Measurement of Banking Efficiency using Dynamic Data Envelopment Analysis Model: Evidence of ten Central and Eastern Europe Countries

Transcription

1 Measurement of Banking Efficiency using Dynamic Data Envelopment Analysis Model: Evidence of ten Central and Eastern Europe Countries Iveta Palečková and Daniel Stavárek Silesian University, School of Business Administration, Univerzitní náměstí 1934/3, Karviná, Czechia Abstract The aim of the paper is to estimate the banking efficiency of ten Central and Eastern Europe Counties using the Dynamic Data Envelopment Analysis. In accordance with the aim of the paper we determine whether banks that belong to financial conglomerates are more or less efficient than other banks in the banking industry. The observed Central and Eastern Europe countries include Bosnia and Herzegovina, Bulgaria, Croatia, Czechia, Hungary, Poland, Romania, Serbia, Slovakia and Slovenia. First, we estimate the relative efficiency using the Dynamic Data Envelopment Analysis, the slack-based measure model with variable return to scale. Next, we analyse the individual banks of financial conglomerates. We investigate banks from four financial conglomerates (Erste Group, KBC Group, Société Générale Group and UniCredit Group). We analyse whether these banks achieve a value above or below the median value in each observed country. Results suggest that most banks were higher efficient than median value of technical efficiency in the banking industry, but there were the differences in banks in financial conglomerates across CEE countries. Keywords: Banking sector; CEE countries; Dynamic DEA; financial conglomerate; intermediation approach Introduction The aim of the paper is to estimate the banking efficiency of ten Central and Eastern Europe Counties using the Dynamic Data Envelopment Analysis. In line with the aim of the paper we determine whether banks that belong to a financial conglomerate are more or less efficient than other commercial banks in the banking sector. First of all the term efficiency is defined. Daraio and Simar (2007) describe efficiency as a distance between the quantity of input and output, and the quantity of input and output that defines a frontier, the best possible frontier for a firm in its industry. Farrell (1957) proposed that the efficiency of any firm consists of two components: technical efficiency and allocative efficiency. Technical efficiency is the ability of the firm to maximize outputs from a given set of inputs. Allocative efficiency is the ability of the firm to use these inputs in optimal proportion given their respective prices. In this paper we estimate the technical efficiency of commercial banks. The observed Central and Eastern Europe (CEE) countries included the banking sectors of transition countries: Bosnia and Herzegovina, Bulgaria, Croatia, Czechia, Hungary, Poland, Romania, Serbia, Slovakia and Slovenia. Bank restructuring and privatization were largely completed in these countries. CEE countries moved from a centrally-planned to a market economy

2 in the beginning of the 1990s. Financial sectors of CEE countries can be characterized as bankbased system and banks play an important role in economy of these countries. Most of analysed CEE countries joined the European Union (except Serbia and Bosnia and Herzegovina that are candidate or potential candidate). Slovenia and Slovakia also entered the Euro area. The CEE countries were strongly affected by the global financial crisis. One of the reason for affecting the CEE countries of the financial crisis was the fact that the CEE banking sector is more than half owned by Western European banks. Therefore it is an area of concern and a potential channel of transmission of economic turmoil. Next, we analyse efficiency of banks from four financial conglomerates, i.e. Erste Group, KBC Group, Société Générale Group and UniCredit Group. We estimated the technical efficiency of the banking sector during the period using the Dynamic Data Envelopment Analysis (DDEA) model. We used the non-oriented slack-based measure (SBM) model with variable return to scale. Banking efficiency is very often investigated using the static methods. This traditional Data Envelopment Analysis (DEA) models treat the efficiency of resources (inputs and outputs) related to making decisions in decision-making units (DMUs) with cross-sectional data. The analysis is performed in only one time period, hampering the measurement of productivity changes when there is more than one time period (Costa et al., 2014). The Dynamic DEA estimates the efficiency of DMU during several periods of time. The Dynamic Data Envelopment Analysis model considers the effect of interconnecting activities (links) between two consecutive terms. We also analyse the individual banks of financial conglomerates. We investigate banks from four financial conglomerates and we analyse whether these banks achieve a value above or below the median value in each observed country. The structure of the paper is follow. Next section presents brief information about previous studies regarding the technical efficiency using non-parametric approach in CEE countries. Third section presents the methodology and next section describes the data set and selection of variables in the model. Then the section empirical analysis and results where are the results of the models registered and last section concludes the paper. Literature Review Several studies estimated the efficiency in selected European countries using different approach. We mention several of them that use the non-parametric approach, namely Data Envelopment Analysis. For instance, Stavárek and Polouček (2004) estimated efficiency and profitability in selected CEE banking sectors using DEA model. Authors found that that the Czech and Hungarian banking sectors were on average evaluated as the most efficient and the Czech banking sector showed itself as the most aligned banking industry among transition countries. Next, Andries and Cocris (2010) analysed the efficiency of the main banks in Romania, Czech Republic and Hungary using parametric and non-parametric approach. Authors analysis showed that the Romanian banks reached low level of technical and cost efficiency. Several researcher evaluated banking efficiency in Visegrad countries and they concluded that the Czech banking sector was highly efficient (e.g. Staníčková and Melecký, 2012). Also Stavárek (2005) estimated commercial banks efficiency in the group of Visegrad countries (Czech Republic, Hungary, Poland, Slovakia) before joining the EU. Author concluded that the Czech banking sector was the most efficient, followed by the Hungarian with a marginal gap.

3 Several studies estimated the efficiency in individual countries of South and Eastern Europe using DEA model. For instance, we can mention Efendić (2011) who estimated banking efficiency in Bosnia and Herzegovina and author found that according to all efficiency indicators, there were a significant potential for efficiency improvements. Similarly, Memić and Škaljić-Memić (2013) estimated performance of banks in Bosnia and Herzegovina using DEA approach. Researchers found that efficiency of individual banks varied throughout the period and not all of the banks were a part of the negative banking sector trend induced by the crisis. Mihajlović et al. (2009) applied DEA approach on banks in Serbia and found that just 9 out of 41 banks in Serbia were efficient in Another 7 inefficient banks have efficiency index very close to 1. Bulajić et al. (2011) used DEA model to estimate efficiency in Serbian banking sector. Authors analysed 30 banks in Serbia operating in the economically turbulent five-year period Authors ranked banks according to their average efficiency and similarities and differences among them were investigated. Also Maletić et al. (2013) estimated the efficiency of banks in Serbia using two DEA models with different input-output indicators. They ranked the banks according to their efficiency and found that the most efficient and superefficient were banks of the public sector. Tochkov and Nenovsky (2011) investigated the efficiency of Bulgarian banks using the non-parametric method within the period. Researchers found that the development of banking efficiency fluctuated and it could be divided into several phases: first the efficiency decreased during , next the average efficiency was increasing within The results showed that large banks were more efficient than small banks. The empirical application of dynamic efficiency of the CEE banking sectors are rare. The Dynamic DEA was applied by e.g. Řepková (2013) who estimated the efficiency of the Czech banking sector during the period The results of the model showed that efficiency slightly increased in the analysed period. Other application of the Dynamic DEA model was in study of Palečková (2015) who estimated the efficiency of the banking sectors in Visegrad countries during the period Researcher found that the average efficiency slightly decreased during the period and significantly decreased in 2012 which was probably as a result of financial crisis. The Czech and Hungarian banking sector was the highest efficient. The similar results were estimated also in the banking sectors in Bosnia and Herzegovina, Croatia, Serbia and Slovenia in study of Palečková (2016). Only a few studies was focused on the efficiency of financial conglomerates. Vander Vennet (2002) examined the cost and profit efficiency of European conglomerates and universal banks and found that conglomerates were more efficient than their specialized competitors. However this study investigated the cost efficiency. Although Casu and Girardone (2004) were focused on different banking sector, authors estimated the efficiency of Italian financial conglomerates in the 1990s. The effect of financial conglomerate estimated examined, for instance, Palečková (2017) who assessed the efficiency and efficiency change using Malmquist index in the group of Visegrad countries during the period. Researcher determined that there were differences in banks in the financial conglomerates across Visegrad countries. Thus it gives us the opportunity to examined the efficiency of CEE banking sectors using the Dynamic DEA and also answer the research question: Are banks that belong to a financial conglomerate more or less efficient than other commercial banks in the banking sector?.

4 Methodology The Data Envelopment Analysis is an approach for evaluating the performance of a set of peer entities (DMUs) which convert multiple inputs into multiple outputs (Cooper et al., 2011). The term DEA was first described by Charnes et al. (1978) and it was model with assumption of constant return to scale. Next, this model was modified by Banker et al. (1984) and became the BCC model which accommodates variable returns to scale. Moreover, the criterion for the classification of DEA models is possible orientations in DEA models. Input-oriented model represent models where the DMUs produce a given quantity of outputs with the minimum possible amount of controllably inputs (Henčlová et al., 2015). Outputoriented model is the model that attempts to maximize outputs while using no more than the observed amount of any inputs (Cooper et al., 2007). Non-oriented (additive) models are based on the optimal mix of inputs and outputs, it is combination both orientations in a single model (Cooper et al., 2007). We employed the non-oriented Slacks-Based Model (SBM) introduced by Tone (2001) who proposed a slacks-based measure (SBM) of efficiency in the Data Envelopment Analysis. The SBM model maximizes the average improvements of relevant factors (inputs / outputs) for the evaluate DMU to reach the frontier (Tone, 2001). Tone and Tsutsui (2010) developed Dynamic DEA model in to a slacks-based measure framework for measuring the dynamic efficiency of relative DMUs over several terms. Authors pointed out a concept of carry-over and accounted the effect of interconnecting activities between two consecutive terms. We adopted the Dynamic DEA model proposed by Tone and Tsutsui (2010). The Dynamic DEA model can easily be written as: subject to max z(t 1) = w (t)λ (t), (1) A (t)λ (t) X (t), (2) λ (t) 0, all t = 0,1,2,, T 1, (3) where z is efficiency of DMU to be estimated, λ (t) is the output vector for each DMU, X is current input, A (t) is the corresponding input coefficient matrices, and w (t) is a non-negative weight vector for the multiple outputs of each DMUj, j indicates the n different DMUs and t denotes time. We estimated the dynamic model in the slacks-based measure (SBM) framework, called Dynamic SBM (DSBM). Data and Selection of Variables The data set used in this study was obtained from the database Orbis Bank Focus and the annual reports of commercial banks during the period We analyse commercial banks of banking sectors in Bosnia and Herzegovina, Bulgaria, Croatia, Czechia, Hungary, Poland, Romania, Serbia, Slovakia and Slovenia. All the data is reported on unconsolidated basis. We

5 analysed only commercial banks because we need homogenous data set. The calculation of the Dynamic DEA requires strictly balanced panel data, thus the data set consists of 173 commercial banks from CEE countries. The observed banks represents, in average, more than 81% of banking sectors assets, thus the sample of banks is representative and results of this paper could be interpreted as results of banking sectors. In order to conduct the Dynamic DEA estimation, inputs and outputs need to be defined. Several major approaches (intermediation approach, production approach, asset approach, profitable approach) were developed in the empirical literature that define the relationship of inputs and outputs in the behavior of financial institutions. Berger and Humphrey (1997) showed that the intermediation approach may be more appropriate for evaluating entire financial institutions because this approach is inclusive of interest expenses, which often account for one-half to twothirds of total costs. Thus, in this paper we adopted intermediation approach for the definition of inputs and outputs. Consistently with this approach, we assume that banks bank collects deposits to transform them in loans. We employed three inputs (labor, deposits and physical capital) and two outputs (loans and other earning assets). We measure labor (L) by the total number of employees and deposits (TD) by the sum of demand and time deposits from customers, interbank deposits and sources obtained by bonds issued. Physical capital (C) is proxied by fixed assets and lands. Loans (TL) are measured by the net value of loans to customers and other financial institutions and other earning assets (OEA) are collected form balance sheet of banks. Moreover, the Dynamic DEA model takes into account the internal heterogeneous organizations of DMUs for which divisions are mutually connected by link variables and trade internal products with each other. Thus, Each DMU has carry-over variables that take into account a positive or negative factor in the previous period. This model has the huge advantage of being able to evaluate the policy effect on the individual divisions of each DMU. As the carry over variable was used total loans in this paper. Descriptive statistics of inputs and outputs are presented is Table 1. Table 1: Descriptive Statistics of Variables (ml EUR) Mean Median Maximum Minimum St.Dev. Total deposits Physical capital Number or employees Total loans Other earning assets Results and Discussions In this paper, we used Dynamic DEA non-oriented SBM model with assumption of variable return to scale. This method is suitable in the banking sector because it can easily handle multiple inputsoutputs producers such as banks. We divided the sample of 10 CEE banks into two parts according to the nominal GDP per capita and the level of financial intermediation. First parts included commercial banks from Czechia, Slovakia, Hungary, Poland and Slovenia and the second part covered commercial banks from Bosnia and Herzegovina, Croatia, Serbia, Bulgaria and Romania.

6 Table 2: Descriptive statistic of technical efficiency in CEE banking sectors (in %) Mean Median Max Min St.Dev. Bosnia and Herzegovina BA Bulgaria BG Czechia CZ Croatia HR Hungary HU Poland PL Romania RO Serbia RS Slovakia SK Slovenia SL Table 2 presents the descriptive statistic of technical efficiency in CEE banking sectors using Dynamic DEA model. The highest average efficiency achieved Croatian commercial banks and the Czech commercial banks. On the other hand, the lowest average efficiency reached commercial banks in Serbia and Bosnia and Herzegovina. Low value of average efficiency especially in Bosnia and Herzegovina is due to the high impact of financial crisis on these commercial banks. Next, we present the development of technical efficiency in observed CEE countries during the period Table 3: Average value of technical efficiency of the commercial banks in CEE countries (in %) Country BA BG CZ HR HU PL RO RS SK SL Table 3 presents the development of efficiency of commercial banks in CEE countries. The development of average efficiency was different in individual countries. Several countries registered the increase in average efficiency, namely Polish, Romanian, Serbian or Slovakian

7 banking industries. On the other hand, average efficiency of banking sector in Bosnia and Herzegovina, Bulgaria, Hungary and Croatia decreased during the period. The development was fluctuating during these eleven years in most of countries. During the financial crisis the efficiency of most banking sectors decreased. It was caused mainly the decreased in provided loans and the excess of total deposits in the balance sheet of most commercial banks. This results are in line with previous studies, e.g. Kevork et al. (2017) or Degl'Innocenti et al. (2017) registered decrease in banking efficiency during the financial crisis and the sovereign debt crisis. In line with the aim of the paper we analyse the individual banks of financial conglomerates. We analyse banks from four financial conglomerates (Erste Group, KBC Group, Société Générale Group and UniCredit Group). We examine the efficiency of banks that belong to a financial conglomerate. We investigate whether these banks achieved a value above or below the median value in each CEE country. We use the technical efficiency score calculated for 10 CEE countries and we calculate mean and median value of technical efficiency for each countries. We compare the value of the banks that belongs to the financial conglomerate with the median value of the banking sector in relevant country. Table 4: The efficiency in banks of Erste Group (in %) Banks of Erste Group Mean value of the bank Median of the banking sector CZ Ceska sporitelna HR Erste & Steiermärkische Bank HU Erste Bank Hungary ZRT RO Banca Comerciala Romana RS Erste Bank a.d. Novi Sad SK Slovenska sporitelna All banks form Central and Eastern Europe countries that belongs to Erste Group achieved higher average values of efficiency than is median value of efficiency in the country (Table 4). The average value of Erste group 75% and Erste Group, in average, the commercial bank from CEE countries that belongs to Erste Group could be market as highest efficient. Table 5: The efficiency in banks of KBC Group (in %) Banks of KBC Group Mean value of the bank Median of the banking sector BG Cibank CZ CSOB HU K&H Bank Zrt SK CSOB SL NLB dd-nova Ljubljanska Banka 77 59

8 The situation in commercial banks with affiliation to KBC Group is similar. Only K&H Bank from Hungary was lower efficient that median efficiency in Hungary (Table 5). The average efficiency of banks belong to KBC Group achieved 70%. Table 6: The efficiency in banks of Société Générale Group (in %) Banks of Société Générale Group Mean value of the bank Median of the banking sector BG Societe Generale Expressbank CZ Komercni banka HR Societe Generale - Splitska Banka PL Euro Bank RO BRD-Groupe Societe Generale RS Societe Generale Banka Srbija SL SKB Banka Commercial banks from Société Générale Group reached average value of 46%. Banks that belong to this financial conglomerate were lowest efficiency (Table 6). And four banks from this financial conglomerate were lower efficient than median efficiency of the country. Only Komercni banka in Czechia, Splitska Banka from Croatia and BRD from Romania were higher efficient than median value of the country. Table 7: The efficiency in banks of UniCredit Group (in %) Banks of UniCredit Group Mean value of the bank Median of the banking sector BA UniCredit Bank BG UniCredit Bulbank CZ UniCredit Bank HR Zagrebacka Banka HU UniCredit Bank Hungary RO UniCredit Bank RS UniCredit Bank Serbia JSC SL UniCredit Banka Slovenija Also all commercial banks that belong to UniCredit Group were highest (or at the same level) efficient than median value in the country (Table 7). The average efficiency of these banks was 71% and they are, after Erste Group, the second highest efficient commercial banks. We can conclude that all banks that belongs to Erste Group and UniCredit Group were more efficient than median value of the country. Nevertheless we cannot unambiguously conclude that the banks that belongs to the financial group are more efficient that other banks in the banking sector. The affiliation to the financial conglomerate should be one of the determinants of technical

9 efficiency in CEE countries for several commercial banks. Similarly as the study of Palečková (2017) was found that in CEE countries were differences in banks in the financial conglomerates. Conclusions The aim of the paper is to estimate the technical efficiency of 10 CEE banking sectors using the Dynamic Data Envelopment Analysis. Applying the Dynamic DEA, SBM model with variable return to scale, the results registered the decrease in the technical efficiency during the financial crisis in most of CEE countries. The Croatian and Czech banking sectors were the most efficient banking sector. Conversely the banking sectors in Serbia and Bosnia and Herzegovina were the lowest efficient. The finding also indicated the decrease in average efficiency during and after the financial crisis. In accordance with the aim of the paper we determine whether banks that belong to a financial conglomerate are more or less efficient than other commercial banks in the banking sector. We found that there were the differences in efficiency in banks in financial conglomerates across CEE countries. This study confirms the results of Vander Vennet (2002) who found that conglomerates were more efficient than their specialized competitors. We found that most banks were higher efficient than median value of technical efficiency in the banking industry. But we cannot state that all banks in the financial conglomerate are more efficient than other commercial banks. We found that there were differences in efficiency in banks that belongs to the financial conglomerate and other banks in banking methods. Thus, in further research we would like to estimates these differences using matching methods, especially propensity score matching method. In further research we also would like to focus on the determinants of the technical efficiency of the commercial banks. We would like to estimate if the inefficiency is connected with the affiliation of the financial group or other bank-specific and macroeconomic factors. We will analyse the effect of affiliation of the financial conglomerate using panel data analysis. Acknowledgement Research behind this paper was supported by the Czech Science Foundation within the project GAČR S Affiliation with financial conglomerate as a determinant of performance and risk of banks. References Andries, A.M., Cocris, V. (2010). A Comparative Analysis of the Efficiency of Romanian Banks. Romanian Journal of Economic Forecasting, 4: Banker, R.D., Charnes, A., Cooper, W.W. (1984). Some Models for Estimating Technical and Scale Inefficiencies in Data Envelopment Analysis. Management Science 30(9): Berger, A.N., Humphrey, D. (1997). Efficiency of Financial Institutions: International Survey and Directions for Future Research. European Journal of Operational Research 98: Bulajić, M., Savić, G., Savić, S., Mihailović, N., Martić, M. (2011). Efficiency assessment of banks in Serbia. Technics Technologies Education Management 6(3):

10 Casu, B., Girardone, C. (2004). Financial Conglomeration: Efficiency, Productivity and Strategic Drive. Applied Financial Economics 14(10): Cooper, W., Seiford, L.M., Tone, K. (2007). Data Envelopment Analysis: A Comprehensive Text with Models Applications. New York: Springer Science. Cooper, W. W., Seiford, L. M., Zhu, J. (2011). Data Envelopment Analysis: History, Models, and Interpretations. Handbook on Data Envelopment Analysis. New York: Springer US. Costa, E.M., Ramos, F.S., Souza, H.R. (2014). Static versus Dynamic DEA in Federal Higher Education Institutions IFES, Análise Econômica 32(61): Daraio, C., Simar, L. (2007). Advanced Robust and Nonparametric Methods in Efficiency. New York: Springer. Degl'Innocenti, M. Kourtzidis, S. Sevic, Z., Tzeremes, N. (2017) Bank productivity growth and convergence in the European Union during the financial crisis. Journal of Banking and Finance 75: Efendić, V. (2011). Efficiency of the Banking Sector of Bosnia Herzegovina with Special Reference to Relative Efficiency of the Existing Islamic Bank. In 8th International Conference on Islamic Economics and Finance. Qatar: Center for Islamic Economics and Finance. Farrell, M. J. (1957). The Measurement of Productive Efficiency. Journal of the Royal Statistical Society. Series A (General) 120(3): Hančlová, J., Chytilová, L., Ramík, J., Toloo, M., Zapletal, F., Babaee, S., Hermans, E. (2015). Optimization problems in economics and finance (Vol. 41). Ostrava: VŠB-Technical University of Ostrava. Charnes, A., Cooper, W. W., Rhodes, E. (1978). Measuring the Efficiency of Decision Making Units. European Journal of Operational Research 2(6): Kevork, I.S., Pange, J., Tzeremes, P., Tzeremes, N. (2017) Estimating Malmquist productivity indexes using probabilistic directional distances: An application to the European banking sector, European Journal of Operational Research 261(3): Maletić, R., Kreća, M., Maletić, P. (2013). Application of DEA Methodology in Measuring Efficiency in the Banking Sector. Economics of Agriculture 60(4): Memić, D., Škaljić-Memić, S. (2013). Performance Analysis and Benchmarking of Commercial Banks Operating in Bosnia and Herzegovina: a DEA Approach. Business Systems Research 4(2): Mihailović, N., Bulajić, M., Savić, G. (2009). Ranking Of Banks in Serbia. Yugoslav Journal of Operations Research, 19(2): Palečková, I. (2015) Banking Efficiency in Visegrad Countries: A Dynamic Data Envelopment Analysis, Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis 63(6): Palečková, I. (2016). Estimation of Relationship between the Dynamic Efficiency and Profitability in the Banking Sector. In: Suresh, P. (ed.) Proceedings of the IRES International Conference Budapest, Hungary. Bhubaneswar: Institute of Research and Journal: 1 5.

11 Palečková, I. (2017). Efficiency Change of Banking Sectors and Banks in the Financial Conglomerates in Visegrad Group Countries. Ekonomický časopis 65(1): Řepková, I. (2013) Estimation of Banking Efficiency in the Czech Republic: Dynamic Data Envelopment Analysis, DANUBE: Law and Economics Review 4(4): Staníčková, M., Melecký, L. (2012). Assessment of Efficiency in Visegrad Countries and Regions Using DEA Models. Ekonomická revue Central European Review of Economics Issues 15(3): , Stavárek, D. (2005). Restrukturalizace bankovních sektorů a efektivnost bank v zemích Visegrádské skupiny. Karviná: Silesian University, School of Business Administration. Stavárek, D., Polouček, S. (2004). Efficiency and Profitability in the Banking Sector. In: POLOUČEK, S. (ed.): Reforming the Financial Sector in Central European Countries. Hampshire: Palgrave Macmillan Publishers. Tochkov, K., Nenovsky, N. (2011). Institutional Reforms, EU Accession, and Bank Efficiency in Transition Economies: Evidence from Bulgaria. Emerging Markets Finance and Trade 47(1): Tone, K. (2001) A slacks-based measure of efficiency in Data Envelopment Analysis, European Journal of Operational Research 130: Tone, K., Tsutsui, M. (2010) Dynamic DEA: A slacks-based measure approach, Omega 38: Vander Vennet, R. (2002) Cost and Profit Efficiency of Financial Conglomerates and Universal Banks in Europe. Journal of Money, Credit, and Banking 34(1):