Challenges Facing the Polish Banking Industry: A Comparative Study with uk Banks Catarina Figueira Joseph G. Nellis David Parker In 2004 Poland entered the eu. This paper investigates the performance of the Polish banking industry over the period 1999-2004, by looking specifically at its comparative efficiency in relation to one of the largest banking sectors in the eu namely, that of the uk. Based on a range of efficiency measures, the empirical results reveal a surprising degree of relative efficiency in the Polish banking industry, no doubt reflecting the substantial economic changes introduced in Poland since 1989. The findings suggest that the Polish banking sector should be able to with- stand the new competitive pressures that it faces following entry into the banking sector of the eu. Key Words: Poland, uk, banking, efficiency, performance jel Classification: C52, F36, G21 Introduction Poland entered the eu as one of a number of new Member States in 2004. Entry into the eu implies increased competitive pressures for the Polish corporate sector created by the European Single Market and eu competi- tion law. This is particularly true for Poland's financial services that until recently were state owned and protected from competition. The banking industry has been transformed in the eu during the last decade as a result of three major developments: (a) the establishment of a Single European Market in financial services, which has intensified competitive pressures and forced the pace of rationalization across the industry; (b) the im- pact of developments concerning information technology and the con- Catarina Figueira is Research Fellow in Economics at the Cranfield School of Management, Cranfield University, United Kingdom. Joseph G. Nellis is Professor of International Management Economics at the Cranfield School of Management, Cranfield University, United Kingdom. David Parker is Professor ofPrivatisation and Regulation at the Cranfield School of Management, Cranfield University, United Kingdom. Managing Global Transitions 5 (1): 25-44 sequences for the delivery of financial products and services, as well as new product development (involving, for example, internet banking and money transmission services); and (c) extensive merger activity, bring- ing about closer integration and, to a large degree, the globalization of financial markets. This has created a business environment in which in- stitutional investors are now challenging the dominant positions of com- mercial banks in both deposit-taking and loan-financing facilities. Also, Poland is by far the largest country amongst the new eu Member States and, therefore, can be expected to attract considerable attention from the European financial services industry, as its economy develops and gravi- tates towards the eu average. The purpose of this paper is to consider the likely competitive pres- sures facing the Polish banking industry in the future. The contributions of the paper are in terms of identifying the relative competitiveness of Poland's banking sector and in applying a number of measures includ- ing stochastic cost frontier analysis. To make the research manageable, the efficiency of Poland's banking sector is compared with the efficiency of banking in the uk. The uk's banking sector is one of the largest in the eu1 and is generally recognized as internationally competitive. It there- fore provides a useful benchmark for comparing the efficiency of Polish banks with those of the eu in general. The alternative, of comparing Polish banking with the average perfor- mance of banking across the eu, would arguably be less satisfactory be- cause average eu performance masks differences within the eu. A num- ber of performance measures are used, namely financial ratios, including profitability, and figures for operating and financing costs. Later in the paper, performance differences are also investigated using a stochastic cost frontier analysis. The paper concludes that the Polish banking sec- tor seems already broadly comparable to the uk's banking sector in many areas of performance. It is, however, still relatively small scale, and com- petition is not as developed as it is in the uk when measured in terms of the number of competing banks. It also suffers from a relative weakness in terms of liquidity and poorly performing loans. The structure of the remainder of the paper is as follows: in the second section, we describe developments in the Polish banking system since 1989, to provide an appropriate context for the statistical analysis. The third section details the various performance measures used to assess ef- ficiency differences between the Polish and uk banking sectors, the data used and provides results using descriptive statistics and tests of signifi- cance between means. In the fourth section, relative performance using a stochastic cost frontier analysis is reported. Finally, in the fifth section, we draw together conclusions and consider some implications for future research. Developments in the Polish Banking System Since 1989 Prior to 1989 Poland's banks were state owned and competition was lim- ited. In 1989 the sector was primarily composed of co-operative banks. By 1993 there were still 1653 co-operative banks out of a total of 1740 banks in the country.2 With the collapse of communism and the introduction of Poland's economic reform program to create a market economy, the Polish banks underwent privatization, so that by 2000 most of the banks had been transferred to the private sector. By then the industry consisted of 754 banks, however around 680 were still small co-operative units. A total of 47 of the larger banks had come under foreign ownership, with banking organizations in eu Member States being the largest single set of foreign owners. Since the end of the 1980s, the Polish banking sector has experienced three main stages of development. Firstly, from 1989 to 1992 there was a dramatic increase in competitive pressures, but still lacking was the necessary institutional underpinning to develop a sound market-based banking system. In particular, a robust legal and regulatory framework was missing. Secondly, between 1992 and 1997 a restructuring of fi- nancial institutions occurred including recapitalization of the banks, privatization,3 and new legal reforms that led to a more orderly compet- itive environment.4 Thirdly, since 1998 strategic investors have become progressively more active, taking advantage of the benefits brought about by privatization and market liberalization. In other words, during the 1990s the banking sector became more commercially orientated, involv- ing significant restructuring in parallel with restructuring changes going on elsewhere in the Polish economy. In recent years, the pace of competition within the banking sector in Poland has intensified, in both the corporate finance and retail sec- tors. This has resulted largely from an influx of foreign-controlled banks. In fact, more than 75% of the capital in the Polish banking industry is now foreign-owned - German, Austrian and Dutch investors domi- nate (Balcerowicz and Bratkowski 2001). The consequence has been the development of new competitive strategies, the promotion of new hu- man resource skills5 and the expansion of systems to identify and cap- ture new markets (Balcerowicz and Bratkowski 2001; Figueira, Nellis, and Schoeneberg 2007).6 In the retail sector there has been extensive devel- opment of branch networks and the use of it in money transmission services.7 There has also been an improvement in the public perception of the banking industry in general, as the less popular and less efficient banks have either been closed or been merged with more efficient banks.8 How- ever, it appears that there are still some areas of the financial market which remain under-developed, especially the housing market. Very few Polish banks seem to specialize in providing mortgages9 and, those which do, impose a number of conditions which restricts the number of people eligible to apply for a mortgage.10 This compares unfavourably with the position in the eu Member States and especially the uk with its well- developed mortgage market supplied by banks and building societies. Previous studies have compared banks operating in Poland and West- ern Europe according to a range of efficiency ratios and concluded that, in 1997, Polish banks were less efficient. However, given the continuing changes in the Polish banking sector it seems timely to assess this per- formance again, using a wider range of performance measures including econometric analysis, particularly given Poland's recent entry into the eu. Performance Measurement, Data and Initial Findings In recent years, several studies have focused on performance in the bank- ing sector, however many of them have concentrated on a particular country and the analysis of scale and scope economies. For example, Berger (1993) analyzed us banks between 1980 and 1989 and concluded that management of resources is critical to achieving efficiency, while scale differences played a relatively minor role. Additional studies that have evaluated the performance of us banks include those by Peristiani (1996), Berger and Mester (1997), Mukherjee, Subhash, and Miller (2001), Barr et al. (2002) and Akhigbe and McNulty (2002). Other performance studies of banking include those by Gough (1979), Hardwick (1989; 1990), Drake (1992), Dietsch (1993) and Lang and Welzel (1996). Altunbas et al. (2001) extended the existing literature on modelling costs in banking systems by estimating scale economies, inefficiencies and technical change. In their study a sample of eu countries was used and efficiency was measured using stochastic cost frontier (scf) tech- niques (for an explanation of scf, see below). The results revealed that production inefficiencies were larger than scale inefficiencies, a finding consistent with the majority of us studies. The study also concluded that inefficiencies tend to vary across countries and over time. Since then, other studies have focused on cost and profit efficiency issues related to eu banking, such as Maudos et al. (2002) and Weill (2004). However, despite the recent entry of a number of Central and Eastern European countries into the eu, there appear to have been few studies of the performance of banks in these countries. The majority of stud- ies tend to be descriptive and a number are restricted to a comparison of accounting ratios, such as return on assets or return on equity (Weller 2000; Marek and Baun 2002; Keren and Ofer 2002). Although a few stud- ies have applied econometric modelling including scf analysis (Mertens and Urga 2001; Hasan and Marton 2003), the literature lacks a direct comparison between the banking systems in these countries and mem- bers of the eu pre-2004. As Berger and Humphrey (1997) conclude from a survey of studies of efficiency of financial institutions, international comparisons deserve additional attention. In this paper the performance of Poland's banking sector is com- pared with performance in uk banks. Bank performance can be mea- sured along a number of dimensions, including charges, financial ratios and costs of operation. Economists usually differentiate between alloca- tive efficiency and productive efficiency when assessing economic per- formance. Allocative efficiency is concerned with price-cost margins, and productive efficiency with costs of production. A distinction is also made between static efficiency gains, which are gains at a point in time or in the short-run, and dynamic efficiency gains, which are more concerned with longer-term economic performance improvements, usually associ- ated with innovation in products and processes. In this study, for reasons of data availability, the concern is with per- formance over the period 1999-2004, and with efficiency in the provision of outputs. Data do not exist to discuss price-cost margins and therefore allocative efficiency (although the existence of competition in uk bank- ing and the growing competition in Polish banking implies a high degree of allocative efficiency) or longer-term dynamic gains. The focus is there- fore on relative static efficiency using measures of productive efficiency.n The main measures used are profitability (since in a competitive mar- ketplace profits reflect cost control as well as revenue maximisation), other financial ratios and costs of production. The data are drawn from the Bankscope data base which contains balance sheet and income state- table 1 Data sample - uk and Polish banks, 2004 United Kingdom Poland Total assets (us$m) 10,703.266 168.099 Sample assets (us$m) 6,814.344 149.299 % assets included 64 89 % of commercial bank assets included 72 89 Total number of banks 140 23 Commercial banks 66 20 Savings banks 2 1 Co-operative banks 0 2 Real estate and mortgage banks 58 0 Investment banks and securities houses 14 0 ment data published by the London-based International Bank Credit Analysis Ltd. The sample used comprises 163 banks, 140 of which are uk banks and the remaining Polish. Prior to 1999 the data in Bankscope are incomplete, thus preventing analyses prior to that year. The banks examined in the Bankscope data base fall into the following categories: commercial, savings, co-operative, real estate and mortgage as well as investment banks and securities houses, with the majority being commercial banks. For the uk, around 41% of the banks are real estate and mortgage banks and 10% are classified as investment banks and se- curities houses. In contrast, the Bankscope data base has no Polish banks classified as investment banks and securities houses. This means that for Poland the classification 'commercial banks' includes banks that provide services which in the uk are mainly offered by specialist real estate and mortgage banks and investment banks. This introduces a potential lack of homogeneity in the classification of banks' activities across the two countries. However, banks in the Bankscope data base are categorized ac- cording to their primary activity or, more precisely, the activity to which more than 50% of operations relate. This means that heterogeneity in ac- tivities is limited and should not constitute a significant problem when comparing banks in Poland and the uk. The information in table 1 highlights other important differences in the two countries' banking systems. In particular, there are many more banks in the uk than in Poland, and each of the banks has much larger average assets - averaging over $6.814 billion in the uk as against more than $149 billion in Poland. Performance results may therefore be af- fected by firm size or scale of operation, something we test for later in the paper. It is also clear from the table that the commercial banks dom- inate both banking systems. For this reason, in the discussion below we concentrate upon the relative performance results for the commercial banks. As can be seen from table 1, more than 60% of the total assets of the banks in both countries are included in the study and over 70% of commercial banks' assets, which suggests that the sample used is suf- ficiently large to offer a fairly representative picture of performance in the uk and Polish banking sectors, especially with respect to commercial banking. Table 2 presents the results for a range of performance measures for banks in the two countries. The indicators are chosen to reflect key bank- ing metrics, namely asset quality ratios, capital ratios, operations ratios and liquidity ratios. Standard deviations are given in parentheses and in- dicate that for some of the measures, such as profitability, no major dif- ferences exist in data dispersion between Polish and uk banks, permit- ting a focus on the mean figures. For other indicators, such as impaired loans (defined as loans with suspended interest), there is a noticeable difference in the data dispersion, which means that both the means and standard deviations should be considered together. Two-tailed t-tests were undertaken to determine whether the difference between means for each of the performance measures was statistically significant at the 10% level. The results are provided in the final column of the table. Starting with the asset quality ratios, it is clear from the information presented in table 2 that in Poland the ratio of impaired loans to total loans is significantly higher than in the uk, confirming that Poland has a more serious problem with underperforming loans in its banks' bal- ance sheets (Polish banks also record higher average loan loss reserves to gross loans, and the difference between means is statistically signifi- cant at the 10% level). This result is almost certainly a legacy of the eco- nomic restructuring of the 1990s and the greater difficulty in assessing a borrower's credit worthiness in Poland than in the uk, with a less-well developed system of credit referencing in the former. In terms of capi- tal ratios, however, banks in Poland are not obviously under-capitalized, as suggested by the mean value shown in the table. Moreover, the dif- ference between the banks in the two countries is only just statistically significant at the 10% level for the ratio of equity to liabilities. Looking at the standard deviations, it is clear that loan loss reserves vary more table 2 Cost and profitability ratios of banks in the uk and Poland" (average values 1999-2004) Banks United Kingdomb Polandb (1) Asset quality ratiosc Loan loss reserves/gross loans 2.108 (3.987) 5.984 (3.177) Yes Impaired loans/gross loans 4.001 (7.313) 17.198 (12.029) Yes Capital ratios Equity/total assets 10.437 (9.772) 9.840 (3.116) No Equity/liabilities 14.551 (22.001) 11.169 (3-989) Yes Operations ratios Net interest margin 2.602 (2.745) 4.195 (2.518) Yes Average profit (profit/assets) 0.014 (0.028) 0.015 (0.008) No Return on assets employed 1.035 (2.253) 0.840 (1.016) No Return on equity 9.325 (11.042) 8.022 (10.565) No Average costs (costs/assets) 0.065 (0.071) 0.106 (0.016) Yes Average operational costs 0.032 (0.074) 0.051 (0.014) Yes Average financial costs 0.033 (0.010) 0.055 (0.008) Yes Cost to income ratio 68.968 (23.251) 70.676 (17.640) No Liquidity ratios Net loans/total assets 56.786 (27.903) 47.080 (13.807) Yes Liquid assets/total deposits & borrowingc 38.541 (39.976) 16.684 (10.818) Yes notes (1) Difference statistically significant (2-tailed test; 10% level). a Notethatthe results reported in this table are based on a 'balanced' panel data set - i. e. the same sets of banks are analysed in each year. b Standard deviations in parentheses. c The ratios were constructed with data from 140 uk banks and 23 Polish banks, with the exception of the following ratios where fewer banks were considered, due to data limitations: loan loss reserves/gross loans (130 uk and 18 Polish banks), impaired loans/gross loans (40 uk and 17 Polish banks) and liquid assets/total deposits and borrowing (61 uk and 20 Polish banks). across uk banks, although the reverse is true for impaired loans. On bal- ance, the standard deviations do not detract from the general conclusion that Poland has a greater problem with underperforming loans. With re- gard to the equity financing ratios, there is a wider dispersion around the mean figure for the uk. Turning to the operations ratios, profitability is conventionally mea- sured as a return on assets employed and as a return on equity. The profitability figures in table 2 suggest that for banks in Poland and the uk, profits on assets employed vary little between the two. Also, while on first inspection the descriptive statistics may suggest that returns on assets employed and return on equity are higher in the uk banks than in their Polish equivalents, the mean differences proved statistically in- significant (again at the 10% level). The conclusion is that the Polish and uk banking sectors have similar profitability. By contrast, costs of production in relation to assets employed are lower in the uk and this result is statistically significant, while the cost to income ratio is slightly higher in Polish banks (though this difference is not statistically significant) than in the uk counterparts. 12 This leads to the conclusion that banks in Poland have higher costs in relation to asset size than in the uk. These higher costs seem to be compensated for by higher revenues in relation to assets employed (note the higher net in- terest margin for Poland's banks), probably reflecting the lower level of competition in Polish banking. In turn, this suggests that as competition puts downward pressure on bank charges, the Polish banks will need to reduce their asset base, probably through further consolidation, if they are to remain competitive. In banking, costs of production can be divided between the costs of operating the bank, including branch networks, and the cost of raising loanable funds. It is therefore useful to explore performance differences separately in terms of operational costs and financial costs. Table 2 pro- vides figures on operational and financial costs in relation to assets em- ployed in banks in Poland and the uk. Both operational and financial costs in relation to assets employed are on average much higher in Poland - a mean figure of 0.051 and 0.055 respectively compared with 0.032 and 0.033, and these differences are statistically significant. This finding is consistent with the notion that Poland's financial market is less advanced and competitive than the uk's. This suggests that, in general, it costs Pol- ish banks more to raise loanable funds than is the case for uk banks with an equivalent asset base. However, with Poland's membership of the eu and the creation of single money and capital markets, this differential is likely to be eroded. This may be expected to improve the competitiveness of Polish banks in terms of raising finance. Finally, the liquidity ratio figures in table 2 suggest that Poland's banks are more exposed in terms of liquid assets with respect to total deposits and borrowing. This finding is of particular concern when set alongside the ratio for impaired loans. Together the results suggest that a number of Poland's banks are likely to be less able to absorb the impact of a financial crisis than banks in the uk. A Cost Frontier Analysis of Banking Performance So far, the relative performance of Polish and uk banks has been mea- sured using descriptive statistics. Here we assess performance using econometrics and specifically a stochastic cost frontier approach. Cost functions provide a more comprehensive analysis of performance than the simpler ratio analysis reported above. A cost function relates the costs of production observed in the data period - in this case 1999-2004 - to input and output variables, and derives directly from the theory of the firm (Varian 1992). We would have liked to have included an assess- ment of Polish and uk banks performance also based on a profit frontier analysis. However, like Bos (2002) and Bikker (2004), we found that while one single cost frontier exists when comparing across countries, this does not hold true for the profit frontier, probably due to different market conditions. Hence, the profit function approach does not allow for satisfactory comparisons across countries or regions. Cost efficiency is the ratio between the minimum cost (Cmin) neces- sary to achieve a desired level of output and the observed total cost (C). Total costs are therefore a function of the output (y), the price of inputs (w) and a set of other factors, which we here decompose into two parts: the level of cost inefficiency in production (u) and a random part (v). The latter accounts for measurement error and other random factors, such as the effects of strikes, etc., on the value of the output variables, to- gether with the effects of unspecified input variables in the cost function (see Coelli, Rao, and Battese 1998). Assuming that u and v are multiplica- tively separable from the other variables of the function and also that the variables are expressed in logarithms, then the cost function can be writ- ten as: lnC = f(y, w) + lnu + lnv. (1) Cost efficiency for an individual bank can then be described by the function: — = exp[/Q/,w)]-exp(lnv) = ^ C ex p [/(}', vi') I • exp(lnv) • exp(lnu) The model employed in this paper is a standard translog functional form (Casu and Girardone 2002; Figueira, Nellis, and Parker, forthcom- ing). Hence the cost equation to be estimated is: 3 13 3 InC = a + ^ßilnwi + - ^T • In(wf) 2 i=1 ;=1 2 2 + ^ rJn(yn) + JnnMyn) • In (jm) n=1 n=1 m=1 3 2 1 2 YjPinWwd ■ In(y„) + SEHE) + -ÖeeME)2 i=1 n=1 2 23 ^ ÄEnln(E) ■ ln(yn) + ^ TEiln(E) ■ ln(w*) + lnv + lnu, (3) 3 + 2 + ^T ÄEnln(E) ■ ln(yn) + = 1 i=1 where restrictions of symmetry and linear homogeneity have been im- posed on input prices. The variables included in the model are total costs (C), which include financial and operating costs, input prices de- scribed as price of loanable funds or the costs of raising funds to lend out (w1), the price of labour (w2) and the price of physical (fixed) capi- tal e. g. buildings (w3), and the quantity of outputs, which are deposits, including loans (y1) and other earning assets (y2) and financial capital (E), which is a proxy for banks' insolvency risk.13 The price of loanable funds is obtained by dividing financial cost by the corresponding liabil- ities, which include deposits, money market funding and other fund- ing. The price of labour would ideally be the marginal cost of employ- ing labour, but in the absence of these data an approximation was used based on the ratio between personnel expenses and total assets. The ra- tionale for this approximation is that it crudely represents the labour cost per worker adjusted for variations in labour productivity (Altunbas et al. 2001). 14 Finally, the price of physical capital is approximated by dividing expenditures on plant and equipment (non-labour costs) by fixed as- sets (Bikker and Haaf 2002; Maudos et al. 2002). One possible difficulty relating to the analysis is aggregation bias because of the mixing of dif- ferent sizes of banks in the two countries. We tested for this by including the logarithm of total assets. However, this proved to be insignificant in the explanation of total costs. Therefore, the mixing of different sizes of banks in uk and Poland does not seem to affect the results. In common with some of the earlier studies of bank performance re- viewed above, we estimate an efficient frontier for the banking industry. A bank's performance is then assessed by measuring how efficient it is, based on its distance from the efficient frontier, a concept that dates back to Farrell (1957). Such values are sometimes referred to as measures of x-inefficiency (Berger 1993). Here the frontier is estimated by amalga- mating data from the Polish and uk banking sectors and again draw- ing from the Bankscope data base. In this stage of the analysis all banks in Poland and the uk were included in the data set so as to maximise the degrees of freedom and provide a more robust estimate of the cost frontier. To model the frontier we used stochastic cost frontier analysis (scf), as proposed by Aigner, Lovell, and Schmidt (1977), and equation 3 above. 15 scf breaks down the error term into the two distinct parts al- ready referred to, namely vi or the random error, which is assumed to be independently and identically distributed following a normal distribu- tion, and ui. This is a non-negative inefficiency term and assumed to be independently and identically distributed and to follow a truncated nor- mal or exponential distribution. The estimated inefficiency for any firm is taken as the conditional mean of the distribution of the inefficiency term, given the observation of the composed error term. The model proposed by Battese and Coelli (1995) is used in this paper and is close to that proposed by Aigner, Lovell, and Schmidt (1977). It dif- fers in imposing allocative efficiency and allows the use of panel data.16 The estimation of the model occurs in three main steps. The first involves the estimation of the function by Ordinary Least Squares (oLs). The pa- rameters obtained are all unbiased with the exception of ß0 (intercept) and < (sum of the variance of ui and v*). The second step is carried out with the estimation of a likelihood function based on Battese and Corra (1977),17 which is evaluated for a series of values of 7 between zero and one - where 7 equal to zero means that the deviations from the frontier are due only to noise, while a value of one indicates that the deviations are due entirely to inefficiency. The estimates for < and ßo are adjusted, with the remaining coefficients unchanged. The final step uses the best estimates from the second step as starting values in an iterative procedure to achieve the final Maximum Likelihood estimates. An individual bank's cost efficiency is then predicted from the esti- mates of the stochastic cost frontier. Battese and Coelli (1988) point out that the best predictor of exp(-ui) is given by: 1 _ E[ex p(-«0|e;] = ———exp(ye, + —), (4) X — (P—- ^ (zt)] i=1 i=1 where z; = (lnyi -xß)/cr5 1 - 7) and <£(•) is the distribution func- tion of the standard normal random variable. 18 Details of the calculations at each stage of the scf analysis are available from the authors on request. References Aigner, D. J., C. A. K. Lovell, and P. Schmidt. 1977. Formulation and esti- mation of stochastic frontier production function models. Journal of Econometrics 6:21-37. Akhigbe, A., and J. McNulty. 2002. The profit efficiency of small us com- mercial banks. Journal of Banking and Finance 27 (2): 307-26. Altunbas, Y., E. P. M. Gardener, P. Molyneux, and B. Moore. 2001. Effi- ciency in European Banking. European Economic Review 45:1931-55. Balcerowicz, E., and A. Bratkowski. 2001. Restructuring and development of the banking sector in Poland: Lessons to be learnt by less advanced transition countries. Case report 44, Centre for Social and Economic Research, Warsaw. Barr, R., K. A. Killgo, T. F. Siems, and S. Zimmel. 2002. Evaluating the pro- ductive efficiency performance of us commercial banks. Managerial Finance 28 (8): 3-25. Battese, G. E., and T. J. Coelli. 1988. Prediction of firm-level technical ef- ficiencies with a generalised frontier production function and panel data. Journal of Econometrics 38 (3): 387-99. -. 1995. A model for technical inefficiency effects in a stochastic frontier production function for panel data. Empirical Economics 20 (2): 325-32. Battese, G. E., and G. S. Corra. 1977. Estimation of a production frontier model: With application to the pastoral zone of Eastern Australia. Aus- tralian Journal of Agricultural Economics 21 (3): 169-179. Berger, A. N. 1993. Distribution-free estimates of efficiency in the us bank- ing industry and tests of the standard distributional assumptions. Journal of Productivity Analysis 4 (3): 261-92. Berger, A. N., and D. B. Humphrey. 1997. Efficiency of financial institu- tions: International survey and directions for future research. Euro- pean Journal of Operational Research 98 (2): 172-212. Berger, A. N., and L. J. Mester. 1997. Inside the black box: What explains differences in the efficiencies of financial institutions? Journal of Bank- ing and Finance 21 (4): 898-947. Bikker, J. A.. 2004. Competition and efficiency in a unified European banking market. Cheltenham: Elgar. Bikker, J. A., and K. Haaf. 2002. Competition, concentration and their re- lationship: An empirical analysis of the banking industry. Journal of Banking and Finance 26 (11): 2191-14. Bos, J. W. B. 2002. European banking: Market power and efficiency. phü diss., Maastricht University. Casu, B., and C. Girardone. 2002. A comparative study of the cost effi- ciency of Italian bank conglomerates. Managerial Finance 28 (9): 3-23. Chiquier, L. 1999. Regulation of mortgage banks and mortgage bonds in Poland. A report prepared for the East European Regional Housing Sector Assistance Project. Project 180-0034, us Agency for Interna- tional Development. Coelli, T., D. S. P. Rao, and G. E. Battese. 1998. An introduction to efficiency and productivity analysis. London: Kluwer. Dietsch, M. 1993. Economies of scale and scope in the French commercial bank industries. Journal of Productivity Analysis 4 (1-2): 35-50. Drake, L. 1992. Economies of scale and scope in uk building societies: An application of the translog multiproduct cost function. Applied Finan- cial Economics 2 (4): 211-19. Farrell, M. J. 1957. The measurement of productive efficiency. Journal of the Royal Statistical Society 120 (3): 253-90. Figueira, C., J. Nellis, and D. Parker. Forthcoming. Does ownership affect the efficiency of African banks? Journal of Developing Areas. Figueira, C., J. Nellis, and R. Schoenberg. 2007. Travel abroad or stay at home? Investigating the patterns of bank industry m&as in the eu. European Business Review 19 (1): 23-39. Golajewska, M., and M. Jozefowska. 2001. Banki komercyjne. In Banko- wosc na swiecie i w Polsce, ed. L. Oreiziak and B. Pietrzak. Warszawa: Olympus. Gough, T. J. 1979. Building society mergers and size efficiency relationship. Applied Economics 11 (2): 185-94. Hardwick, P. 1989. Economies of scale in building societies. Applied Eco- nomics 21:1291-304. -. 1990. Multi-product cost attributes: A study of uk building soci- eties. Oxford Economic Papers 42 (2): 446-61. Hasan, I., and K. Marton. 2003. Development and efficiency of the bank- ing sector in a transitional economy: Hungarian experience. Journal of Banking and Finance 27 (12): 2249-71. Kempny, M. 2002. Three years of mortgage banking in Poland. Paper pre- sented at the oecd workshop in Warsaw, Poland, 5-6 December. Keren, M., and G. Ofer. 2002. Globalization and the role of foreign banks in economies in transition. tiger Working Paper 32. Lang, G., and P. Welzel. 1996. Efficiency and technical progress in banking: Empirical results for a panel of German cooperative banks. Journal of Banking and Finance 20 (6): 1003-23. Marek, D., and M. Baun. 2002. The eu as a regional actor: The case of the Czech Republic. Journal of Common Market Studies 40 (5): 895-919. Maudos, J., J. H. Pastor, F. Perez, and J. Quesada. 2002. Cost and profit ef- ficiency in European banks. Journal of International Financial Markets, Institutions and Money 12 (1): 33-58. Mertens, A., and G. Urga. 2001. Efficiency, scale and scope economies in the Ukrainian banking sector in 1998. Emerging Markets Review 2 (3): 292-308. Mukherjee, K., C. R. Subhash, and S. M. Miller. 2001. Productivity growth in large us commercial banks: The initial post-deregulation experi- ence. Journal of Banking and Finance 25 (5): 913-39. Opiela, T. 2001. Assessing the efficiency of Polish commercial banks. Working Paper 18, National Bank of Poland. Peristiani, S. 1996. Do mergers improve the x-efficiency and scale efficiency of us banks? Evidence from the 1980s. Research Paper 9623, Federal Reserve Bank of New York. usaid. 2000. An assessment and rating of the Polish banking system. Re- port, usaid mission to Poland. Varian, H. R. 1992. Microeconomic analysis. 3rd ed. New York: Norton. Weill, L. 2004. Measuring cost efficiency in European banking: A com- parison of frontier techniques. Journal of Productivity Analysis 21 (2): 133-52. Weller, C. 2000. Multinational banks in developing transition economies. Technical Paper 241, Economic Policy Institute.