EFFICIENCY OF ISLAMIC BANKS IN MALAYSIA

MARIANI ABDUL MAJID*
NOR GHANI MD. NOR**
FATHIN FAIZAH SAID

ABSTRACT

In recent years, Malaysian Islamic banks have to operate in an increasingly competitive environment. This trend is expected to continue as the competition from conventional banks picks up, partly in response to the ASEAN Free Trade Agreement (AFTA), but also in response to the general globalization of markets. How Islamic banks will be affected by the increased competitive pressures depends in part on how efficiently they are run. This paper examines the productive efficiency of Malaysian commercial (Islamic and conventional) banks over the 1993 to 2000 time period. The goal of the analysis is to identify the efficiency level of Islamic banks compared to other commercial banks in Malaysia. Using the stochastic frontier cost function approach, efficiency scores were determined for each bank to determine their relative positions on the efficiency ladder. An inefficiency model is estimated to link the inefficiency of input use to produce output to other factors. The results show that the efficiency level of Islamic banks is not statistically different from the conventional banks. In addition there is no evidence to suggest that bank efficiency is a function of ownership status i.e. public or private, and foreign or local.
1. INTRODUCTION
During the early years after independence, foreign banks dominated the Malaysian banking scene. For instance, two years after independence (1959), there were only 8 local banks out of the 26 commercial banks in operation. Twenty years later, the role played by the local banks improve considerably when the number of commercial banks increased to 38 with 21 domestic banks and 16 foreign banks in 1982. The number shrank considerably (although the split between local and foreign banks is maintained) in the 1990s following a spate of ‘mergers and take-overs’ as banks try to secure particular niches and economies of scale in order to stay competitive. By August 2001, there were 27 commercial banks in total including 2 Islamic banks.

Along side the conventional banking system, the Islamic banking system has steadily evolved to become a major player on the Malaysian financial landscape. Its birth could be traced back to the establishment of the Pilgrimage Fund and Management Council (LUTH) in 1969; although its function was then limited to mobilizing savings for Muslims who wish to perform Hajj in Mecca. The establishment of Bank Islam Malaysia Berhad (BIMB) in 1983 really marks the beginning of a new era in Islamic banking in Malaysia. The year witnessed the creation of a full-fledged Islamic banking institution offering financial services in accordance with Islamic jurisprudence. The Islamic banking system received a further boost when a scheme where conventional banks can open windows for Islamic banking products called Interest-free Banking Scheme was launched in 1993. Another important event for Islamic banking in Malaysia occurred in October 1999 when the second Islamic bank, Bank Muammalat Malaysia Berhad (BMMB) opened its door to the public. By the end of 2001, there were 36 conventional banking institutions offering Islamic banking products in Malaysia (14 commercial banks, 10 finance companies, 5 merchant banks and 7 discount houses) besides 2 full-fledged Islamic banks (Bank Negara Malaysia, 2001).

Despite its growth, the share market of the Islamic banking system in Malaysia is relatively small compared to the size of the overall banking system. Herein lies one of the reasons for assessing the efficiency of Islamic banks relative to their conventional counterparts. Efficient operation will not only ensure its survival through growth but will also practically demonstrate that the Islamic financial system is the alternative to the currently adopted system. By end of October 2001, the total assets of RM59.0 billion belonging to Islamic banking institutions only represents 8.2 percent of the total banking assets (RM717 billion). Similarly, the role of Islamic banks in financing activities is still small compared to the conventional banks where it accounted for a mere 6.5% (RM28 billion) from a total of RM432 billion for the whole banking system. Total deposits in the Islamic banking system amounted to RM47 billion (9.5 percent) compared to RM495 billion total deposits in the entire system (Abdul Kader, 2002).

The gale of economic liberalization currently sweeping across the globe provides another reason for looking into the efficiency of Islamic banks since it poses a further competitive challenge to the Islamic banking institutions. It is expected that by the year 2007, trade liberalization will require Malaysian domestic banks to compete with other global players on an equal playing field. Such a change implies that local Islamic banking institutions will have to be efficient, innovative, competitive and resilient players in the market. This is particularly important since foreign banks that are arguably more efficient may also offer Islamic banking products to take advantage of brisk demand for Islamic banking products.

The policy relevance of this study can be viewed from the target set by the Malaysian central bank (Bank Negara Malaysia) for the Islamic banking institutions to command a 20 percent market share by year 2010 as outlined in its financial sector master plan. How the increased competition expected in the future affects Islamic banks partly depends on how efficient the banks operate in the new competitive environment. Efficient banks will gain competitive advantage while the inefficient ones may be driven out of the market. Islamic banks are arguably placed at a very disadvantaged position since they are already currently operating in a restricted environment. The opportunity set available to the managers of Islamic banks is smaller in the sense that conventional banks can adopt any business strategy available to Islamic banks and not vice versa. Of course, this may not necessarily mean that the Islamic banks are going to be less efficient but it does open up the possibility of such state of affair. Despite its importance, the study of bank efficiency particularly Islamic banks in Malaysia is clearly lacking in the literature.

This paper aims at determining the relative efficiency of Islamic banks to other commercial banks in Malaysia, using the stochastic cost frontier approach. The paper unfolds as follows. The following section reviews several efficiency studies on banking services. Section 3 describes the model, estimation technique and the efficiency indicators used. Section 4 presents the empirical results and the paper ends with some final remarks in Section 5.
2. LITERATURE REVIEW ON EFFICIENCY STUDIES IN BANKING SERVICES
A survey of the literature indicates that there are numerous empirical studies that try to look into the relative efficiency of financial institutions operating within a country (and in some cases between countries). Generally, these studies differ in the specific methodology used and the type of banks considered in the study sample. This section reviews some of these studies by conveniently dividing them into works on bank efficiency in general and comparative efficiency between Islamic and conventional banks.

Using stochastic cost frontier methodology, Mendes and Rebelo (1999) study the performance of 221 Portugese banks in the 1990-1995 period. They found that increased competition faced by banks during the period under study did not lead to an improvement in the overall cost efficiency. The study also found no clear evidence on the existence of a predictable association between size and cost efficiency, and banks could in fact remain competitive by exerting firm control over cost. In a similar study (Zaim, 1995) found that the financial reform introduced in Turkey in the mid-eighties had succeeded in prodding commercial banks into taking positive measures to enhance technical as well as allocative efficiencies. By comparing performance between the pre and post liberalization era, the study observes that banks in Turkey recorded some improvement in efficiency during the 1980-1990 period. Lovell and Grifell-Tatje (1997) used Spanish banking data over 1986-1993 period to show that savings banks reacted to banking deregulation by engaging in efficiency-enhancing merger activities. During the same period, the commercial banks instead chose to adjust their loan and deposit interest rates as a way of fending off increased competition from savings banks. Savings banks were also found to have a more efficient form of organizational structure compared to the commercial banks. Chang et. al. (1998) found that foreign-owned multinational banks operating in the U.S. were significantly less efficient than their local counterparts. The study used 1984-1989 banking data and translog stochastic cost frontier method to estimate the cost inefficiency scores. In another study, Mester (1993) uses the stochastic econometric cost frontier approach to investigate efficiency of banks operating in the Third Federal Reserve District in the U.S. using 1991-1992 data. Findings from the study suggests that there is less to be gained in terms of cost savings from changing output size or mix than from using inputs more cost efficiently.

Kwan and Eisenbeis (1996) used semi-annual data for a sample of 254 bank holding companies from 1986 to 1991 in order to find the relationship between scale and efficiency. The banks were grouped into size-based quartile to allow for different production technologies for each size class. Separate cost functions was estimated for each size quartile using the method of maximum likelihood. After controlling for other factors, smaller banking firms on average were found to be less efficient than larger banking firms. Smaller banking firms tended to exhibit larger variations in inefficiencies than larger firms do. These findings suggest that on average, large banking firm operates closer to its respective efficient frontier than the small banking firms. Average inefficiency appears to decline over the period 1986 to mid 1990, apparently responding to the increased competition in banking wrought by market and regulatory changes. Persistence of inefficiency rankings suggests that relatively efficient (inefficient) banking firms tend to stay relatively efficient (inefficient) over a fairly long period.

In one cross border efficiency study, Abd. Karim (2001) found that larger banks tend to be more cost efficient compared to their smaller rivals although the degree of advantage obtained from scale economies decreases for a select ASEAN countries (Indonesia, Malaysia, the Philippines and Thailand). The study employed the stochastic cost frontier approach.

In a study on the risk of failure faced by the inefficient banks Cebenoyan, Cooperman and Register (1993) found that banks and S&Ls (Savings and Loans) in the U.S. with low efficiency estimates face failure at greater rates compared to those with higher efficiency estimates. Similarly, in a study of bank failures during 1920s, Wheelock and Wilson (1995) find that the less technically efficient a bank was, the greater likelihood of failure. Such findings are also reported by Berger & Humprey (1997) in their survey of financial institutions efficiency literature where many studies support that most bank failures are directly related to poor management quality, a low capital position, a large number of problem loans and a weak or negative cash flow.

A number of studies on the relative efficiency of Islamic banks efficiency or productivity could also be found in the literature. Unlike the current study, these studies tend to use general comparisons of accounting ratios as a method of analysis. ; either Islamic bank is more efficient or productive than the conventional banks, or the reverse. Hamid (1999) evaluates the performance of one of the five Islamic banks in Bangladesh, IBBL in comparison with two comparable conventional banks, ABBL and PBL using accounting ratios. IBBL is found to be more productive in terms of total income compared to the other two conventional banks. However, one of the two conventional banks under study, ABBL is better than the Islamic bank, IBBL in terms of the labour productivity. He reasoned that the comparatively new Islamic bank (IBBL) incurred higher cost because of her strategy of hiring experienced bankers from the conventional banks. In another study (Abdus Samad & Hassan, 1999), the profitability performance of Islamic bank in Malaysia (BIMB) was compared to 8 conventional banks using the financial ratios. The study found that the average profit of BIMB is significantly lower than the conventional banks. The authors argued that this is to be expected since the investment opportunity set for BIMB in stocks and securities is smaller due to religious constraint. Sarker (1999) did a study on all Islamic based banks in Pakistan by developing the so-called Banking Efficiency Model. Efficiency measures on productivity, operation, allocation, distribution and stability were computed as financial ratios. The study concluded that Islamic banks couldn’t achieve maximum efficiency under the conventional banking framework because of constraints that they face. Metwally (1997) compared the performance of 15 conventional banks and 15 “riba” free or interest-free banks from all over the world. The study tested for structural difference between the two groups of banks in terms of liquidity, leverage, credit risk, profit and efficiency. The study found that “riba” free banks face more difficulties in attracting deposits compared to the conventional banks. Secondly, “ riba” free banks tend to be more conservative in utilizing funds for lending and are disadvantaged in terms of investment opportunities. As a result, this bank has higher cash to deposits ratio compared to conventional banks. However, statistical results suggest that profitability and efficiency differences are not statistically significant between the two types of banks.
3. METHODOLOGY AND DATA
In estimating cost frontiers, two general approaches can be found in the literature, namely non-parametric and parametric methods. The non-parametric data envelopment analysis (DEA) approach determines which bank in the sample produces a particular output combination at the given input prices at least cost. This defines the best practice bank for that output/ input prices combination. A bank’s relative efficiency is measured by the ratio of its own cost to the cost of the ‘best practice’ bank that faces the same input prices and produces the same output bundle. The advantage of using the DEA is that it doesn’t require an assumption on the distribution of the error terms. The disadvantage of using DEA is that it attributes any unfavourable influence, even those that are beyond a bank’s control, to inefficiency.

An alternative method, the one used in this study, is the parametric stochastic frontier approach developed by Aigner et al. (1977) and Meusen and Broeck (1977). This method postulates a two-component error structure; a controllable factor and a random uncontrollable factor. Consider the following cost function of a firm;








where CIE represents the total cost of the firm; yj represents various products or services produced by the firm; wk represents the prices of inputs used by the firm in the production of the products and services; represents random disturbance term, which allows the cost function to vary stochastically. The uncertainty being captured by the in the cost function can further be decomposed as follows.



where v represents random uncontrollable factors that affect the total cost (such as machine performance, weather and luck). They are identically distributed as normal variates and the value of the error term in the cost relationship is, on average equal to zero. On the other hand, the u represents individual firm cost deviations or errors that are due to factors that are under the control of the firm such as technical and allocative efficiency.

The stochastic frontier model postulates that a firm’s observed cost will deviate from the cost frontier because of random noise (uncontrollable component), vi, and possible inefficiency (controllable component), ui. The controllable inefficiencies is one-sided since it can only increase costs above the frontier or best practice levels while that for random fluctuation can either increase or decrease cost. Since uncontrollable factors are assumed to be symmetrically distributed, the cost frontier, f(yj,wk) + , is therefore stochastic. Thus when estimating a cost function with a composite error term, it is assumed that the sum of two-sided error represents random fluctuations in cost and a one-sided positive error represents inefficiency (see Mester (1996) for details).

In this paper the translog cost function is chosen since this is the most frequently used functional forms in banking efficiency studies (Abd. Karim, 2001). The translog stochastic cost frontier can be written as follows.

(3)
where,
ln C = the natural logarithm of the total cost;
ln y yt = natural logarithm of the jth output (j = 1, 2, 3,…n);
ln w kit = natural logarithm of the kth input price (k = 1, 2,…m);
t = year of observation; and
Bs are the coefficients to be estimated.


where,
In C = the natural logarithm of the total cost;
In y yt = natural logarithm of the jth output (j = 1, 2, 3,…n);
In w kit = natural logarithm of the kth input price (k = 1, 2,…m);
t = year of observation; and
Bs are the coefficients to be estimated.

The vit s are random variables due to measurement errors in the input variable or the effect of unspecified explanatory variables in the model, and the uit s are non-negative random variables, associated with inefficiency of input used, given the levels of outputs and the inputs. Most studies on inefficiency use what has now become to be known as the intermediation approach in measuring a bank flow of services. Under this approach output is measured in terms of the dollar volume of loans. Deposits are treated as inputs and loans as outputs (Sealey & Lindley, 1977). It is a popular approach partly because of the ease of acquiring required data. The intermediation approach includes both operating and interest expense or income distributed to the depositors in the case of Islamic bank. In equation 3, total cost (C) is defined to include all labour and capital expenses and interest expense or income distributed to depositors for the Islamic bank (see Al-Habshi (1999) for further discussion). The outputs are Ringgit amounts of loans, advances and financing; Ringgit amounts of deposits, placement with other financial institutions and cash and short-term funds; and Ringgit amount of securities and investments. Input prices include staff expenses per employee, expenses on land, building and equipment per Ringgit of assets, and expenses on interest or income distributed to depositors per Ringgit of deposits.

Following Jondrow et al. (1982), a measure of firm’s inefficiency can be written as;



where,



 = the standard normal density function;
 = the cumulative normal density function; and
 = 2u/2.

This measure of controllable inefficiency for ith firm (i.e. Ui), can then be used to examine the inefficiency effects (Chang et al., 1998). The cost inefficiency effects are assumed to be defined by;






where

D1 is the dummy variable for public/private ownership: D1=1 if an observation involves a government-owned bank, zero otherwise;
D2 is the dummy variable for Islamic bank: D2=1 if an observation involves an Islamic bank, zero otherwise;
D3 is dummy variable for foreign/local ownership. D3=1 if an observation involves a foreign owned bank, zero otherwise;
Z1 = logarithm of the size of bank
Z2 = square of logarithm of the size of bank
Z3 = interaction between bank size and public/private ownership
Z4 = interaction between bank size and foreign/local ownership
Z5 = year of observation

The specification (4) above implies that besides considering the effect of inefficiency in the use of input for producing the output, variations in the inefficiency effects are also influenced by other variables. This study allows for variations in the inefficiency effect to be different for bank ownership type (public/private as well as foreign/local), type of bank (Islamic/conventional), bank size and year of observation.

The data set is created using information obtained from banks’ annual reports and ABM Bankers Directory. The latter information source is used since some annual reports do not provide information on the number of employees. All monetary amounts are in Ringgit Malaysia (RM). Excluded from the sample were those banks with incomplete data. The eventual dataset contains complete data for 34 banks (24 local and 10 foreign) from a total of 55 commercial banks in operation during the period under study (i.e. from 1993 to 2000).

4. EMPIRICAL RESULTS

Before discussing the results of estimated cost frontier and inefficiency model, it maybe worthwhile to first look at some descriptive statistics (Table 1) for the inefficiency measures computed for several pertinent bank categories/classifications.

Table 1

Descriptive Statistics for Inefficiency Measures for
Various Bank Categories

Mean Std. Dev. Minimum Maximum
Conventional 0.302 0.240 0.030 1.230
Islamic 0.280 0.162 0.090 0.600

Local 0.321 0.248 0.050 1.230
Foreign 0.254 0.199 0.030 1.080

Asset Size > RM12 billion 0.319 0.209 0.110 0.860
Asset Size RM6-12 billion 0.290 0.269 0.090 1.200
Asset Size RM3-6 billion 0.409 0.338 0.120 1.230
Asset Size < RM3 billion 0.226 0.206 0.030 0.810

Notice that Islamic banks does marginally better than conventional banks in terms of efficiency although both produce at a cost that is 30.2% and 28% respectively higher than necessary. The slight edge achieved by the Islamic banks over conventional banks is not however statistically significant. This observation provides no evidence that Islamic banks may be less efficient compared to their conventional counterparts due to a more restrictive business environment.

It is also interesting to note that foreign banks are generally more efficient than local banks. Further test suggests that the difference is statistically significant at the 5% level. This finding has important implication on the future prospects of local Islamic (and conventional) banks when the banking market is open to more foreign bank participation.

There is no clear conclusion that can be derived from Table 1 regarding the effect of size on inefficiency. There is simply no clear observable pattern linking these two variables although the smallest bank appears to be the most efficient. However, as is the case for any bivariate relationship (including those discussed above) that fails to control for the effects of other relevant variables, this does not imply that no relationship exists between the variables. Indeed, results of the inefficiency model as displayed in Table 3 provide a more reliable representation of the relationship between inefficiency and several structural variables.

The stochastic cost frontier (Equation 3) was estimated using the maximum likelihood method and the results of the estimation are given in Table 1. Results of the inefficiency model (Equation 4) are reported in Table 2.


Table 2

Maximum Likelihood Estimates for the Translog Stochastic
Cost Frontier Model

Coeff. t-ratio
Constant -19.9081 -3.75
Deposits 0.4492 0.93
Investment -0.2797 -0.24
Loan 3.9513 2.57
Price Capital 0.8922 1.56
Price Labor -3.1278 -2.99
Price Borrowed Fund -1.8497 -1.41
Deposits2 -0.0263 -1.13
Loan2 -0.0640 -0.54
Investment2 0.0747 0.86
PriceCapital2 0.0061 0.40
PriceLabor2 0.0625 1.77
PriceBorrowedFund2 -0.2568 -2.72
Deposits X Loan -0.0469 -0.72
Deposits X Investment 0.0359 0.53
Loan X Investment -0.0932 -0.45
Deposits X Price Capital -0.0158 -0.36
Deposits X Price Labor 0.0640 0.87
Deposits X Price Borrowed Fund -0.1095 -1.08
Loan X Price Capital -0.0881 -0.87
Loan X Price Labor 0.0416 0.22
Loan X Price Borrowed Fund -0.0414 -0.18
Investment X Price Capital 0.0490 0.59
Investment X Price Labor 0.0835 0.54
Investment X Price Borrowed Fund 0.2736 1.46
Price Capital X Price Labor 0.0755 0.86
Price Capital X Price Borrowed Fund 0.0903 0.87
Price Labor X Price Borrowed Fund -0.0699 -0.35

Lambda 3.0157 3.99
Sigma 0.4193 11.54

Table 3
OLS Estimates of the Inefficiency Model

Coeff. t-ratio
Constant 16.7311 0.14
Islamic Bank 0.1241 1.36
BankSize -2.3763 -6.77
BankSize2 0.0802 7.55
Private 0.3247 0.39
Private X Bank Size -0.0172 -0.33
Foreign 0.6610 1.31
Foreign X bank Size -0.0404 -1.23
Year 0.1241 0.01

To the extent that there is an overriding interest in the empirically estimated inefficiency model, much of this section is devoted to discussing the results presented in Table 3.

The coefficient on the Islamic Bank dummy is positive but not statistically significant, suggesting that Islamic banks are no less efficient than their conventional counterparts. This finding is comforting to proponents of Islamic banking since it implies that Islamic banks are equally cost efficient (or inefficient!) even when the business environment is some sense more restrictive.

Both bank size coefficients (Bank Size and BankSize2) are statistically significant. Not only does bank size influence inefficiency, it also does so in a non-linear fashion. Taken together, negative coefficient for Bank Size and positive for BankSize2, it appears that increasing size initially provides some scale economies before diseconomies of scale sets in once a critical size is reached. Such a relationship between size and inefficiency provide some evidence for a U-shaped average cost function.

Ownership status (public versus private) appears to make no difference in terms of inefficiency since the coefficients on Private and Private interacted with Bank Size variables are not statistically significant. This finding strengthens the consensus in the literature over the relative efficiency of private and public firms that there are no clear superiority of one form of organization over another when it comes to efficiency.

The Foreign and Foreign X Bank Size interaction variables are also not statistically significant suggesting that foreign banks are no more efficient than local banks. This finding augurs well for the future prospect of locally owned banks when the market is fully liberalized. However, such a statement of optimism must be taken with cautioned since it is not known whether foreign banks may become more efficient once the existing restrictions on foreign banks are lifted (e.g. limits on the number of branches).
5. CONCLUSION
This study attempts to determine the relative efficiency of Islamic banks to other commercial banks in Malaysia by employing the stochastic cost frontier approach as the means of analysis. An inefficiency model is estimated to link the inefficiency of input use to produce output to other factors including, most importantly perhaps, whether a bank is Islamic or otherwise. This study finds that there is no statistically significant difference in the level of efficiency between Islamic and conventional banks operating in Malaysia based on data from 1993 to 2000. There is also no evidence to suggest that bank efficiency is a function of ownership status (public/private or foreign/local). This study does however find that inefficiency is related to bank size and in a non-linear fashion. Increasing size initially provides some scale economies before diseconomies of scale sets in once a critical size is reached thus suggesting a U-shaped average cost function.



REFERENCES

Abd Karim, M. Z. (2001), ‘Comparative Bank Efficiency across Select ASEAN Countries’, ASEAN Economic Bulletin, 18 (3), 289-304.
Abdul Kader, M. R. (2002), ‘The Operation of Islamic Banking Master Plan’, Experts Meeting on Towards 20 percent Market Share, Bank Negara Malaysia Strategy, Kuala Lumpur.
Abdus Samad and M. K. Hassan (1999), ‘The Performance of Malaysian Islamic Bank during 1984-1997: An Exploratory Study’, International Journal of Islamic Financial Services, 1 (3), http://islamic-finance.net/journal.html
Aigner, D. J., C. A. K. Lovell and P. Schmidt (1977), ‘Formulation and Estimation of Stochastic Frontier Production Function Models’, Journal of Econometrics, 6, 21-37.
Association of Banks in Malaysia (Various issues), ABM Bankers Directory.
Bank Negara Malaysia (2001), Annual Report.
Berger, A. N. and D. B. Humprey (1997), Efficiency of Financial Institutions: International Survey and Directions For Future Research, USA: Wharton Financial Institutions Centre, The Wharton School, University of Pennsylvania.
Cebenoyan, A. S., Cooperman E. S. and C. A. Register (1993), ‘Firm Efficiency and the Regulatory Closure Of S & Ls: An Empirical Investigation’, Review of Economics and Statistics, 75, 540-545.
Chang, C. E., I. Hasan and W. C. Hunter (1998), ‘Efficiency of Multinational Banks: an Empirical Investigation’, Applied Financial Economics, 8, 689-696.
Farrell, M. J. (1957), The Measurement of Productive Efficiency’, Journal of the Royal Statistical Society, Series A (120), 253-281.
Habshi-al, S. M. (1999), ‘Financial Performance Measurement and Distribution Policy of Islamic Financial Institutions’, International Conference on Islamic Economics in the 21st Century, Kuala Lumpur.
Hamid, M. A. (1999), ‘Islamic Banking in Bangladesh Expectations and Realities’, International Conference on Islamic Economics in the 21st Century, Kuala Lumpur.
Jondrow, J., C. A. K. Lovell, I. S. Materov and P. Schimidt (1982), ‘On the Estimation of Technical Efficiency in the Stochastic Frontier Production Function Model’, Journal of Econometrics, 19, 233-238.
Kumbhakar, S. C. and C. A. K. Lovell (2000), Stochastic Frontier Analysis, Cambridge: Cambridge University Press.
Kwan, S. H and R. A. Eisenbeis (1996), ‘An Analysis Of Inefficiencies In Banking: A Stochastic Cost Frontier Approach’, Federal Reserve Bank of San Francisco Economic Review, 2, 16-26.
Lovell, C. A. K. and E. Grifell-Tatje (1997), ‘The Sources of Productivity Change in Spanish Banking’, European Journal of Operational Research, 98, 364-380.
Mendes, V. and J. Rebelo (1999), ‘Productive Efficiency, Technological Change and Productivity in Portuguese Banking’, Applied Financial Economics, 9, 513-521.
Mester, L. J. (1993), Efficiency of Banks in the Third Federal Reserve District, USA: Wharton Financial Institutions Centre, The Wharton School, University of Pennsylvania.
Mester, L. J. (1996), ‘A Study of Bank Efficiency Taken into account Risk- Preference’, Journal of Banking and Finance, 20, 1025-1045.
Metwally, M. M. (1997), ‘Differences between the Financial Characteristics of Interest-free Banks and Conventional Banks’, European Business Review, 97 (2), 92.
Meusen, W. and V. D. Broeck (1977), ‘Efficiency Estimation from Cobb-Douglas Production Functions with Composed Error’, International Economic Review, 18(2), 435-444.
Sarker, M. A. A. (1999), ‘Islamic Banking in Bangladesh: Performance, Problems and Prospects’, International Journal of Islamic Financial Services, 1(3), http://islamic-finance.net/journal.html
Sealey, C. W. and J. T. Lindley (1977), ‘Inputs, Outputs, and A Theory of Production and Cost at Depository Financial Institutions’, The Journal of Finance, XXXII (4), 1251-1266.
Wheelock, D. C. and Wilson P. W. (1995), ‘Evaluating the Efficiency of Commercial Banks: Does Our View of What Banks Do Matter?’, Review (Federal Reserve Bank of Saint Louis), 77(4), 39-52.
Zaim, O. (1995), ‘The Effect of Financial Liberalization on the Efficiency of Turkish Commercial Banks’, Applied Financial Economics, 5, 257-264.