See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/264815079

The impact of competition on bank performance in Bangladesh: an empirical

study

Article  in  International Journal of Financial Services Management · January 2014

DOI: 10.1504/IJFSM.2014.062293

CITATIONS READS

7 2,164

2 authors:

S. M. Sohrab Uddin Yasushi Suzuki

University of Chittagong Ritsumeikan Asia Pacific University
23 PUBLICATIONS   58 CITATIONS    94 PUBLICATIONS   191 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Banking and Economic Rent in Asia View project

Green Society Bangladesh View project

All content following this page was uploaded by Yasushi Suzuki on 17 October 2016.

The user has requested enhancement of the downloaded file.

Int. J. Financial Services Management, Vol. 7, No. 1, 2014 73

The impact of competition on bank performance in

Bangladesh: an empirical study

S. M. Sohrab Uddin*
Department of Finance and Banking,
University of Chittagong,
Chittagong 4331, Bangladesh
Email: smsohrabcu@yahoo.com
*Corresponding author

Yasushi Suzuki
Graduate School of Management,
Ritsumeikan Asia Pacific University,
1-1 Jumonjibaru, Beppu, Oita, Japan
Email: szkya@apu.ac.jp

Abstract: This paper attempts to assess bank performance, competition and

their relationship empirically by using banking sector data (1983–2011) and
individual bank data (2001–2011) from Bangladesh. Return on assets and data
envelopment analysis grounded bank efficiency are used as bank performance
measures, whereas eight structural measures are employed for assessing
competition followed by the adoption of regression analysis for identifying its
impact. The findings report an improvement of bank performance with a few
fluctuations in between the sample periods under study. On the other hand, the
level of competition has been consistently increasing in the banking sector, as
pointed out by all structural measures. The regression result shows evidence
of a negative relationship between competition and bank performance. It is
required to revisit the structural changes of the banking sector. In particular, the
regulatory authorities need to ensure necessary incentives for banks, particularly
for private banks, to improve their performance in terms of profitability and
efficiency.

Keywords: financial repression; financial deregulation; competition;

concentration; bank performance; data envelopment analysis; efficiency; return
on assets; profitability; Bangladesh.

Reference to this paper should be made as follows: Uddin, S.M.S. and Suzuki, Y.
(2014) ‘The impact of competition on bank performance in Bangladesh:
an empirical study’, Int. J. Financial Services Management, Vol. 7, No. 1,
pp.73–94.

Biographical notes: S. M. Sohrab Uddin is an Associate Professor at the

Department of Finance and Banking, University of Chittagong, Bangladesh. He
obtained BBA and MBA from University of Chittagong, and MBA and PhD
from Ritsumeikan Asia Pacific University, Oita, Japan. His research area
focuses on financial market, institutions and instruments, with particular
emphasis on banking sector development in developing countries.

Copyright © 2014 Inderscience Enterprises Ltd.

74 S.M.S. Uddin and Y. Suzuki

Yasushi Suzuki, PhD (Economics), LLM (International Economic Law), is a

Professor at the Graduate School of Management, Ritsumeikan Asia Pacific
University, Oita, Japan. He has an undergraduate degree from Waseda
University in Tokyo and a doctorate from the University of London. His main
interest is on the institutional political economy of financial development. His
works have appeared in Review of Political Economy, Asian-Pacific Economic
Literature, Journal of Comparative Asian Development and Journal of Global
Ethics.

This paper is a revised and expanded version of a paper entitled ‘Competition

and bank performance in Bangladesh: an empirical study’ presented at the
‘2nd International Conference on Management and Economics’, University of
Ruhuna, Sri Lanka, 26–27 February 2013.

1 Introduction

The banking sector, initially in developed countries and later on in developing countries,
has experienced a paradigm shift to financial deregulation from financial repression
during 1980s and 1990s with an aim of accelerating the level of competition. Under the
changing scenario, a significant number of studies have been undertaken in different
parts of the world, particularly in the context of developed countries, to deal with the
empirical assessment of competition and its impact on the banking sector. According to
Ataullah and Le (2006, p.656), no study prior to their own focuses on the explicit
evaluation of the relationship between bank performance in the form of efficiency and
the level of banking sector competition from the perspective of developing countries.
However, the measurement of competition and its impact on bank performance carries
important policy implications for the banking sector regulator of any country irrespective
of its stage of development. For example, such analysis can be used as an indirect way
of analysing the impact of financial deregulation (Zhao and Murinde, 2011), and
accordingly, it provides necessary information for further regulation or deregulation of
the banking sector.
Concerning the banking sector of Bangladesh, the objective of accelerating
competition and banking sector performance has been accomplished by the government
by adopting different policies including introduction of private sector banks in 1982,
initiation of denationalising state-owned banks in 1983, and declaration of Financial
Sector Reform Program (FSRP) for deregulating the banking sector in 1989 by
introducing relaxation of reserve requirements, withdrawal of state directed credit policy,
development of legal infrastructure, adoption of international standard for loan screening
and monitoring, and liberalisation of deposit and lending rates (Debnath, 2004). The
policy initiative undertaken by the government during 2000 for merging or closing down
of unproductive branches of state-owned banks played significant role in creating level
playing ground for the private and foreign banks that emerged into the market during the
liberalisation framework. Bangladesh Bank, the central bank with the responsibility of
regulating and supervising the banking sector, also asked private and foreign banks to
increase their rural branches in proportion to urban branches by changing the urban to
rural ratio from 4:1 to 1:1 through issuing Banking Regulation and Policy Department
(BRPD) circular letter no. 13 in December 2011. The objective was to enhance the
 The impact of competition on bank performance in Bangladesh 75

financing access to agriculture and Small and Medium Enterprise (SME) sectors, even
though banks in developing countries do not have the incentive to do so in a liberalised
and competitive framework as argued by Hellmann et al. (1997). Does the level of
competition under the changing circumstances in Bangladesh provide necessary impetus
for banks to improve their performance? This issue has become worth investigating and
the present study is an attempt in this aspect.
The existing studies focusing either on the South Asian region or on the banking
sector of Bangladesh attempt to address this issue by assessing the level of competition
and its impact on the banking sector. For instance, Perera et al. (2006) use the data of 26
banks with a sample period from 1995 to 2003 and adopt the Panzar–Rosse algorithm for
assessing the nature of competition while making a comparative study among South
Asian countries. But, they have not explicitly focused on the performance aspect. On the
other hand, Samad (2008) tests the validity of either of two distinct paradigms, namely
Structure Conduct Performance (SCP) or Efficient Structure (ES) during the period
1999–2001 by using three-bank concentration ratio as a proxy for competition. However,
employing a single predictor does not always reflect the actual scenario of the market due
to its inherent limitations. As a result, a list of proxies has been used in this study.
In addition, the adoption of traditional measure as well as Data Envelopment Analysis
(DEA) grounded efficiency for evaluating bank performance is another area of
differentiation from Samad (2008). It is also important to note that many private banks
started to expand their branch network after the time period covered by the above studies
due to the relaxed policy promulgated by Bangladesh Bank and the liberalisation and
development of stock market during recent time. Thus, this study, by using the data
of pre- and post-liberalisation periods, fills in the gap with regard to the impact of
competition on banking sector and provides necessary policy implications for the
regulatory authorities in Bangladesh. This paper also contributes to existing literature in
developing countries by placing country-specific banking market evidence.
The later part of this paper is organised in the following way. Section 2 deals with the
review of the related literature concerning SCP and ES highlighted by the banking
literature. Section 3 provides a brief overview of the banking sector of Bangladesh.
Section 4 highlights the data and methodology adopted for computing bank performance
and competition, and for assessing the impact of competition. Section 5 represents the
empirical findings and corresponding analyses, and Section 6 concludes the paper with
some policy implications.

2 Review of the related literature

The banking literature remains divided over the conflicts arising out of the SCP
paradigm, also known as the structure performance paradigm, and the ES paradigm. The
SCP hypothesis, which, according to Park (2009, p.654) and Seelanatha (2010, p.21),
dates back to Mason (1939), is the oldest and traditional hypothesis. It states that the
performance of banks largely depends upon the structure of the market such as the
number of banks and the market shares of banks; and the profitability of banks decreases
with the increase of competition. In other words, the higher the concentration ratio, the
higher will be the profitability of banks, reflecting a positive association between market
share of a bank and its performance. This is because, the SCP hypothesis is highly
76 S.M.S. Uddin and Y. Suzuki

governed by the traditional neoclassical theory on the ground that due to collusive or
monopolistic reason banks, irrespective of their efficiencies, in a concentrated market are
always in a position to charge prices above the marginal cost in order to increase profit.
For example, various regulatory restrictions including ceilings on interest rates and entry
barriers facilitate collusive behaviour of banks even in a low-concentrated market
(Lloyd-Williams et al., 1994). Therefore, the SCP paradigm stresses on the importance of
increased concentration with a view to reducing competitive behaviour of banks (Bikker
and Haaf, 2002), and many scholars express their views in favour of the concentrated
market structure. For instance, Hellmann et al. (1997) point out the importance of
concentrated market by arguing that competition creates disinclination towards the
achievement of optimal scale. Boyd et al. (2004) and Beck et al. (2006) find that
financial or banking crisis is less likely to occur in a concentrated market. Casu et al.
(2010) opine that increased competition in the banking market provides less incentive for
banks in delegating monitoring and screening of loans; therefore, badly affects the
financial stability. At the same time, many proponents such as Rose and Fraser (1976),
Heggestad and Mingo (1977), Berger and Hannan (1989), Lloyd-Williams et al. (1994)
and Samad (2005) express their opinions in favour of the SCP hypothesis. Nevertheless,
the survey conducted by Gilbert (1984) is quite influential in this regard, since he
summarises 44 studies focusing on the relationship between market concentration and
bank performance. According to Gilbert’s findings, 32 studies report significant positive
impact of concentration on bank performance.
In contrast, many scholars argue in the opposite direction to support the competitive
market structure. Calem and Carlino (1991) point out that a market with higher
concentration is more vulnerable to crisis; and thereby is less competent and equitable.
Berger et al. (2004) mention that government intention to restrict competition through
foreign bank entry regulation and state ownership of banks generates adverse effect and
ultimate poor economic efficiency in a country. Furthermore, it is highly likely that banks
in a concentrated atmosphere can engage in non-competitive deeds to generate higher
revenue with lower benefits for consumers (Abbasoglu et al., 2007; Wong et al., 2008);
thus, produce monopoly and corresponding inefficiencies (Suzuki et al., 2008).
Importantly, arguments highlighting the possible benefits of competitive market actually
emerge from the application of the standard industrial organisation economics to the
financial sector, particularly the banking sector. Moreover, these arguments show their
inclination towards the alternative hypothesis of the SCP, that is, the ES which states
that enhanced performance of banks leads to higher market share which in turn results
into market concentration associated with superior efficiency. That is, bank-specific
efficiency difference in a particular market leads to uneven proportion of market size and
corresponding high intensity of concentration. In fact, this hypothesis does not consider
market concentration as a random event; instead, it is the result of greater efficiency
of the dominant banks (Smirlock, 1985). This is possible because a bank with either
superior management or production technology in a competitive market can lower cost to
increase profit and to attain higher market share (Berger, 1995). On the other hand, a
bank with higher efficiency than its competitors can also maximise profit either by
maintaining the current market size and pricing policies or by accommodating size
expansion and price reduction strategies (Lloyd-Williams et al., 1994, p.437). It means
that banks under such a market mechanism strive for achieving dual objectives of
maximising profits and minimising costs and prices, and as a consequence, the highest
 The impact of competition on bank performance in Bangladesh 77

amount of credit will be allocated (Northcott, 2004). Therefore, according to the ES

hypothesis, profitability of banks greatly depends upon the efficiency rather than the
market structure of the banking sector. In this way, the emergence of the ES not only
challenges the traditional SCP hypothesis but also puts forward an alternative way of
analysing the different dynamics of the banking sector. Studies of Demsetz (1973),
Brozen (1982), Samad (2008) and Seelanatha (2010) support the ES hypothesis.
Therefore, based on the literature reviewed above, it can be argued that the nature of
the relationship between competition and bank performance is rather ambiguous.
According to Wanniarachchige and Suzuki (2010), the relationship is country-specific
in nature and as such they suggest to conduct more studies at country level. Focusing
on the findings of earlier studies, concentrating on the banking sector of Bangladesh
such as Samad (2008), the impact of competition on bank performance in the form of
profitability and efficiency cannot be generalised since the results derived from pool and
annual data portray different results. Because of this anomaly, he urges for further studies
to explore the impact of changing market structure on bank performance.
With regard to the development of models for assessing the impact of competition,
previous studies concentrating on developing countries adopt bank-specific, industry/
country-specific, or a combination of both types of variables. For instance, Wanniarachchige
and Suzuki (2010) use industry/country-specific variables only, whereas Ataullah and
Le (2006) and Samad (2008) use both variables for their studies. This study also adopts
similar approaches for selecting the necessary variables in the regression.

3 Overview of the banking sector of Bangladesh

In spite of the existence of different forms of financial institutions and intermediation, the
banking sector dominates the process of channelisation of funds in Bangladesh. The size
of banking sector assets relative to Gross Domestic Product (GDP) was 69.76% in 2010
compared to the size of market capitalisation of the stock market relative to GDP of
32.79% and the size of non-banking sector assets relative to GDP of 5.98% (Bangladesh
Bank, 2011; Uddin and Gupta, 2012). By the end of 2011, the banking sector composed
of 47 scheduled banks out of which four were state-owned banks, four were state-owned
development financial institutions, 30 were private banks, and nine were foreign banks
(Bangladesh Bank, 2011). Private banks are sub-divided into denationalised banks,
Islamic banks and non-Islamic banks, and there were two denationalised banks, seven
full-fledged Islamic banks and 21 non-Islamic banks. However, out of the 21 non-Islamic
banks, 16 maintained separate Islamic banking windows for their customers leading to a
total of 23 banks that engaged in the Islamic banking activities either partly or fully. The
adoption of financial deregulation policies for accelerating the number of banks and
branches under private and foreign ownerships results into a shift of market share from
state-owned scheduled banks to private and foreign banks. For instance, in 1976, state-
owned scheduled banks controlled 94.5% of deposits whereas by the end of 2011 the
share became 32.1%.1 The share of private and foreign banks, in contrast, has been in an
increasing tread. To be specific, the share of private banks has increased from 3.6% in
19832 to 61.6% in 2011 and the share of foreign banks has increased by 14.5% during
1976–2011. Similar changes can also be observed in the credit market. The permission
issued by Bangladesh Bank for the inception of nine new private banks in 2012
accelerates the share of private banks further in future.
78 S.M.S. Uddin and Y. Suzuki

Table 1 represents the composition of different types of banks at the end of 2011.
Table 1 also indicates that private banks dominate the market with respect to both
deposits and credits. To be specific, the market share of private banks in terms of
deposits and credits is 61.6% and 66.6%, respectively, whereas state-owned banks and
development financial institutions collectively hold 32.1% of deposits and 27.8% of
credits. On the other hand, foreign banks control 6.3% and 5.6% of banking sector
deposits and credits, respectively. Private banks also account for 60.0% of the banking
sector assets compared to 34.0% of state-owned scheduled banks and 6.0% of foreign
banks. Moreover, private banks maintain 94.1% of the banking sector Automated Teller
Machines (ATMs). However, in aggregate, state-owned scheduled banks hold the
majority of the branch network by maintaining 60.8% of the total number of branches,
out of which state-owned banks account for 43.2% and development financial institutions
retain 17.6%. Both state-owned banks and development financial institutions maintain
more branches in the rural areas than in the urban areas, since the corresponding
percentage of their rural branches are 63.8% and 88.2%. In contrast, out of their 38.4%
branch network relative to the banking sector, private banks focus more on urban areas
by establishing 63.4% of their branches in the urban areas. Foreign banks, on the other
hand, concentrate solely on the urban areas with 100.0% urban branches.
Table 1 Composition of different types of banks

Types of No. of No. of No. of Deposits Credits Assets

Banks Banks Branches ATMs Amount % Amount % Amount %
SB 4 3437 52 1214.30 27.3 747.60 21.3 1660.76 28.0
DFI 4 1406 5 213.00 4.8 229.30 6.5 326.47 6.0
PB 30 3055 3386 2746.10 61.6 2332.70 66.6 3556.80 60.0
FB 9 63 155 278.60 6.3 195.50 5.6 379.47 6.0
DB 2 590 5 192.07 4.3 148.54 4.2 254.47 4.0
IB 7 655 155 742.01 16.7 681.63 19.4 950.84 16.0
NIPB 21 1810 3226 1812.02 40.6 1502.53 43.0 2351.49 40.0
Notes: SB, DFI, PB, FB, DB, IB, and NIPB, respectively, stand for state-owned banks,
development financial institutions, private banks, foreign banks, denationalised
banks, Islamic banks, and non-Islamic private banks. Amounts are in BDT (BDT
stands for Bangladesh Taka, and Taka is the local currency of Bangladesh)
billion.
Source: Constructed by the authors based on October–December issue 2011 of
scheduled bank statistics, a quarterly publication of Bangladesh bank. Few of
the data are collected from the Statistics Department of Bangladesh bank

4 Data and methodology

In order to measure bank performance, both traditional measure and frontier-based

measure are used. Bauer et al. (1998) and Sathye (2003) urge for the simultaneous
adoption of both measures for making the results more meaningful and presentable. In
this regard, Return on Assets (ROA) is the traditional measure and DEA-based efficiency
is the frontier measure. ROA is used as a non-frontier-based measure due to its ability to
reflect the earnings and profitability of banks. It is considered as a major indicator by
 The impact of competition on bank performance in Bangladesh 79

investors, managers and business analysts for appraising the performance of a bank on
the whole (Avkiran, 1999). This is because, it indicates the managerial capability of a
bank to generate revenue by utilising its resources and thus, any change in ROA over
time provides necessary evidence for judging the increase or decrease of performance
(Neal, 2004, p.180). Fraser and Fraser (1990) also urge for assessing bank performance
from a profitability point of view and regard ROA as an important tool for doing so. On
the other hand, the application of parametric techniques is rather complicated in this case
since, according to Bhattacharyya et al. (1997, p.335); the continuation of financial
repression policies for a relatively long period of time and the existence of market
imperfections in developing countries greatly distort the input and output prices and
quantities. It is also highly likely that the adoption of parametric techniques requires prior
estimation of the functional form for estimating the efficiency of banks along with a large
sample size, which is not always possible to execute precisely from the perspective of a
developing country like Bangladesh. Avkiran (1999), Ataullah et al. (2004) and Neal
(2004) also employ similar approach by using ROA for comparing the results generated
by DEA in their studies. The yearly ROA figures during the period 1980–2011 represent
the banking sector performance as a whole, whereas the efficiency is calculated for a
sample of 39 banks including four state-owned banks, 30 private banks and five foreign
banks for a period of 11 years from 2001 to 2011.

4.1 DEA
The term DEA was first developed and used by Charnes et al. (1978) as a linear
programming technique for measuring the efficiency of non-profit organisations in the
public sector (Casu and Molyneux, 2003, p.1866; Neal, 2004, p.175; Ruggiero, 2007,
p.259; Sathye, 2001, p.617; Sathye, 2003, p.665). Generally, efficiency means the ability
of a firm to convert its inputs into outputs in the best possible manner. According to Casu
and Molyneux (2003, p.1866) and Neal (2004, p.175), Charnes et al. (1978) use the
efficiency concept developed by Farrell (1957) where he opines that efficiency is
composed of technical efficiency and allocative efficiency. Technical efficiency can be
regarded either as an output-oriented approach or as an input-oriented approach. If the
focus is on maximising outputs for a given level of inputs, it is considered as an output-
oriented approach; whereas if the focus is on minimising inputs for a given level of
outputs, it is treated as an input-oriented approach. On the other hand, allocative
efficiency refers to the use of inputs in respect to their given prices optimally. The
introduction of DEA encourages the researchers to measure the DEA grounded efficiency
of firms in different sectors of the economy by extending the original methodology of
DEA. But, concerning the banking sector, Sherman and Gold (1985) were the pioneers in
measuring the DEA-based efficiency (Molyneux et al., 1996; cited in Sathye, 2001,
p.617). Later on many other studies use DEA for their respective banking sector
efficiency analyses.
Under DEA, each firm within the sample is regarded as a Decision-Making Unit
(DMU) and the efficiency is calculated for each and every DMU. This study treats each
bank of the sample as a DMU. DEA generally provides relative efficiency score of each
DMU rather than generating any form of absolute measure; thus, it is commonly
regarded as a benchmarking tool as the most efficient firms (banks) lie on the frontier and
80 S.M.S. Uddin and Y. Suzuki

envelope other low-performing firms (banks) of the sample (Neal, 2004, p.176). The
efficiency score generated by the DEA measure ranges from 0 to 1. A DMU with an
efficiency score of 1 is treated as the most efficient within the sample. Similarly, an
efficiency score of 0.95 means that the DMU is 95% efficient compared to the best
practiced DMUs. DEA is unit invariant, which means that the variables used for
measuring bank efficiency are not required to represent in a common unit of measure. It
can accommodate numerous variables along with small sample size. In addition, it does
not require any form of distributional proposition such as the residuals in the regression
analysis (Wu, 2005). Moreover, being a non-parametric technique, the application of
DEA does not require prior estimation of the functional form of the efficient frontier,
which is the secret of its wider application compared to other parametric and non-
parametric techniques, particularly in the context of developing countries. At the same
time, the specification of the weights of different inputs and outputs is not required in
advance under DEA, rather such weights are mechanically produced by the model itself
so that maximum possible score for a particular DMU can be attained (Cooper et al.,
2007). However, in spite of these advantages, there are a few limitations of the DEA
concept. Firstly, DEA as a deterministic model cannot avoid the estimation error
situation, and the existence of too many diversified DMUs within the sample makes the
results ambiguous. This is because no adjustment is made by the DEA framework to
incorporate bias generated by heterogeneity, external shocks, measurement error and so
on, as it is a non-parametric and non-stochastic approach (Worthington, 2000). Secondly,
no hypothesis testing is done under DEA due to the non-existence of the distributional
form. Thirdly, any violation of the data integrity makes the DEA results unreliable
(Sathye, 2003).
The specification of inputs and outputs is very crucial for measuring the efficiency
through DEA and the specification largely depends upon the selection of an approach
indicating the behaviour of bank as a legal entity. Generally, two main approaches are
available for specifying the inputs and outputs of banks. The key differences between the
two approaches rely on how they treat deposits and interest expenses. The production
approach regards deposit as output whereas the intermediation approach treats deposit as
input. Similarly, the production approach completely ignores interest expenses from the
list of inputs and outputs whereas the intermediation approach includes interest expenses
in the input side. The present study selects the intermediation approach since studies
including Berger and Humphrey (1997) and Casu and Molyneux (2003) regard this
approach as superior to the production approach. Accordingly, interest expenses and non-
interest expenses divided into overhead expenses excluding staff expenses and personnel
expenses are on the input side whereas interest income and non-interest income are on
the output side. Non-interest income, which includes various fees and commissions
received from different non-banking operations of banks including stock market
activities, is accommodated due the intensification of such activities under financial
deregulation. Many other studies including Avkiran (1999, 2000), Sathye (2003), Sturm
and Williams (2004), Galagedera and Edirisuriya (2005) and Uddin and Suzuki (2011)3
also employ similar approach for their respective studies. However, the specification
done under this study significantly differs from these studies as it takes interest income
rather than net interest income as output. This is because inclusion of net interest income
 The impact of competition on bank performance in Bangladesh 81

(which is the difference between interest income and interest expenses) in the output side
leads to double counting of interest expenses as it is included in both input and output
sides’ calculations. Moreover, although in the banking literature both ‘flow’ and ‘stock’
concepts have been used simultaneously for the specification of inputs and outputs, this
study uses flow concept mainly to avoid the limitation of using both concepts in the same
model. In their respective studies Sathye (2003), Drake et al. (2009) and Hadad et al.
(2010) also acknowledge the flow concept, which is also corresponding to the
revenue/profit-based efficiency originally featured by Berger and Mester (1997).
This study uses Slack-Based Measure (SBM), which was first coined by Tone
(2001),4 due to its superiority over other traditional measures such as Charnes–Cooper–
Rhodes (CCR) measure and Banker–Charnes–Cooper (BCC) measure. Usually, unlike
the CCR and BCC measures, the SBM measure introduces slacks (‘inputs excesses’ and
‘output shortfalls’) of the DMUs in estimating bank efficiency, which has been a lacking
prior to its emergence. As a result, the efficiency score generated by SBM never exceeds
the score provided by CCR measure and thus, a DMU is SBM-efficient only if it is CCR-
efficient. Furthermore, concerning the units of data, SBM is invariant which irrespective
of the compatibility makes it highly distinguished from other models. Drake et al. (2009)
and Hadad et al. (2010) also use SBM model for measuring DEA-based bank efficiency.
The present study also employs Constant Returns to Scale (CRS)5 assumption suggested
by Charnes et al. (1978),6 instead of Variable Returns to Scale (VRS) assumption to
ensure maximum discrimination of DMUs; since VRS only compares similar-sized
DMUs; and accordingly, fails to distinguish between low-performing DMUs with dissimilar
sizes.

4.2 Competition measurement techniques

On the other hand, this study uses structural competition measures, namely the k-bank
concentration ratios (one-bank, three-bank, five-bank and seven-bank), Herfindahl–
Hirschman Index (HHI), Hall–Tideman Index (HTI), Comprehensive Industrial
Concentration (CCI) and entropy measure with a sample period of 29 years from 1983 to
2011 to measure the level of competition. Generally, structural measures link competition
to concentration. It means that the lower the level of concentration in the banking sector,
the higher will be the competition. The fundamental reason for choosing the structural
approaches over other available approaches is that the present study attempts to ascertain
the impact of changing market structure on bank performance. In order to accomplish the
objective, structural measures are more appropriate, as it is argued by Seelanatha (2010,
p.21) that other approaches, for instance non-structural approaches, focus on factors other
than changing market structure that affect the competitive behaviour of banks. The
competition is calculated by using the loan figures mainly due to the availability during
the whole period under study, and all banks operating in a particular year are considered
for the computation. The data are collected from different issues of the ‘Bank O Arthik
Pratisthansamuher Karjaboli’, which is in Bengali language (the English transformation
of the publication is ‘Bank and Financial Institutions’ Activities’) and published by the
Ministry of Finance of the Government of Bangladesh. Table 2 provides a comparative
summary of the competition measures adopted in this study.
82 S.M.S. Uddin and Y. Suzuki

Table 2 Summary of different structural measures of competition

Name of the Focus of the Mathematical

Range
measure measure form
The market share of the k
k
K-bank largest bank(s) depending
CRk = ∑ si Zero and 1
concentration ratios upon the arbitrary selection i =1
process.
n
The market share of all banks
HHI HHI = ∑ si 2 1/n and 1
in an industry. i =1

Assign more focus on number ⎛ n ⎞

HTI HTI = 1 ⎜ 2∑ isi − 1⎟ Zero and 1
of banks in an industry. ⎝ i =1 ⎠
Accommodate both absolute n
Decimal
CCI and relative concentrations CCI = s1 + ∑ si 2 (1 + (1 − si ) )
fraction and 1
with equal importance. i=2

Smaller absolute weight is n Zero and

Entropy assigned to the largest bank E = ∑ si log 2 si
log 2 n
(s) with higher market share. i =1

Source: Constructed by the authors based on Bikker and Haaf (2002)

4.3 Competition and bank performance

In order to assess the impact of all of the competition measures on ROA, the following
regression equation will be executed eight times by using each of the competition
measures as independent variable in each case and by maintaining the dependent variable
and other independent variables unchanged:
ROAt =α 0 + β1COMPt + β 2 GDPGt + β 3 BRANCH t + ε t (1)

By using similar approach as above, the following regression equation will be used for
assessing the impact of competition on efficiency:
EFFit =α 0 + β1COMPit + β 2 GDPGit + β3 BRANCH it + β 4 SIZEit + β5 NPLit + β 6 AGEit +ε it (2)

where ROA represents the return on assets of the banking sector as a whole, COMP
stands for competition, GDPG means growth rate of GDP at current market price
(computed by dividing the difference of current year’s GDP and previous year’s GDP by
the GDP of the previous year), BRANCH stands for total number of branches of the
banking sector per million of people in the first case and number of branches of a
particular bank in the second case, EFF represents efficiency, SIZE means total assets of
a bank relative to total banking sector assets, NPL expresses non-performing loans of a
bank relative to total loans, AGE means number of years in operations, α denotes
intercept, β represents regression coefficients, i and t denote bank and time, respectively,
and ε represents disturbance term. GDPG, BRANCH, SIZE, NPL and AGE are
used to observe the influence of the macroeconomic environment, the impact of the
outreach of cliental-base, the effect of banks’ relative size, the influence of associated
credit risk of loans and the impact of experiences gained over the years on bank
performance, respectively. Unlike equation (2), equation (1) includes industry/country-
specific variables only, as the ROA figures represent the banking sector as a whole.
 The impact of competition on bank performance in Bangladesh 83

The relationship between GDPG and bank performance indicators can be either
positive or negative. The growing economy in a developing country like Bangladesh
possibly creates greater room for banks to attain higher profit/efficiency by extending
their business. At the same time, it is also highly likely that the frequent instability in the
form of political unrest in the country may offset the favourable atmosphere created by
the macroeconomic environment from time to time. Moving to the variable BRANCH,
most of the banks, particularly private banks, are in their expansionary phase and thus,
can attain more aggregate profit by utilising their expanded branch network leading to a
positive association between ROA and BRANCH. Also, there is no guarantee that a new
branch operates with the same level of efficiency of an old branch, which can make
the relationship between efficiency and BRANCH negative. Among other variables,
SIZE and NPL are expected to have a positive and a negative impact on efficiency,
respectively. Finally, AGE brings experiences in doing business for banks and thus, its
positive impact on efficiency is predicted.
The sample period used for assessing the relationship between competition and ROA
is 1983–2011, leading to a total of 29 years. On the other hand, the panel data set in
equation (2) contains the data of 39 banks including four state-owned banks, 30 private
banks and five foreign banks. The sample period is 2001–2011 mainly due to the
unavailability of bank level data prior to 2001. Thus, covering 39 banks with a sample
period of 11 years leads to a balanced panel of 429 observations. Data relating to ROA
are collected from different annual reports of Bangladesh Bank. On the other hand, data
relating to GDPG and BRANCH (per million of people) are collected from different
December issues of the Monthly Economic Trends, a publication of Bangladesh Bank.
All the bank level data are collected from respective banks’ annual reports.

5 Empirical findings

5.1 Changes in bank performance

Figure 1 portrays the trend of the banking sector ROA during the period 1980–2011. It
shows that prior to the declaration of FSRP; the ROA of the sector was low but stable in
nature. Immediately after the declaration in late 1980s, the ROA was drastically reduced;
and during the period 1991–1993, the ROA was negative. Although it showed a slight
improvement in 1994 and onward, but from time to time the ROA of the sector became
either negative or close to zero. To be specific, the ROA of the year 1995 was negative
and that of the year 2000 and 2007 were close to zero. However, after 2007 the ROA has
started to increase by showing a rate of more than 1.00%, which was the first time in
the banking history of Bangladesh, and reached to its highest rate, that is 1.80% in 2010.
It has shown a declining trend again in 2011.
On the other hand, Figure 2 reflects the efficiency of the sample banks during the
period 2001–2011. Although the overall efficiency of the sample banks has increased by
13.79% in 2011 compared to 2001, the banking sector experienced few ups and downs in
between. The efficiency of the sample banks reached its highest score of 0.77 in 2009
and reduced to 0.73 in 2010, which has reduced further to 0.66 in 2011.
84 S.M.S. Uddin and Y. Suzuki

Figure 1 ROA of the banking sector (see online version for colours)

Figure 2 Efficiency of the sample banks (see online version for colours)

In order to have an idea about the efficiency scores of different types of banks,
Table 3 reports bank category-wise efficiency scores. Table 3 also states that foreign
banks are the best performing banks whereas private banks are the least performing
banks. The efficiency score of state-owned banks has increased to 0.77 in 2011 from 0.74
in 2001 by reaching its highest score of 0.83 in 2010. On the other hand, the efficiency
score of private banks has increased by 15.09% in 2011 compared to 2001. However,
after reaching the highest score of 0.74 in 2009, the efficiency of private banks has been
in a decreasing trend. In contrast, the efficiency scores of foreign banks are more than
0.90 in most cases except in 2001 and 2002. The added advantages of foreign banks in
the form of technological advantage and concentration only in the big cities and that of
state-owned banks in the form of general public’s credibility and safety due to the
government affiliation, favourably contribute towards the upholding of their upper hands
over private banks.
Table 3 Bank category-wise efficiency scores

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
State-owned
0.74 0.71 0.70 0.39 0.69 0.78 0.76 0.66 0.79 0.83 0.77
banks
Private banks 0.53 0.50 0.62 0.53 0.59 0.67 0.68 0.70 0.74 0.69 0.61
Foreign banks 0.75 0.79 0.91 0.90 1.00 1.00 1.00 1.00 1.00 0.97 0.93
 The impact of competition on bank performance in Bangladesh 85

5.2 Changes in competition level

Figures 3 and 4 represent various competition ratios during the period 1983–2011. It is
clear from the graphs that the calculated values of competition not only vary across ratios
but also differ within a particular ratio. The one-bank ratio indicates the market share of
the leading bank, which was a state-owned bank until 2009. Its share has reduced to 9.6%
in 2009 from 26.0% in 1983. However, in 2010, one of the private Islamic banks took
the leading position by capturing 8.8% of the total market share, although the share has
reduced to 8.6% in 2011. Both the rise of Islamic banking and the policy initiative
undertaken by the government in 20007 with a view to merging or closing down of
branches of state-owned banks may result in such transfer of market leadership. The
three, five and seven banks’ ratios have reduced by 58.46%, 56.35% and 54.39%,
respectively, in 2011 compared to 1983. It means that the level of banking sector
competition has intensified during the period under study in spite of the fact that the top
three, five and seven largest banks still collectively maintain 23.8%, 32.3% and 39.0%,
respectively, of the total market share. The emergence of many new private and foreign
banks and the approximate double declining reduction of the market share of larger state-
owned banks under financial deregulation have contributed considerably towards the
increase in the level of competition. Even though the banking sector has experienced few
mergers and acquisitions among existing banks, their impact is not reflected in the
change of the market structure mainly due to the fact that these were constituted among
banks with smaller market shares to affect the structure of the market in a noticeable
way.

Figure 3 Changes in the competition ratios (excluding entropy) (see online version for colours)

Figure 4 Changes in the entropy measure (see online version for colours)
86 S.M.S. Uddin and Y. Suzuki

Similar results are also depicted by other measures. Starting with the HHI, it can be
observed that it has reduced to 0.038 in 2011 from 0.145 in 1983. Linking this ratio to the
horizontal merger guideline of the US Department of Justice, it can be concluded that the
banking sector of Bangladesh has shifted to a very low concentrated market from a
moderately concentrated market. As a matter of fact, the HHI index in 1983 was 1450
(computed as 0.145 × 10,000), and by the end of 2011 the index becomes 3808
(computed as 0.038 × 10,000). Likewise, the HTI has reduced to 0.037 in 2011 from
0.143 in 1983 and the CCI has reduced to 0.145 in 2011 from 0.403 in 1983. On the other
hand, the entropy measure has increased to 5.076 in 2011 from 3.222 in 1983. Yet, it
indicates an increasing trend of competition because of its reverse nature of relationship.
For instance, HHI, HTI and CCI are positively related with concentration whereas
entropy measure is inversely related with concentration.

5.3 Relationship between competition and bank performance

Table 4 accumulates the descriptive statistics of the variables used for measuring the
relationship between competition and ROA. The average HHI score is 0.0881 during the
period 1983–2011 with a maximum and a minimum value of 0.1446 and 0.0376,
respectively. The maximum and minimum values of the one-bank, three-bank, five-bank,
seven-bank, HTI, CCI and entropy ratios indicate the degree of changes of the level of
competition in the banking sector during the period under study. On average, the GDP
grows by 4.97% by reaching to a maximum rate of 6.67% and a minimum rate of 2.16%.
The lowest number of branches per million is 44.94 million and the highest number
of branches is 54.56; thus, indicate an overall increase of 21.41% during the period
1983–2011. The ROA figures are transformed by using the Van Der Waerden’s formula
to remove the normality problem, as shown in Table A1 of Appendix A. According to the
transformation technique, ranking is created by using the normal scores. The formula
used for the transformation is r/(N + 1), where w is the sum of the case weight and r is
the rank. Alkdai and Hanefah (2012) also use similar technique for the transformation.
Table 4 Descriptive statistics of the variables used for assessing the relationship between
competition and ROA

Variables N Minimum Maximum Mean Std. Deviation

T
ROA 29 0.0023 0.0667 0.0344 0.0196
HHI 29 0.0376 0.1446 0.0881 0.0311
CRk1 29 0.0863 0.2605 0.1917 0.0500
CRk3 29 0.2382 0.5733 0.4222 0.1005
CRk5 29 0.3229 0.7401 0.5609 0.1278
CRk7 29 0.3896 0.8549 0.6390 0.1378
HTI 29 0.0370 0.1434 0.0776 0.0315
CCI 29 0.1451 0.4033 0.2849 0.0744
Entropy 29 3.2216 5.0798 4.1702 0.5885
GDPG 29 0.0216 0.0666 0.0497 0.0119
BRANCH 29 44.9400 54.5600 49.7717 3.1825
 The impact of competition on bank performance in Bangladesh 87

The regression results of the model for measuring the impact of HHI on ROA are
displayed in Table 5. The results are adjusted for autocorrelation problem. However,
there is no multi-collinearity problem among the independent variables as shown by the
Variance Inflation Factors (VIFs) in Table A2 of Appendix A. The HHI has a negative
impact on ROA, which is significant at 1% level. Similar to the HHI, all other measures
influence the ROA of the banking sector. In all cases, the negative relationship is
significant at 1% level, as summarised in Table 6 (also supported by the correlations
among different competition proxies reported in Table 3 of Appendix A). The F-statistic
indicates the significance of each model at either 1% level or 5% level.
Table 5 Relationship between HHI and ROA

Model β Std. Error t

ά 0.065 *** 0.276 2.862
HHI –0.488 *** 0.174 –2.802
GDPG –0.189 0.405 –0.467
BRANCH 0.001 0.001 0.302
R-Squared 0.392
Durbin-Watson Statistic 2.05
F-Statistic 3.71 **
Notes: The results are obtained after making proper adjustment for autocorrelation
problem by adopting the autoregressive (AR) estimation method. The symbols
(***) and (**) represent statistical significance at 1% and 5% level,
respectively.

Table 6 Relationships between other competition measures and ROA

Model β Std. Error t F-Statistic

CRk1 –0.260 *** 0.096 –2.71 3.45 **
CRk3 –0.138 *** 0.049 –2.84 3.69 **
CRk5 –0.119 *** 0.038 –3.12 4.15 ***
CRk7 –0.114 *** 0.038 –3.02 4.08 ***
HTI –0.597 *** 0.202 –2.95 4.10 ***
CCI –0.196 *** 0.067 –2.92 3.85 **
Entropy 0.030 *** 0.009 3.39 4.68 ***
Notes: The results are obtained after making proper adjustment for autocorrelation
problem by adopting the autoregressive (AR) estimation method. The symbols
(***) and (**) represent statistical significance at 1% and 5% level,
respectively.
The descriptive statistics of all the variables used for measuring the relationship between
competition and efficiency are displayed in Table 7. It shows that the average efficiency
score of the sample banks during the period 2001–2011 is 0.6651. The score of the least
efficient sample bank is 0.13, whereas the score associated with the most efficient bank
or banks is 1.00. On average, the largest bank holds 14.14% of the total market share
whereas the largest three, five and seven banks, correspondingly, hold 31.16%, 42.10%
and 49.17%, respectively, of the total market share during the sample period. The number
of branches and the age of business operation of the sample banks range from 1 to 1291
and 2 to 40, respectively.
88 S.M.S. Uddin and Y. Suzuki

Table 7 Descriptive statistics of the variables used for assessing the relationship between
competition and efficiency

Variables N Minimum Maximum Mean Std. Deviation

EFF 429 0.1300 1.0000 0.6651 0.2249
HHI 429 0.0376 0.0810 0.0543 0.0145
CRk1 429 0.0863 0.2067 0.1414 0.0413
CRk3 429 0.2382 0.4113 0.3116 0.0582
CRk5 429 0.3228 0.5351 0.4210 0.0743
CRk7 429 0.3896 0.6123 0.4917 0.0800
HTI 429 0.0370 0.0585 0.0442 0.0072
CCI 429 0.1451 0.2802 0.2046 0.0461
Entropy 429 4.4168 5.0798 4.8345 0.2259
GDPG 429 0.0442 0.0666 0.0590 0.0065
BRANCH 429 1 1291 139.1935 264.0833
SIZE 429 0.0003 0.1984 0.0234 0.0288
NPL 429 0.0001 0.8099 0.0851 0.1211
AGE 429 2 40 16.9744 9.4286

Table 8 represents the summary of the regression results between HHI and bank
efficiency. There is no severe multi-collinearity problem existed among the independent
variables used for the regression (see Table A2). However, the Breusch–Pagan test
indicates a p-value of 0.01 and the Wooldridge test reports a p-value of 0.00, which
notify the existence of heteroscedasticity and autocorrelation problems. In this regard, the
Prais–Winsten panel-estimation method is adopted, since it automatically corrects both
problems. Adhikary (2011) also adopts similar approach in his study. It is from Table 8
that the HHI has a strong negative impact on efficiency, which is significant at 5% level.
Among other variables, NPL has a negative effect on efficiency whereas age has a
reverse impact. Both are significant at 5% level. It means that the higher the NPL of a
bank, the lower will be the efficiency and vice versa. It is also evident from the findings
that banks’ experiences gained over the years from the business operations have also
positively influenced their efficiencies, which follows the replication of the idea of
‘learning by doing’. The other independent variables do not possess any statistically
significant relationship with bank efficiency.
On the other hand, Table 9 depicts the relationships between other competition
measures and bank efficiency. Similar to the HHI, all other measures indicate their
respective negative relationship with bank efficiency. The negative relationship of the
one-bank and three-bank ratios is significant at 10% level. On the other hand, the five-
bank, seven-bank, HTI, CCI and entropy’s negative relationship with bank efficiency is
significant at 5% level. In all cases, there is the existence of heteroscedasticity and
autocorrelation problems and thus, results are adjusted for eliminating such fundamental
problems by adopting the Prais–Winsten panel-estimation method. The Wald Chi2
statistic indicates the significance of each of the models at 1% level.
 The impact of competition on bank performance in Bangladesh 89

Table 8 Relationship between HHI and efficiency

Model β Std. Error z

ά 0.864 *** 0.157 5.50
HHI –2.499 ** 1.174 –2.13
GDPG –2.398 1.816 –1.32
BRANCH –0.000 0.001 –0.50
SIZE 1.133 1.321 0.86
NPL –0.324 ** 0.159 –2.03
AGE 0.005 ** 0.002 2.34
No. of banks 39
R-Squared 0.1029
Wald Chi2 25.37 ***
Notes: The symbols (***) and (**) represent statistical significance at 1% and 5%
level, respectively.

Table 9 Relationships between other competition measures and efficiency

Model β Std. Error z Wald Chi2

CRk1 –0.728 * 0.399 –1.83 24.18 ***
CRk3 –0.534 * 0.297 –1.80 24.16 ***
CRk5 –0.465 ** 0.233 –1.99 24.73 ***
CRk7 –0.514 ** 0.229 –2.25 25.78 ***
HTI –5.729 ** 2.464 –2.32 26.43 ***
CCI –0.715 ** 0.369 –1.94 24.55 ***
Entropy 0.182 ** 0.077 2.35 26.45 ***
Notes: The symbols (***), (**) and (*) represent statistical significance at 1%, 5%
and 10% level, respectively.

6 Conclusion

This study aimed at investigating the impact of competition on bank performance to fill
the gap in the existing banking literature in Bangladesh. Based on the empirical evidence,
it can be concluded that competition negatively influences the performance of banks in
terms of profitability and efficiency. The continuous reduction of bank spread margin
under high rate of inflation (see Suzuki and Adhikary, 2010), and the degree of
accelerated competition contribute to such negative impact on performance. Moreover, it
is also highly likely that the non-price competition in the form of branch expansion and
advertisement suggested by Hellmann et al. (2000, p.157) also exists in the banking
sector of Bangladesh as most of the private banks have been experiencing the expansion
phase. Such non-price competition may favourably affect the financial deepening of a
developing country like Bangladesh where financial sector is mostly underdeveloped in
nature. However, due to the absence of deposit rate controls under liberalisation, the
inefficiencies created by the non-price competition may outperform the gains from
financial deepening by creating inferior substitutes for customers and by eroding the
franchise value of banks, particularly private banks.
90 S.M.S. Uddin and Y. Suzuki

Therefore, based on the empirical evidence, it can be argued that the regulatory
authority needs to ensure a structure for the banking sector with necessary incentive for
banks, particularly for private banks, to improve their profitability and efficiency. No
doubt, various changes under financial deregulation have been successful in accelerating
the level of competition. The issue is whether this changing level of competition can
enhance the banking sector performance in a developing country like Bangladesh. It is
also evident that Bangladesh Bank is going to deregulate the banking sector further by
permitting nine new private banks to enter into the market, as reported in Section 3. Such
decision accelerates the competition level in the future, which may result in a decreased
performance. Moreover, the issuance of BRPD circular letter no. 13 (mentioned in
Section 1) illustrates the importance of encouraging private and foreign banks to
stimulate agriculture and SME sectors for the economic prosperity of the country. How
can it be done? Based on the findings, there is no incentive for banks at this point of time.
Under these circumstances, it is vital for the regulatory authorities to revisit the structural
changes in order to avoid the occurrence of banking sector instability argued by the
existing literature.

References
Abbasoglu, O.F., Aysan, A.F. and Gunes, A. (2007) Concentration, Competition, Efficiency and
Profitability of the Turkish Banking Sector in the Post-Crises Period, MPRA Paper No. 5494.
Available online at: http://mpra.ub.uni-muenchen.de/5494/5491/MPRA_paper_5494.pdf
Adhikary, B.K. (2011) ‘Foreign direct Investment, governance, and economic growth: a panel
analysis of Asian Economies’, Asia Pacific World, Vol. 2, No. 1, pp.72–94.
Alkdai, H.K.H. and Hanefah, H.M.M. (2012) ‘Board of Director’s characteristics and the earnings
management in Malaysian Shariah compliant listed companies’, Paper presented at the 2nd
International Conference on Arts, Social Sciences & Technology, Penang, Malaysia.
Ataullah, A., Cockerill, T. and Le, H. (2004) ‘Financial liberalization and bank efficiency:
a comparative analysis of India and Pakistan’, Applied Economics, Vol. 36, No. 17,
pp.1915–1924.
Ataullah, A. and Le, H. (2006) ‘Economic reforms and bank efficiency in developing countries:
the case of the Indian banking industry’, Applied Financial Economics, Vol. 16, No. 9,
pp.653–663.
Avkiran, N.K. (1999) ‘The evidence on efficiency gains: the role of mergers and the benefits to the
public’, Journal of Banking and Finance, Vol. 23, pp.991–1013.
Avkiran, N.K. (2000) ‘Rising productivity of Australian trading banks under deregulation
1986–1995’, Journal of Economics and Finance, Vol. 24, No. 2, pp.122–140.
Bangladesh Bank (2011) Annual Report 2010–2011, Bangladesh Bank, Dhaka, Bangladesh.
Bauer, P.W., Berger, A.N., Ferrier, G.D. and Humphrey, D.B. (1998) ‘Consistency conditions for
regulatory analysis of financial institutions: a comparison of frontier efficiency methods’,
Journal of Economics and Business, Vol. 50, pp.85–114.
Beck, T., Demirguc-Kunt, A. and Levine, R. (2006) ‘Bank concentration, competition and crises:
first results’, Journal of Banking and Finance, Vol. 30, pp.1581–1603.
Berger, A.N. (1995) ‘The profit-structure relationship in banking: tests of market-power and
efficient-structure hypotheses’, Journal of Money, Credit and Banking, Vol. 27, No. 2,
pp.404–431.
Berger, A.N., Demirguc-Kunt, A., Levine, R. and Haubrich, J.G. (2004) ‘Bank concentration
and competition: an evolution in the making’, Journal of Money, Credit and Banking, Vol. 36,
No. 3, pp.433–451.
 The impact of competition on bank performance in Bangladesh 91

Berger, A.N. and Hannan, T.H. (1989) ‘The price-concentration relationship in banking’,
The Review of Economics and Statistics, Vol. 71, No. 2, pp.291–299.
Berger, A.N. and Humphrey, D.B. (1997) ‘Efficiency of financial institutions: international survey
and directions for future research’, European Journal of Operational Research, Vol. 98,
pp.175–212.
Berger, A.N. and Mester, L.J. (1997) ‘Inside the black box: what explains differences in the
efficiencies of financial institutions?’, Journal of Banking and Finance, Vol. 21, pp.895–947.
Bhattacharyya, A., Lovell, C.A.K. and Sahay, P. (1997) ‘The impact of liberalization on the
productive efficiency of Indian commercial banks’, European Journal of Operational
Research, Vol. 98, pp.332–345.
Bikker, J.A. and Haaf, K. (2002) ‘Measures of competition and concentration in the banking
industry: a review of the literature’, Economic and Financial Modelling, Vol. 9, pp.53–98.
Boyd, J.H., Nicolo, G.D. and Smith, B.D. (2004) ‘Crises in competitive versus monopolistic
banking systems’, Journal of Money, Credit, and Banking, Vol. 36, No. 3, pp.487–506.
Brozen, Y. (1982) Concentration, Merger and Public Policy, Macmillan, New York.
Calem, P.S. and Carlino, G.A. (1991) ‘The concentration/conduct relationship in bank deposit
markets’, The Review of Economics and Statistics, Vol. 73, No. 2, pp.268–276.
Casu, B., Ferrari, A. and Zhao, T. (2010) ‘Financial reforms, competition, and risk in banking
markets’, in Fiordelisi, F., Molyneux, P. and Previati, D. (Eds): New Issues in Financial and
Credit Markets, Palgrave Macmillan, Basingstoke, pp.111–120.
Casu, B. and Molyneux, P. (2003) ‘A comparative study of efficiency in European banking’,
Applied Economics, Vol. 35, No. 17, pp.1865–1876.
Charnes, A., Cooper, W.W. and Rhodes, E. (1978) ‘Measuring the efficiency of decision making
units’, European Journal of Operational Research, Vol. 2, pp.429–444.
Cooper, W.W., Seiford, L.M. and Tone, K. (2007) Data Envelopment Analysis: A Comprehensive
Text with Models, Applications, References and DEA-Solver Software, 2nd ed., Springer,
New York.
Debnath, R.M. (2004) Banks and Legal Environment, Nabajuga Prokashani, Dhaka, Bangladesh.
Demsetz, H. (1973) ‘Industry structure, market rivalry, and public policy’, Journal of Law and
Economics, Vol. 16, No. 1, pp.1–9.
Drake, L., Hall, M.J.B. and Simper, R. (2009) ‘Bank modelling methodologies: a comparative non-
parametric analysis of efficiency in the Japanese banking sector’, Journal of International
Financial Markets, Institutions and Money, Vol. 19, No. 1, pp.1–15.
Farrell, M.J. (1957) ‘The measurement of productive efficiency’, Journal of the Royal Statistical
Society, Vol. 120, No. 3, pp.253–290.
Fraser, D.R. and Fraser, L.M. (1990) Evaluating Commercial Bank Performance: A Guide to
Financial Analysis, Bankers Publishing Co., New York.
Galagedera, D.U.A. and Edirisuriya, P. (2005) ‘Performance of Indian commercial banks
(1995–2002)’, South Asian Journal of Management, Vol. 12, No. 4, pp.52–74.
Gilbert, R.A. (1984) ‘Bank market structure and competition: a survey’, Journal of Money, Credit
and Banking, Vol. 16, No. 4, pp.617–645.
Hadad, M.D., Hall, M.J.B., Kenjegalieva, K.A., Santoso, W. and Simper, R. (2010) ‘Banking
efficiency and stock market performance: an analysis of listed Indonesian banks’, Review of
Quantitative Finance and Accounting, Vol. 37, No. 1, pp.1–20.
Heggestad, A.A. and Mingo, J.J. (1977) ‘The competitive condition of U.S. banking markets and
the impact of structural reform’, The Journal of Finance, Vol. 32, No. 3, pp.649–661.
Hellmann, T.F., Murdock, K.C. and Stiglitz, J.E. (1997) ‘Financial restraint: toward a new
paradigm’, in Aoki, M., Kim, H-K. and Okuno-Fujiwara, M. (Eds): The Role of Government
in East Asian Economic Development: Comparative Institutional Analysis, Clarendon Press,
Oxford, pp.163–207.
92 S.M.S. Uddin and Y. Suzuki

Hellmann, T.F., Murdock, K.C. and Stiglitz, J.E. (2000) ‘Liberalization, moral hazard in banking,
and prudential regulation: are capital requirements enough?’, The American Economic
Review, Vol. 90, No. 1, pp.147–165.
Lloyd-Williams, D.M., Molyneux, P. and Thornton, J. (1994) ‘Market structure and performance in
Spanish banking’, Journal of Banking and Finance, Vol. 18, pp.433–443.
Mason, E.S. (1939) ‘Price and production policies of large-scale enterprise’, The American
Economic Review, Vol. 29, No. 1, pp.61–74.
Molyneux, P., Altunbas, Y. and Gardener, E.P.M. (1996) Efficiency in European Banking, John
Wiley & Sons, London.
Neal, P. (2004) ‘X-efficiency and productivity change in Australian banking’, Australian Economic
Papers, Vol. 43, No. 2, pp.174–191.
Northcott, C.A. (2004) Competition in Banking: A Review of the Literature, Bank of Canada
Working Paper 2004–24. Available online at: http://www.bankofcanada.ca/en/res/wp/
2004/wp2004-2024.pdf
Ondrich, J. and Ruggiero, J. (2001) ‘Efficiency measurement in the stochastic frontier model’,
European Journal of Operational Research, Vol. 129, pp.434–442.
Park, K.H. (2009) ‘Has bank consolidation in Korea lessened competition?’, The Quarterly Review
of Economics and Finance, Vol. 49, No. 2, pp.651–667.
Perera, S., Skully, M. and Wickramanayake, J. (2006) ‘Competition and structure of South Asian
banking: a revenue behaviour approach’, Applied Financial Economics, Vol. 16, No. 11,
pp.789–801.
Rose, P.S. and Fraser, D.R. (1976) ‘The relationships between stability and change in market
structure: an analysis of bank prices’, The Journal of Industrial Economics, Vol. 24, No. 4,
pp.251–266.
Ruggiero, J. (2007) ‘A comparison of DEA and the stochastic frontier model using panel data’,
International Transactions in Operational Research, Vol. 14, pp.259–266.
Samad, A. (2005) ‘Banking structure and performance: evidence from Utah’, Review of Business
Research, Vol. 5, No. 2, pp.151–156.
Samad, A. (2008) ‘Market structure, conduct and performance: evidence from the Bangladesh
banking industry’, Journal of Asian Economics, Vol. 19, No. 2, pp.181–193.
Sathye, M. (2001) ‘X-efficiency in Australian banking: an empirical investigation’, Journal of
Banking and Finance, Vol. 25, pp.613–630.
Sathye, M. (2003) ‘Efficiency of banks in a developing economy: the case of India’, European
Journal of Operational Research, Vol. 148, No. 3, pp.662–671.
Seelanatha, L. (2010) ‘Market structure, efficiency and performance of banking industry in
Sri Lanka’, Banks and Bank Systems, Vol. 5, No. 1, pp.20–31.
Sherman, H.D. and Gold, F. (1985) ‘Bank branch operating efficiency: evaluation with data
envelopment analysis’, Journal of Banking and Finance, Vol. 9, pp.297–315.
Smirlock, M. (1985) ‘Evidence on the (non)relationship between concentration and profitability in
banking’, Journal of Money, Credit and Banking, Vol. 17, No. 1, pp.69–83.
Sturm, J-E. and Williams, B. (2004) ‘Foreign bank entry, deregulation and bank efficiency:
lessons from the Australian experience’, Journal of Banking and Finance, Vol. 28, No. 7,
pp.1775–1799.
Suzuki, Y. and Adhikary, B.K. (2010) ‘A ‘bank rent’ approach to understanding the development
of the banking system in Bangladesh’, Contemporary South Asia, Vol. 18, No. 2, pp.155–173.
Suzuki, Y., Miah, M.D. and Yuan, J. (2008) ‘China’s non-performing bank loan crisis: the role of
economic rents’, Asian-Pacific Economic Literature, Vol. 22, No. 1, pp.57–70.
Tone, K. (2001) ‘A slacks-based measure of efficiency in data envelopment analysis’, European
Journal of Operational Research, Vol. 130, No. 3, pp.498–509.
 The impact of competition on bank performance in Bangladesh 93

Uddin, S.M.S. and Gupta, A.D. (2012) ‘Concentration and competition in the non-banking sector:
evidence from Bangladesh’, Global Journal of Management and Business Research, Vol. 12,
No. 8, pp.81–88.
Uddin, S.M.S. and Suzuki, Y. (2011) ‘Financial reform, ownership and performance in banking
industry: the case of Bangladesh’, International Journal of Business and Management, Vol. 6,
No. 7, pp.28–39.
Wanniarachchige, M.K. and Suzuki, Y. (2010) ‘Bank competition and efficiency: the case of Sri
Lanka’, Asia Pacific World, Vol. 1, No. 1, pp.117–131.
Wong, J., Fong, T.P-W., Wong, E.T-C. and Choi, K-F. (2008) ‘Determinants of the performance of
banks in Hong Kong’, in Genberg, H. and Hui, C-H. (Eds): The Banking Sector in Hong
Kong: Competition, Efficiency, Performance and Risk, Palgrave Macmillan, Basingstoke,
pp.50–65.
Worthington, A.C. (2000) ‘Cost efficiency in Australian non-bank financial institutions: a non-
parametric approach’, Accounting and Finance, Vol. 40, No. 1, pp.75–97.
Wu, S. (2005) ‘Productivity and efficiency analysis of Australian banking sector under
deregulation’, Proceedings of the Australian Conference of Economists 2005, University
of Melbourne. Available online at: http://editorialexpress.com/cgi-bin/conference/
download.cgi?db_name=ACE2005&paper_id=2192
Zhao, T. and Murinde, V. (2011) ‘Bank deregulation and performance in Nigeria’, African
Development Review, Vol. 23, No. 1, pp.30–43.

Notes
1 See October–December Issues 1976 and 2011 of the Scheduled Bank Statistics, a quarterly
publication of Bangladesh Bank.
2 1983 is the year in which both private and denationalised banks started their operations
simultaneously.
3 These studies consider interest expenses and non-interest expenses as inputs and net interest
income and non-interest income as outputs.
4 See Drake et al. (2009, p.7) and Hadad et al. (2010, p.4).
5 Ondrich and Ruggiero (2001), Galagedera and Edirisuriya (2005) and Ruggiero (2007) also
use the CRS assumption for their studies.
6 See Ruggiero (2007, p.259).
7 232 branches of the state-owned banks have either merged or closed down under the policy
initiative.
8 According to the horizontal merger guideline, if the index is less than 1000; the market can be
regarded as a very low concentrated market; if the index is between 1000 and 1800, the
market can be treated as a moderately concentrated market; and if the index is more than
1800, the market can be considered as a highly concentrated market.
 94 S.M.S. Uddin and Y. Suzuki

Appendix A

Table A1 The Jarque-Bera statistic

Before Transformation
ROA HHI GDPG BRANCH
Jarque-Bera Statistic 14.2886 1.1699 1.8421 2.6600
Probability 0.0008 0.5571 0.3981 0.2645
After Transformation
ROAT HHI GDPG BRANCH
Jarque-Bera Statistic 1.7334 1.1699 1.8421 2.6600
Probability 0.4203 0.5571 0.3981 0.2645

Table A2 Variance Inflation Factors (VIFs)

HHI and ROA HHI and Efficiency

Variables VIF Variables VIF
HHI 2.06 HHI 2.15
GDPG 2.05 GDPG 1.94
BRANCH 1.49 BRANCH 6.05
SIZE 5.65
NPL 1.34
AGE 1.93

Table A3 Correlations among different competition proxies

HHI CRk1 CRk3 CRk5 CRk7 HTI CCI Entropy

HHI 1.0000
CRk1 0.9489 1.0000
CRk3 0.9893 0.9763 1.0000
CRk5 0.9901 0.9640 0.9907 1.0000
CRk7 0.9907 0.9535 0.9840 0.9973 1.0000
HTI 0.9753 0.8605 0.9355 0.9523 0.9625 1.0000
CCI 0.9918 0.9794 0.9962 0.9959 0.9922 0.9436 1.0000
Entropy –0.9907 –0.9227 –0.9749 –0.9875 –0.9876 –0.9806 –0.9798 1.0000

View publication stats