study applies recent two-step system GMM dynamic panel data techniques. The empirical
results indicate that the effect of increasing bank capital on profit (risk) is significantly positive
(negative), supporting the hypotheses of SCP and moral hazard. It is also found that different profit-
ability variables have different results on the persistence of profit. More specifically, ROA and ROE do
not show persistence of profit, but NIM and NR do. Hence, persistence of profit is greatly affected by
different profitability variables. At the same time, all of the three risk variables have persistence: VROA
has the largest persistence of risk, followed by VROE and LLR.
The extant literature, which ignores influence factors, notes that the impact of bank capital on
profitability (risk) is ambiguous. Thus, in consideration of the effects of the influencing factors, three
conclusions are reached herein. First, investment banks have the lowest positive capital effect on
NIM and NR, while commercial banks reveal the highest reverse capital effect on VROA. Investment
banks have the highest profitability persistence in terms of NIM and NR. As for the persistence of
risk, cooperative banks occupy the highest value of LLR, with the lowest value falling into the group
of commercial banks. Therefore, the persistence of risk for LLR differs among different bank
categories.
Second, banks in upper-middle-income countries have the greatest positive capital effect on ROE,
but those in high-income countries have the lowest positive capital effect on NR. Hence, in low-income
countries banks have a higher capital effect of profitability. Banks in lower-middle income countries
also have the highest reverse capital effect on risk (VROA), but banks in high-income countries have the
lowest values (in terms of VROA and VROE). This provides evidence that banks in high-income countries
probably are associated with a more sound financial supervisory and technology. Thus, their risk may
be lower than those for developing countries and emerging markets (Claessens, 2009). Hence, the
capital effects of profitability and risk do incur different impacts by different country income levels.
Third, it is found that banks in Middle Eastern countries own the highest and positive capital effect
on ROE. Asian banks have the largest reverse capital effect on VROE and LLR, while the lowest value of
VROA is for banks in Middle Eastern countries. This means that banks there are accompanied by the
lowest risk level, probably because the Middle East is the historical origin of Islam. The provisions of
Islamic precepts note that a bank cannot receive interest for no reason, nor can it invest in gambling
and other activities. Thus, the level of risk-taking is limited by principal of the religion in the Middle
East. Worth noting is that banks in the Middle East not only have the greatest persistence of risk, but
also have the persistence of profit, meaning that banks in this area enjoy dominant competition.
Fourth, we confirm with robustness tests that our results for regulatory quality and institutional
developments are independent of differences in profitability and risk.
Fifth, we further confirm that our result for the Asian crisis is independent. It is found that during
the Post-Crisis period, bank capital can generate an increase in bank profitability and a decline in bank
risk versus that during the Pre-Crisis period. Thus, not only is SCP supported, but also is the financial
accelerator effect.
Important policy implications emerge from our empirical results. First, does increasing capital raise
profit or decrease risk? We show evidence that there is a positive (negative) relationship between capital
and profitability (risk) through the whole Asian banking sample, thus supporting both the SCP and moral
hazard hypotheses. Furthermore, the negative relationship between risk and the level of capital also
suggests that the regulator should apply closer monitoring to prohibit those banks from gambling in
excessively risky undertakings. Nevertheless, this rough conclusion counters some phenomena. For
example, banks in some countries have a strong capital structure, though their profit and risk are not
satisfied. Thus, this study further considers different profitability and risk variables, as well as different
conditions. Conducting these efforts enables us to resolve the puzzle of the extant literature, which
ignores critical factors related to banks’ specializations, income levels, and geographic regions. The
government or authorities should take these influencing factors into consideration. Second, different
profitability (risk) variables present different patterns with capital. Hence, the authorities should realize
that using a single profitability (risk) variable may result in a totally wrong policy. Specifically, invest-
ment banks are playing an important role in the Asian banking industry, since a positive persistence of
profit is found. Asian countries should improve their banking system by mending the financial efficiency
of investment banks to implement the suggestions proposed by Basel III.
Acknowledgment
We would like to thank Professor McNelis and two anonymous referees for their highly constructive
comments. The usual disclaimer applies and views are the sole responsibility of the authors. Chien-
Chiang Lee thanks the National Science Council of Taiwan for financial support.
References
Aggarwal, R., Goodell, J.W., 2009a. Markets versus institutions in developing countries: national attributes as determinants.
Emerging Markets Review 10 (1), 51–66.
Aggarwal, R., Goodell, J.W., 2009b. Markets and institutions in financial intermediation: national characteristics as determi-
nants. Journal of Banking and Finance 33 (10), 1770–1780.
Aggarwal, R., Jacques, K.T., 1998. Assessing the impact of prompt corrective action on bank capital and risk. Federal Reserve Bank
of New York Economic Policy Review, 23–32.
Agoraki, M., Delis, M.D., Pasiouras, F., 2011. Regulations, competition and bank risk-taking in transition countries. Journal of
Financial Stability 7 (1), 38–48.
Agusman, A., Monroe, G.S., Gasbarro, D., Zumwalt, J.K., 2008. Accounting and capital market measures of risk: evidence from
Asian banks during 1998–2003. Journal of Banking and Finance 32 (4), 480–488.
Akhavein, J.D., Swamy, P.V., Taubman, S.B., 1997. A general method of deriving the efficiencies of banks from a profit function.
Journal of Productivity Analysis 8 (1), 71–93.
Altunbas, Y.S., Carbo, E., Gardener, P.M., Molyneux, P., 2007. Examining the relationships between capital, risk and efficiency in
European banking. European Financial Management 13 (1), 49–70.
Arellano, M., Bover, O., 1995. Another look at the instrumental variables estimation of error-component models. Journal of
Econometrics 68 (1), 29–51.
Barth, J.R., Caprio Jr., G., Levine, R., 2001. The regulation and supervision of banks around the world: a new database. In: Litan, R.
E., Herring, R. (Eds.), Integrating Emerging Market Countries into the Global Financial System. Brookings-Wharton Papers in
Financial Services, Brooking Institution Press, pp. 183–240.
Barth, J., Caprio, G., Levine, R., 2006. Rethinking Bank Regulation: Till Angels Govern. Cambridge University Press, Cambridge.
Barth, J., Caprio, G., Levine, R., 2008. Bank regulations are changing: for better or worse? Comparative Economic Studies 50 (4),
537–563.
Beck, T., Demirguc-Kunt, A., Levine, R., 2000. A new database on financial development and structure. World Bank Economic
Review 14 (3), 597–605.
Beck, T., Levine, R., 2004. Stock markets, banks, and growth: panel evidence. Journal of Banking and Finance 28 (3), 423–442.
Berger, A.N., 1995. The relationship between capital and earnings in banking. Journal of Money, Credit and Banking 27 (2), 432–456.
Berger, A.N., Bonime, S.D., Covitz, D.M., Hancock, D., 2000. Why are bank profits so persistent? The roles of product market
competition, informational opacity, and regional/macroeconomic shocks. Journal of Banking and Finance 24 (7), 1203–1235.
Bernanke, B.S., 1983. Non-monetary effects of financial crises in the propagation of the great depression. The American
Economic Review 73 (3), 2547–2576.
Blundell, R., Bond, S., 2000. GMM estimation with persistent panel data: an application to production functions. Econometric
Reviews 19 (3), 321–340.
Brewer Jr., E., Lee, C.F., 1986. How the market Judges bank risk, economic perspectives. Federal Reserve Bank of Chicago, 25–31.
Carbo, S., Humphrey, D., Maudos, J., Molyneux, P., 2009. Cross-country comparisons of competition and pricing power in
European banking. Journal of International Money and Finance 28 (1), 115–134.
Casu, B., Girardone, C., 2006. Bank competition, concentration and efficiency in the single European market. The Manchester
School 74 (4), 441–468.
Claessens, S., 2009. Competition in the Financial Sector: Overview of Competition Policies. IMF Working Paper No. 09/45.
Claessens, S., Demirgüç-Kunt, A., Huizinga, H., 2001. How does foreign entry affect the domestic banking markets? Journal of
Banking and Finance 25 (5), 891–911.
Claessens, S., Klingebiel, D., 2001. Competition and scope of activities in financial services. World Bank Research Observer 16 (1),
18–40.
Claessens, S., Laeven, L., 2003. Competition in the financial sector and growth: a cross-country perspective. In: Good-
hart, Charles (Ed.), Financial Development and Economic Growth: Explaining the Links. Palgrave.
Claessens, S., Laeven, L., 2004. What drives bank competition? Some international evidence. Journal of Money, Credit, and
Banking 36 (3), 563–583.
Deesomsak, R., Paudyal, K., Pescetto, G., 2004. The determinants of capital structure: evidence from the Asia Pacific region.
Journal of Multinational Management 14 (4/5), 387–405.
Delis, M.D., Molyneux, P., Pasiouras, F., 2011. Regulations and productivity growth in banking: evidence from transition econ-
omies. Journal of Money, Credit and Banking 43 (4), 735–764.
Demirgüç-Kunt, A., Huizinga, H., 2000. Financial Structure and Bank Profitability. The World Bank Policy Research Working
Paper, No. 2430.
Demirgüç-Kunt, A., Kane, E., 2002. Deposit insurance around the world: where does it work? Journal of Economic Perspectives
16 (2), 175–195.
Demirgüç-Kunt, A., Karacaovali, B., Laeven, L., 2005. Deposit Insurance Around the World: A Comprehensive Database. World
Bank Policy Research Working Paper No. 3628.
Demirgüç-Kunt, A., Huizinga, H., 1999. Determinants of commercial bank interest margins and profitability: some international
evidence. World Bank Economic Review 13 (2), 379–408.
Doytch, N., Uctum, M., 2011. Does the worldwide shift of FDI from manufacturing to services accelerate economic growth? A
GMM estimation study. Journal of International Money and Finance 30 (3), 410–427.
Driffill, J., Psaradakis, Z., Sola, M., 1998. Testing the expectations hypothesis of the term structure using instrumental variables.
International Journal of Finance and Economics 3 (4), 321–325.
Fernández de Guevara, J., Maudos, J., Pérez, F., 2007. Integration and competition in European banks. Journal of International
Money and Finance 26 (1), 26–45.
Festic, M., Kavkler, A., Repina, S., 2011. The macroeconomic sources of systemic risk in the banking sectors of five new EU
member states. Journal of Banking and Finance 35 (2), 310–322.
Goddard, J., Liu, H., Molyneux, P., Wilson, J.O.S., 2010. Do bank profits converge? European Financial Management. doi:10.1111/j.
1468-036X.2010.00578.x.
Goddard, J., Molyneux, P., Wilson, J.O.S., 2004. The profitability of European banks: a cross-sectional and dynamic panel analysis.
The Manchester School 72 (3), 363–381.
Hannan, T.H., Berger, A.N., 1991. The rigidity of prices: evidence from the banking industry. American Economic Review 81 (4),
938–945.
Hassan, M.K., Sanchez, B., Yu, J.-S., 2011. Financial development and economic growth: new evidence from panel data. The
Quarterly Review of Economics and Finance 51 (1), 88–104.
Heggestad, A.A., 1977. The competitive condition of US banking markets and the impact of structural reform. Journal of Finance
32 (4), 649–661.
Hsieh, M.-F., Lee, C.-C., 2010. The puzzle between banking competition and profitability can be solved: international evidence
from bank-level data. Journal of Financial Service Research 38 (2), 135–157.
Hughes, J.P., Mester, L., 1998. Bank capitalization and cost: evidence of scale economies in risk management and signaling.
Review of Economics and Statistics 80 (2), 314–325.
Hughes, J.P., Moon, C., 1995. Measuring Bank Efficiency When Managers Trade Return for Reduced Risk. Department of
Economics Rutgers University Working Paper, No. 1995-20.
Iannotta, G., Nocera, G., Sironi, A., 2007. Ownership structure, risk and performance in the European banking industry. Journal of
Jacques, K., Nigro, P., 1997. Risk-based capital, portfolio risk and bank capital: a simultaneous equations approach. Journal of
Economics and Business 49 (6), 533–547.
Jahankhani, A., Lynge Jr., M.J., 1980. Commercial bank financial policies and their impact on market-determined measures of
risk. Journal of Bank Research 11 (3), 169–178.
Kahane, Y., 1977. Capital adequacy and the regulation of financial intermediaries. Journal of Banking and Finance 1 (2), 207–218.
Karels, G.V., Prakash, A.J., Roussakis, E., 1989. The relationship between bank capital adequacy and market measures of risk.
Journal of Business Finance and Accounting 16 (5), 663–673.
Kennedy, P., 2008. A Guide to Econometrics. Blackwell Publishing, Malden, MA.
Kim, D., Santomero, A.M., 1988. Risk in banking and capital regulation. Journal of Finance 43 (5), 1219–1233.
Koehn, M., Santomero, A., 1980. Regulation of bank capital and portfolio risk. Journal of Finance 35 (5), 1235–1245.
Kwan, S., Eisenbeis, R.A., 1997. Bank risk, capitalization, and operating efficiency. Journal of Financial Services Research 12 (2–3),
117–131.
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 1998. Law and finance. Journal of Political Economy 106 (6), 1113–1155.
Lepetit, L., Nys, E., Rous, P., Tarazi, A., 2008. Bank income structure and risk: an empirical analysis of European banks. Journal of
Maudos, J., Solís, L., 2011. Deregulation, liberalization and consolidation of the Mexican banking system: effects on competition.
Journal of International Money and Finance 30 (2), 337–353.
Neumark, D., Sharpe, S.A., 1992. Market structure and the nature of price rigidity: evidence from the market for consumer
deposits. Quarterly Journal of Economics 107 (2), 657–680.
Pettway, R.H., 1976. Market tests of capital adequacy of large commercial banks. Journal of Finance 31 (3), 865–875.
Rime, B., 2001. Capital requirements and bank behavior: empirical evidence for Switzerland. Journal of Banking and Finance 25
(4), 789–805.
Sanyal, P., Shankar, R., 2011. Ownership, competition, and bank productivity: an analysis of Indian banking in the post-reform
period. International Review of Economics and Finance 20 (2), 225–247.
Shen, C.-H., Lee, C.-C., 2006. Same financial development yet different economic growth – why? Journal of Money, Credit and
Banking 38 (7), 1907–1944.
Shim, J., 2010. Capital-based regulation, portfolio risk and capital determination: empirical evidence from the U.S. property-
liability insurers. Journal of Banking and Finance 34 (10), 2450–2461.
Short, B., 1979. The relation between commercial bank profit rates and banking concentration in Canada, Western Europe and
Japan. Journal of Banking and Finance 3 (4), 209–219.
Shrieves, R.E., Dahl, D., 1992. The relationship between risk and capital in commercial banks. Journal of Banking and Finance 16
(2), 439–457.
Smirlock, M., 1985. Evidence on the (non) relationship between concentration and profitability in banking. Journal of Money,
Credit and Banking 17 (1), 69–83.
Windmeijer, F., 2005. A finite sample correction for the variance of linear efficient two-step GMM estimators. Journal of
Econometrics 126 (1), 25–51.

The impact of bank capital on profitability and risk in Asian banking

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The impact of bank capital on profitability and risk in Asian banking

Uploaded by

Copyright:

Available Formats

Journal of International Money and Finance 32 (2013) 251–281

Contents lists available at SciVerse ScienceDirect

Journal of International Money

The impact of bank capital on proﬁtability and risk in Asian

* Corresponding author. Tel.: þ886 7 5252000x4825; fax: þ886 7 5254899.

2.1. Relationship between capital and risk

Authors (year) Period Countries Methodologies Empirical results

Authors (year) Period Countries Methodologies Empirical results

2.2. Relationship between capital and proﬁtability

3. Methodological issues and data sources

pit ¼ a0 þ a1 pit1 þ a2 CPit þ a0Fit þ li þ hit ; ci; t: (1)

Vit ¼ b0 þ b1 Vit1 þ b2 CPit þ b0Fit þ mi þ vit ; ci; t: (2)

3.2. Descriptions and sources of data

No. Code Country Bank number Income level Deposit insurance

4.1. Benchmark results

Classiﬁcation Variable Descriptions Sources

Classiﬁcation Variable Descriptions Sources

Classiﬁcation Variable Descriptions Sources

(1) (2) (3) (4)

(1) (2) (3)

(1) (2) (3) (4) (5) (6) (7)

(1) (2) (3) (4) (5) (6) (7)

(1) (2) (3) (4) (5) (6) (7)

(1) (2) (3) (4) (5) (6) (7)

4.2. Robustness analysis

(1) (2) (3) (4) (5) (6) (7)

(1) (2) (3) (4) (5) (6) (7) (8)

(1) (2) (3) (4) (5) (6)

(1) (2) (3) (4) (5) (6) (7)

(1) (2) (3) (4) (5) (6) (7)

You might also like