Market structure and concentration in Islamic and conventional banking Ali CHKIR* & Amira HAKIM** Abstract This paper investigates the market structure and the degree of concentration of Islamic banking industry. This paper reports an empirical assessment of competitive conditions among the Arab GCC banking industry during the years of 2005-2010. We opt for a methodology as proposed by Panzar and Rosse based on a nonstructural estimation of the H-statistic and the most frequently applied measures of concentration k-bank concentration ratio (CRk), Herfindahl–Hirschman Index (HHI), and the measure of entropy. The results show that conventional banking industry is more concentrated than the Islamic banking industry. However the measure of entropy confirms that the market structure of Islamic and conventional banking is almost similar with reference to total deposits and totals loans. The empirical evidence reveals that the banking industry in general and the conventional banking industry in the region operate under perfect competition; the prevailing market structure in Islamic banking industry is mostly monopoly. Keywords: Islamic banking – concentration- market structure- GCC countries- Panzar-Rosse Approach 1 Introduction In the context of the financial globalization, Islamic finance seems to be a strategic issue to guard against economic crises and practice of our ethical values. Disclosure of Islamic banks in the banking sector has changed the structure of the banking market. This is a question mark for the economic policy makers. In the same context of financial globalization and the emergence of new technologies of information and communication, the banking industry is sustained by a process of banking and financial consolidation observed in the worldwide sphere. These worked embrittlement of the barriers to entry of new competitors and an intensification of competition. The study of market structure in the banking industry is one of the controversial subjects which surmount a review of extensive literature treating the subject in several dimensions namely a bank, a country and a geographic agglomeration. The study suggests that the competitive situation emphasize economic concentration for its interdependence with the competitive power of the sector. The study of market structure and the competitive power of the banking industry is an interest in economic policy for its impact on the economy. Several recent studies focus on the competitive structure of the banking industry. However, questions remain on the concentration and conditions of competition that Islamic banks are thriving in an economic environment characterized by the hegemony of conventional banks. Several recent studies focus on the competitive structure of the banking industry. However, questions remain on the concentration and conditions of competition that Islamic banks are thriving in an economic environment characterized by the hegemony of conventional banks. This study provides an overview of the tools at hand to investigate competitive conditions and to calculate the degree of concentration in banking market. To this purpose, we apply a methodology first proposed by Rosse and Panzar to a panel of banks. Section 2 illustrates the Rosse–Panzar test, with particular reference to the theoretical framework, the empirical implications and the existing studies employing this methodology. Section 3 describes our sample used to conduct this study. 2 Section 4 presents an overview of market concentration measures, and compares measures of concentration between Islamic and conventional banks. Section 5 presents and discusses our application to the GCC banking system. Some conclusions are given in the last section. 1- The Panzar and Rosse approach: theory Panzar-Rosse approach provides an assessment of the market structure and competitive conditions through the income elasticity in relation to input prices. H-statistic calculated derived from the estimation of a production function in its reduced form defined by the sum of the income elasticities to variations in input prices. Panzar-Rosse model is a non-structural approach is part of the new theory of industrial organization. This approach emphasizes the impact of prices of production factors on the income of the banking firm to determine the nature of competition in banking systems, which can serve as a measure of competitive behaviour of banks. This is an approach that provides a measure of market power of a bank and the structure of the banking market. The Panzar-Rosse approach is based on a number of assumptions including: The cost structure is homogeneous. The production function is Cobb-Douglass standard. The elasticity of demand relative to the price is higher than unity. The banks operate in their long-term equilibrium. Banks' performance is influenced by the actions of other banks (except in the case of monopoly market structure). The Panzar-Rosse approach has undergone several deepening conducted by contemporary economists. They have added other assumptions that are crucial to apply this approach to banking. And Gelos and Roldos (2002) added two fundamental assumptions. On the one hand banks are firms mono produced in which labor, capital and funds are considered as inputs. On the other hand input prices are assumed to be not associated with high quality services. In contrast, prices could imply high incomes that will biased the value of H-statistic. Panzar-Rosse have drawn a test of competition based on the estimation of income function in its reduced form defined by the sum of elasticities bank revenue compared to variations in input prices that allows assessment of the situation market for each banking system. 3 The test is derived from a general banking market model, which determines equilibrium output and the equilibrium number of banks, by maximising profits at the bank level and the industry level. This implies, first, that bank i maximises its profits, where marginal revenue equals marginal cost: 𝑹∗𝒊 (𝒙∗ , 𝒏∗ , 𝒛∗ ) − 𝑪∗𝒊 (𝒙∗ , 𝒘∗ , 𝒕∗ ) = 𝟎 Ri : equilibrium value of revenue of bank i Ci : costs of bank i (the prime denoting marginal) xi : the output of bank i n: the number of banks wi : a vector of m factor input prices of bank i, zi: is a vector of exogenous variables that shift the banks revenue function ti: a vector of exogenous variables that shift the banks cost function. Market power is measured by the extent to which a change in factor input prices ( 𝝏𝑾𝒌𝒊 ) is reflected in the equilibrium revenues (𝝏𝑹∗𝒊 ) earned by bank i. Panzar and Rosse define a measure of competition H as the sum of the elasticities of the reduced-form revenues with respect to factor prices: 𝒎 𝑯 = ∑(𝝏𝑹∗𝒊 /𝝏𝑾𝒌𝒊 )(𝑾𝒌𝒊 /𝑹∗𝒊 ) 𝒌=𝟏 The strongest assumption of the model is that banks operate in their long-run equilibrium. This test is derived from a general equilibrium model that determines the market equilibrium maximizing profits at each bank in the market. More precisely bank i maximizes its profit when its marginal revenue equals its marginal cost. The condition of long run equilibrium is maintained on the entire market. 𝑹∗𝒊 (𝒙∗ , 𝒏∗ , 𝒛∗ ) − 𝑪∗𝒊 (𝒙∗ , 𝒘∗ , 𝒕∗ ) = 𝟎 The calculation of the statistic H is defined by the sum of elasticities income of bank compared to variations in input prices. This statistic reflects an assessment of the degree of competition in the banking market. 𝒎 𝑯 = ∑(𝝏𝑹∗𝒊 /𝝏𝑾𝒌𝒊 )(𝑾𝒌𝒊 /𝑹∗𝒊 ) 𝒌=𝟏 4 Table-1 Discriminatory power of H- Statistic Values of H Competitive environment H< 0 Monopoly equilibrium: each bank operates independently as under monopoly profit maximisation conditions (H is a decreasing function of the perceived demand elasticity) or perfect cartel. 0<H<1 Monopolistic competition free entry equilibrium (H is an increasing function of the perceived demand elasticity). H =1 Perfect competition. Free entry equilibrium with full efficient capacity utilisation. 2- literature review The aim of this paper is to measure the competitiveness of banking in Islamic banks testing the conditions of competition in the banking sector. To do this, the literature on the nature of competition in the banking industry is divided into two streams which are of structural approaches and non-structural approaches. Structural approaches analyze the market structure. The most popular is the current "Structure-Conduct-Performance" which is developed by Mason (1939) and further refined by Bain (1951). This approach requires that the market structure affects the behavior of firms in a given industry and this in turn affects performance. Actually the market organization can influence the nature of competition and price formation. However, this model has weaknesses in the theoretical and empirical framework. There is also the theory of structure-efficiency which has been developed by Dermetz (1973) and Petzman (1977). This model is the result of a change in the latter approach by announcing that this positive relationship is not the result of market power but high efficiency of firms with market shares of the most important. This model states that efficient firms decrease in size and market share due to their ability to generate more profits which 5 leads to market concentration. The structural approach is a non formal measure of competition. However, it is difficult to identify the most suitable structural model. Non structural models capable of measuring the competition will compensate the shortcomings of theoretical and empirical evidence deficiencies of structural models. Among these models the method of Panzar Rosse (1987) and the method of Bresnahan and etLau (1982). This model is based on the theory of industrial organization. This model was based on the model developed by Iwata (1974) they have solved the problem of econometric identification. The model maximizes the profit of banks in oligopoly. This is a static model that determines in the short term the market power. The conjunctural variation of parameters is determined by the simultaneous estimation of the supply side and the demand based on time series data for the banking industry studied. Panzar Rosse model (1987) is derived from the literature of the new theory of industrial organization. Panzar Rosse model (1987) is derived from the literature of the new theory of industrial organization. Panzar and Rosse suggests a measure of market power and the conditions of competition in the sector from the impact of prices of production factors on the revenues of firms in this sector. They developed an empirical test of competition called the "H-statistic" by estimating the revenue function in its reduced form defined by the sum of the elasticities of production in relation to price variations of production factors to assess the structure banking market. The method of Lerner (1934) also comes from the theory of industrial organization. It measures the market power depending on the elasticity of demand with respect to price. More specifically when the elasticity is higher over the market price, it is close to marginal cost. This measure suggests the extent of monopoly power of a firm in the industry to which it belongs through its rate of profit: the difference between the selling price and the marginal cost of production. The index takes the value zero if the firm operates in a condition of perfect competition and the value 1 if it operates in a monopoly situation. The measure of competition in the banking industry is likely to be evaluated by several methods which differ according to the origin of the theory: the theory of industrial organization and the new 6 theory of industrial organization. To perform an empirical explanation consistent with the proposed issue taking into account the characteristics of banks in each country, the structural approach is most suitable. Without being exhaustive, we will limit the Panzar and Rosse approach to study competition in the banking sector. Studies dealing with this subject in the case of conventional banks based on the method of PanzaretRosse are numerous (Bang Nam Jeon et al. (2011). Rima Turk Ariss (2009), and Eduardo Levy Yeyatt Ale Jundro Micco (2007), Kent Mathews et al. (2007). Saeed al-Muharem et al. (2006) ...). They are configured with different characteristics and in different countries. However, the literature on the comparison between competitiveness and the concentration of Islamic banks and conventional is too limited. Recent studies deal with the conditions of competitiveness for Islamic and conventional banks in the banking market and the differences in profitability. In the same lineage, Laurent Weill (2010) and Rima Turk Ariss (2010) compare the market power of Islamic banks compared to conventional banks in the case of MENA countries. Finally, Muhamed-Zulkhibri et al. (2007) study the market structure of Islamic banking in Malaysia and assess the degree of competition in the sector. Most of the work dealing with the problem of measuring the competitiveness of banking in Islamic banks have looked to the MENA countries. Paper the most recent being that of Rima Turk Ariss (2010). It is based on a sample of 58 Islamic banks and conventional banks 192 for the period 2000-2006 for 13 countries belong to the MENA. The author evaluated the competitive conditions of Islamic banks and conventional banks based on indicators of market power. The model used is the Panzar-Rosse model. The application of this model implies that the monopolistic competition market structure is described by Islamic and conventional banks. The ratios of concentrations are three times higher for Islamic banks. The Lerner index indicates that Islamic banks have market power higher than their peer (conventional banks). The regression results show that Islamic banks do not get higher returns compared to conventional banks. 7 Laurent Weill (2010) conducts his analysis of purchasing power by a comparison between Islamic banks and conventional banks in the case of MENA countries. The paper deals 1166 135 Islamic banks and conventional banks for the period 2000-2007. The models used are Panzar-Rosse model and approach to industrial organization model based on the SCP (Structure-Conduct-Performance). Comparing the Lerner index for Islamic banks and conventional banks, the basic conclusion is the absence of a significant difference in terms of market power between Islamic and conventional banks. Regression of this index and the comparison states suggests that the market power of Islamic banks is lower than the market power of banks. Estimation of the model Panzar-Rosse tends to confirm our conclusion. Muhamed Abdul-Zulkhibri Majidet Fadzlan Sufian (2007) discuss the market structure of Islamic banking in Malaysia and analyzes the degree of competitiveness. The model used is the Panzar-Rosse highlighting 17 Islamic banks for the period 2001-2005. The application of this model to the case of Malaysia confirms previous studies in the case of Malaysia and in other developing countries. The market structure of the Islamic banking system is described by monopolistic competition. Table 2- the Panzar-Rosse application in previous studies Authors Period Countries Results Bang Nam Jeon et al (2011) (1997-2008) 17 countries of Asia and Latin America Monopolistically competitive Rima Turk Ariss(2009) (2000-2006) monopolistic competition Mahmoodul Hasan Khan (2009) (1997-2007) MENA countries Pakistan Kent Mathews et al (2007) (1988-2004) Saeed Al-Muharem et al (2006) (1993_2002) Great Brittan GCC countries monopolistic competition monopolistic competition Kuwait, Saudi Arabia, UAE: perfect competition Bahrain, Qatar, Oman: monopolistic competition EMamatzukis et al (2005) (1998-2002) Coccoresel (2004) (1997-1999) Jaccob A.Bikker et Katharina Haaf (1988-1998) 8 South East Europe countries Italy 23 countries monopolistic competition monopolistic competition monopolistic competition (2002) Phil Molyneux et al (1994) (1986-1989) Germany, France Italy, Spain, UK 8 France, Germany, Spain, UK: monopolistic competition Italy : monopoly Table 3- the Panzar-Rosse Application in Islamic Banks- Previous studies Authors Rima Turk Ariss(2010) Period (2000-2006) Muhamed Abdul-Zulkhibri Majid (2001-2005) &Fadzlan Sufian (2007) Laurent Weill (2010) (2000-2007) Countries Results MENA countries Islamic banking industry and conventional banking 58 Islamic Banks industry operates under monopolistic competition. 192 conventional banks Islamic Banks are less competitive than conventional banks Malaysia 17 Islamic Banks monopolistic competition MENA countries : 135 Islamic Bank and 1166 conventional Banks No significant difference of market power between Islamic banks and conventional banks. the Rosse-Panzar model confirms that Islamic banks are no less competitive than conventional banks. 3- Data description We use annual bank-level balance sheet and income statement data retrieved from the BankScope database to estimate the degree of banking competition and foreign. Our data set covers a total of 6 emerging economies from GCC for the period 2005-2010. This period was chosen for reasons related to the significant economic and crises experienced during this period. The final sample includes 15 Islamic and 47 conventional banks operating in 6 different countries. It is important to mark the structure of banks in our sample which is similar to the functional, institutional, and legal one. Our sample is homogeneous it probably makes more reliable estimates. The table 3 lists the number of banks and observations in each country. The figures shows that the number of Islamic banks operating in countries of the sample is very small compared to the number of conventional banks, which share of conventional banks in the sample is about 76 ℅. This shows that the conventional banking industry has a hegemonic character and the Islamic banking industry is still in its embryonic stage. Table4: Sample size of the Islamic and conventional banks data sets, 2005-2010. Countries SAUDI ARABIA QATAR U.A.E Kuweit Jordan Bahrain Total Islamic Banks Number of Banks Number of Observations 3 18 2 12 3 18 2 12 2 12 3 18 15 90 9 Conventional Banks Number of Banks Number of observations 6 36 4 24 11 66 8 48 9 54 9 54 47 282 4- Banking Market Concentration: an analysis of concentration measures The description of the market structure by the number of banks is a relatively limited .It ignores the distribution of banks in a given market. As the concentration indices, weighted by market share, taking into account both the number of banks and distribution, they are used as a proxy for the intensity of competition in the banking market. We used the ratio of concentration of n-banks, particularly the ratio of the three largest banks and the ratio of the five largest banks, the Herfindahl-Hirschmann, and the entropy measure by referring to the total of debt, total assets and total deposits.There is a need to investigate comparative measures of bank concentration. We first turn to the definition and discussion of three concentration indices. n-bank concentration ratio Simplicity and limited data requirements make the k bank concentration ratio one of the most frequently used measures of concentration in the empirical literature. The N bank concentration measure is simply the ratio between the assets of the N largest banks within a country divided by the total assets in the banking system at a given point in time. Giving equal emphasis to the n leading banks, but neglecting the many small banks in the market. There is no rule for the determination of the value of 𝐶𝑛 . 𝑪𝒏 = ∑𝒏𝒊=𝟏 𝒒𝒊 ⁄𝑸 = ∑𝒏𝒊=𝟏 𝑺𝒊 We present the concentration index of three and five banking firms by reference to total assets, total loans and total deposits as measures of the bank size. The calculation of the concentration ratios in the market shows that there are some banks that dominate the market in the case of Islamic banks and conventional banks. The concentration of conventional banks is two times higher than the concentration of Islamic banks. This result is the same in the case of the concentration ratios of three and five banking firms by reference to total assets. However, if one refers to the total deposits, the concentration of Islamic banks and conventional banks are almost similar for the period 20052008. While for the period 2009-2010, we found that the concentration of conventional banks is 10 galloping in 2009. It decreased in 2010 but it is still too high compared to Islamic banks. The decrease and the evolution towards higher are explained by the subprime crisis which has been sustained by all countries of the world including the Gulf countries which have been affected despite the high concentration of Islamic banks in this agglomeration. C3 (Assets) C5 (Assets) 0.4 0.3 0.2 0.3 Islamic 0.1 Convention al 0 2004 2006 2008 2010 2012 0.8 C3 (deposits) 0.2 Islamic 0.1 Conventional 0 2004 2006 2008 2010 2012 0.6 0.6 0.4 Islamic 0.4 0.2 Conventi onal 0.2 0 2004 C5 (deposits) 0.8 2006 2008 2010 2012 Islamic 0 2004 2006 2008 2010 2012 Source : authors' calculations Herfindahl-Hirschman Index: The HHI is the sum of the squared bank market shares. The Herfindahl-Hirschman index stresses the importance of larger banks by assigning them a greater weight than smaller banks, and it incorporates each bank individually, so that arbitrary cut-offs and insensitivity to the share distribution are avoided. 𝒒 𝟐 𝒊 𝑰𝑯𝑯 = ∑𝑵 𝒊=𝟏 ( 𝑸 ) The Herfindahl-Hirschmann shows the unequal size of the firm. In terms of total assets and total loans, the size of firms in the conventional banking industry is unequal. But compared against the Islamic banking industry, the size of conventional banks is even more unequal than Islamic banks 11 for the entire period. With reference to total deposits, the variation of the Herfindahl-Hirschmann, the variation of the inequality of size compared to conventional banking Islamic banking is marginal for the period 2005-2008. In 2009, the Herfindahl-Hirschmann is almost zero in the case of Islamic banks, while there is an increase in this index in the case of conventional banks. This growth continued in 2010 for both conventional and Islamic banks. IHH (Assets) 0.05 0.15 Index HirschmanHerfindahl (Deposits) 0.04 0.1 0.03 Islamique Islamic 0.02 conventionnell e 0.05 0.01 0 2004 2006 2008 2010 2012 Conventional 0 2004 Evolution of the concentration of conventional banks 0.35 0.8 0.3 0.6 0.25 C3 (deposits) 0.4 C5 (deposits) 0.2 0 2004 IHH (deposits) 2006 2008 2010 2006 2008 2010 2012 Evolution of the concentration of Islamic banks 0.2 C3 (deposits) 0.15 0.1 C5 (deposits) 0.05 0 2004 2012 2006 2008 2010 2012 Source : authors' calculations The ratios of concentrations C3etetC5 in the case of Islamic banks are higher for three periods than conventional banks and the Herfindahl-Hirschmann is six times higher than conventional banking industry. Entropy measure The index ranges between 0 and , and is therefore not restricted to [0, 1], as most of the other measures of concentration presented above. The value of the Entropy varies inversely to the degree 12 of concentration. It approaches zero if the underlying market is monopoly and reaches its highest value, E = log n , when market shares of all banks are equal and market concentration is lowest. The Entropy measure takes the form: 𝒏 𝒏 𝑬 = − ∑ 𝒔𝒊 𝐥𝐨𝐠 𝟐 𝒔𝒊 = −(𝟏⁄𝐥𝐧 𝟐) ∑ 𝐬𝐢 𝐥𝐧 𝒔𝒊 𝒊=𝟏 𝒊=𝟏 The entropy measure determines the unequal distribution of firms in the market. Its values are between 0 and with n is the number of firms in the market. In the case of our sample, the values of the entropy must be between 0 and 0 and for the Islamic banking market, and is between for the case of conventional banking market. With reference to total assets, the concentration is low in the conventional banking industry and market shares of all firms tend to be equal. The market is so competitive. However in the case of Islamic banks, the value of the entropy tends to 0. So the market is monopolistic. The structure of conventional and Islamic banking market is almost similar with reference to total deposits and total loans. E (Assets) E (Deposits) 5 4 4 3 3 Islamic 2 1 0 2004 2 Islamic 1 Conventional Conventional 2006 2008 2010 0 2004 2012 2006 2008 2010 2012 E (Loans) 5 4 3 Islamic 2 Conventional 1 0 2004 2006 2008 2010 2012 Source : authors' calculations 13 5- Evaluation Method of the H- Statistic Presentation of the model We proceed in our study of the model with the presentation which is the fundamental step in our empirical study. The PR methodology allows for the calculation of a measure of market structure, the H-statistic, as the sum of the elasticities of total revenues of the bank with respect to its input prices. We estimate equations banking income which they are explained by the price of production factors and bank-specific factors that have an impact on the long run equilibrium. The production function is Cobb-Douglas type: TRit= β (WL,it) β1 (WF,it)β2 (WK,it)β3 (Y1,it)γ1 (Y1,it)γ2 eεit The logarithm of this function is as follows: ln (TRit) = β + β1 ln(WL,it)+ β2 ln(WF,it) + β3 ln(WK,it) + γ1 ln(Y1,it) + γ2 ln(Y2,it) + εit On the basis of the theoretical assumptions of the model Panzar Rosse we estimate two linear equations by changing the dependent variable for each regression, all other things being equal. The first postulate is that we will refer to the intermediation approach, which states that any banking entity uses inputs in its interim financial function. The second calls for the estimation of the function of total income including interest and non-interest view that growth in other income excluding interest, such as commissions and capital gains is important. So according to theoretical foundations and the work of Molyneux, Thornton, and Lloyd-Williams (1996), Gelos and Roldos (2004) Classens and Leaven (2004), and Rima Turk (2010), the first regression is as follows: ln (TRit)= β +β1ln(WL,it) +β2ln(WF,it) +β3ln(WK,it) +γ1ln(Y1,it) +γ2ln(Y2,it) +εit (1) We will test later the equilibrium of long-term referring to the third theoretical postulate. This assumes that all banks in the industry are in their equilibrium of long-term. With reference to Nathan and Neave (1989) and Shaffer (1982), we tested the hypothesis using the following model: 14 ln (ROAit)= β +β1ln(WL,it) +β2ln(WF,it) +β3ln(WK,it) +γ1ln(Y1,it) +γ2ln(Y2,it) +εit E denotes the statistical equilibrium that it (2) is equal to the sum of input price elasticities with respect to income. If the value of E is different from zero then the market is not in equilibrium. In fact, long-term ROA do not relate to the price of inputs. TRit is the total income of bank i at time t. Modernization of the banking sector was followed by the increase in the share of interest income in the total income of the bank. However, with the emergence of new banking products, the share of non-interest income occupies an important part in the total income of the firm. Total revenue is measured by the ratio of total interest income and noninterest relative to total assets. WL,it is the cost of labor. This variable is represented by the ratio of personnel expenses to total assets. Personal expenses include salaries and wages. WF,it is the cost of funds. This variable is represented by the ratio of interest expense to total deposits. Interest expense includes interest paid on deposits and interest paid on loans. As for the funds, they are made by customer deposits, interbank deposits, subordinated debt and other longterm debts. WK,it is the cost of fixed capital. This variable comprises general and administrative costs, physical depreciation, depreciation of investment securities, and other overheads, This variable is represented by the ratio of other operating and administrative expenses of total assets. Y1,it this variable takes into account the specific risk to the bank. It is measured by the ratio of total net loans and total assets. It reflects both the credit risk and interest rate risk. If the ratio is high it indicates the relative illiquidity of the bank and the t limited ability to grant long-term credits, and the effects of loan quality, loan loss reserve was netted out from total loans. However, it is positively correlated with income. Y2,it : equity is an indicator of the ability of funding. It is expected to have a positive effect on income. if the bank has more equity it has the ability to give more credit and receive more revenue. This variable is measured by the ratio of total equity to total assets. The latter is used as a rough indicator of the size of the bank account for economies of scale. The influence of this variable is indeterminate because all positive influences on income can be offset by the banks of large sizes because they are able to diversify and spread the risk of their business. Empirical illustration We estimated the steps outlined above based on the same data set. The estimates were made by the Winrats program. The procedure was, first, to estimate the model for the whole sample with no distinction between Islamic and conventional banks. Was appointed associate model (1) where the endogenous variable is the total revenue. Thereafter it was entitled to two estimates for the case of Islamic and conventional banks which case the endogenous variable is the total revenue. Statistics H is defined as the sum of elasticities bank revenue to changes in input prices. So the results of the 15 estimation of the banking income allow for the determination of market structure in the banking industry for each category. The H-statistic is written as follows: 𝑯 = ∑𝟑𝒊=𝟏 𝜷𝒊 +∑𝟐𝒊=𝟏 𝜸𝒊 Table 5- the results of H – statistic All Banks Islamic Banks Conventional Banks WL 0.1526253 -0.463053 -0.1801023 WF 1.7361353 0.00000 -0.1745021 WK 0.0270870 0.00000 -0.2890106 Y1 0.0047361 0.00000 -0.0000157 Y2 0.9061875 -0.00000 6.0046954 H-statistic 2.8267712 -0.463053 5.3610647 Appendix 2 shows the H-statistic for each industry derived from individual regressions using bank effects (fixed or random depending on the results of the Hausman test), and the H-statistic calculated from a pooled regression on all countries estimated with country and bank random effects. According to the results, table 7, the average level of H-statistic is positive and varies between 5.3610647 and 2.8267712 in the case of the banking industry in general and the case of the conventional banking industry in particular. These values are well above zero rows. So the regression results are too consistent with the case of the perfect competition. So the perfect competition seems to describe the competitive behavior of the banking industry better than in general and conventional banking industry in particular in the case of the Gulf countries. However in the case of Islamic banks, the value of H-statistic is above zero which is of the order of -0.463053. So the regression results are too consistent with the case of monopoly. The market structure of the Islamic banking industry in the Gulf countries is described by a monopoly. This result is different from the work of Rima Turk (2010) which suggests that monopolistic competition is the best description of the market structure for Islamic and conventional banks. Further to the decomposition 16 of elasticities of various production factors shown in the table above, we noticed that the principal price of production factors inducing the amplitude of H-statistic is the price of the fund for the case of the banking industry in general. This can be justified by the nature of the activity of banks which is essentially the management of funds and diversification of risk. This is very consistent with the intermediation approach. In the case of banking industry in general the price of labor is an important way on the value of Hstatistic. This is explained by the fact that conventional banks in the course hold a competitive advantage over its competitors in research and develop and recruit qualified personnel. the case of the banking industry in general and the case of the conventional banking industry in particular, perfect competition is mainly explained by the homogeneity of products and services offered by Islamic banks. There is a standardized level of conventional financial products. The innovation process in Islamic banks is static. There must be a dynamic level of the innovation process. More specifically banking industry in GCC countries must invest more in research and development and training of these employees. The H-statistic calculated that Islamic banking industry in the GCC region generally operates under conditions of monopoly. The monopoly is explained by the very high concentration of the banking sector. In fact it can occupy a larger market share of the global market in which competition is increasing in this case beyond national borders to international ones. Islamic banks are still in the embryonic stage, even in a community of Muslim countries where Islamic finance has seen its real growth. Table 8: the results of Equilibrium tests E= -0.1315917 E= 6.894867 E= 0.510937 Model Agregate Islamic Model Conventional Model The results of equilibrium tests for all markets shown in Table-8 indicate that all countries are not in an equilibrium state. The calculated E-statistic is not significantly different from 0 for all banking 17 industry considered and for conventional banking industry. This implies that the banking market is in long-run equilibrium. However, the calculated E-statistic is significantly different for Zero for all Islamic banking industry. Therefore we can think of the banking sector in each economy as not being in long-run equilibrium. Conclusion This article sought to assess competitive conditions and concentration in the GCC banking industry as many as 6 countries over 6 years (2005–2010). Our stating idea was to detect the level of penetration of Islamic banks in the country with a majority Muslim population and to study the evolution of the phenomenon of economic concentration of Islamic banks in the banking industry in general. The Panzar–Rosse approach has been applied to obtain a measure of competitive conditions in to GCC countries over a time span of 6 years. The empirical evidence indicates that GCC banks in general and conventional banks earn revenues as if they were under conditions of perfect competition. Therefore, the recent tendency to concentration (usually considered with worry from a theoretical point of view) appears not to have had a significant impact on the conduct of the firms operating in the GCC banking industry. The results for GCC countries also suggest this conclusion, although the data do not represent long-run equilibrium values. Concentration in the banking markets of GCC countries was measured using various k-bank concentration ratios, the Herfindahl index, and the entropy measure. The results of the calculation of concentration ratios show that conventional banks have a very high degree of concentration in the market. While the concentration of Islamic banks is even lower than their pair (conventional banks). This sort of conclusion may help policy makers in determining future measures aimed at improving efficiency of these banking systems. The study of concentration and competitiveness of Islamic banks appears a necessity, or even an emergency to meet the challenges of the banking industry. The penetration of Islamic banks is unavoidable to go to meet the challenges and overcome moments of crisis suffered in the context 18 of financial globalization phase. Our findings suggest that the banking market in the GCC countries tends to success to confront the forces of international banking competition by developing a competitive structure. Acknowledgements The authors are grateful to three anonymous referees for helpful comments and advice. Naturally all remaining errors are ours. 19 References Bang Nam Jeon & al, 2011, « Do foreign banks increase competition? Evidence from emerging Asian and Latin American banking markets», Journal of Banking et Finance, 35 (2011) 856–875 Bain, J.S. (1951), « Relation of profit rate to industry concentration: American manufacturing 1936-1940», Quarterly Journal of Economics 65, 293-324. Bikker, J.A. and J.M. Groeneveld (2000), « Competition and Concentration in the EU Banking Industry», Kredit und Kapital 33, 62-98. Bikker, J.A. and K. Haaf 2002, « Competition, concentration and their relationship: an empirical analysis of the banking industry», Journal of Banking et Finance, (forthcoming). Bresnahan, T.F. 1982, «The Oligopoly Solution Concept is Identified», Economic Letters 10, 8792. Bresnahan, T.F. 1989, «Empirical studies of industries with market power», in: Schmalensee, R., Willig, R.D. (Eds), Handbook of Industrial Organisation, Volume II, 1012-1055 Campbell J.P, 1977, « On the Nature of Organizational Effectiveness » in Goodman P.S., Pennings J.M., New Perspectives on Organizational Effectiveness, Jossey-Bass. Cetorelli, N. 1999, «Competitive Analysis in Banking: Appraisal of the methodologies», Economic Perspectives, Federal Reserve Bank of Chicago, 2-15. Chamberlin, E., 1933, «The Theory of Monopolistic Competition», Harvard University Press, Cambridge, MA. Coccorese, P., 2004, «Banking competition and macroeconomic conditions: A disaggregate analysis»,, Journal of International Financial Markets, Institutions and Money 14, 203–219. Davies, S.W. (1979), « Choosing between Concentration Indices: The Iso-Concentration Curve»,Economica 46, 67-75. Davies, S.W. (1980), « Measuring Industrial Concentration: An Alternative Approach», Review of Economics and Statistics 62, 306-309. De Bandt, O., & Davis, E. P. (2000) «Competition, contestability and market structure in European banking sectors on the eve of EMU», Journal of Banking and Finance, 24, 1045−1066. Demsetz, H., (1973), « Industry structure, market rivalry and public policy», Journal of Law and Economics 16, 1- 10. 20 EMamatzukis & al, 2005, « Competition and concentration in the banking sector of the South Eastern European region», Emerging Markets Review , 6 (2005), 192–209. Eduardo Levy Yeyatt & Ale Jundro Micco, 2007, Concentration and foreign penetration in Latin American banking sectors: Impact on competition and risk, Journal of Banking et Finance, 31 (2007), 1633–1647. Gelos, R. G., & Roldos, J. (2004), « Consolidation and market structure in emerging market banking systems»,, Emerging Markets Review, 5, 39−59. Hannah, L. and J.A. Kay (1977), «Concentration in Modern Industry», MacMillan Press, London. Hall, M. and N. Tideman (1967), «Measures of Concentration», American Statistical Association Journal, 162-168. Rosenbluth, G. (1955), Measures of Concentration, in Business Concentration and Price Policy, National Bureau Committee for Economic Research, Princeton, 57-99 Hart, P.E. (1975), «Moment Distribution in Economics: An Exposition», The Journal of the Royal Statistical Society, series A, 138, 423-434. Hause, J.C, 1977, « The Measurement of Concentrated Industrial Structure and the Size Distribution of Firms», Annals of Economic and Social Measurement 6, 73-107. Iwata, G. "Measurement of Conjectural Variations in Oligopoly," Econornetrica, Vol. 42 (13791, pp. 947-966. Kwoka, J. (1985), «The Herfindahl Index in Theory and Practice», Antitrust Bulletin 30, 915-947. Lau, L.J., 1982, «On identifying the degree of competitiveness from industry price and output data», Economics Letters 10, 93-99. Laurent Weil, 2010, « do Islamic banks have greater market power? », BOFIT Discussion papers 2-2010. Lerner, Abba P. 1934. “The Concept of Monopoly and the Measurement of Monopoly.” Review of Economic Studies, 1: 157-175. Mason, E.S., 1939. Price and production policies of large-scale enterprises», American Economic Review 29 (Supplement), 61–74. Matthews, Murinde, K. V., et Zhao, T. (2007). Competitive conditions among the major British banks», Journal of Banking et Finance, 31(7), 2025–2042. 21 Molyneux, P., D.M. Lloyd-Williams and J. Thornton (1994), «Competitive conditions in European banking, Journal of Banking et Finance 18, 445-459. Molyneux, P., Y. Altunabas and E. Gardener (1996), «Efficiency in European Banking», John Wiley et Sons Ltd. reference Muhamed-Zulkhibri et al., 2007, «Market Structure and Competition in Emerging Market: Evidence from Malaysian Islamic Banking Industry», Journal of Economic Cooperation, 28,2 (2007), 99-121. Muharrami, S., Matthews, K., et Khabari, Y. (2006). Market structure and competitive conditions in the Arab GCC banking system. Journal of Banking et Finance, 30(12), 3487–3501. Panzar, J.C. and J.N. Rosse (1987), «Testing for ‘Monopoly’ Equilibrium», Journal of Industrial Economics 35, 443-456. Panzar, J.C., Rosse, J.N., 1982, «Structure, Conduct and Comparative Statistics», Bell Laboratories Economic Discussion Paper No. 248. Peltzman, S. (1977),«The gains and losses from industrial concentration», Journal of Law and Economics 20, 229-263. Shaffer, S. (1983),«Non-structural measures of competition», Economic Letters 12, 349-353. Shaffer, S., and J. DiSalvo (1994), «Conduct in a banking duopoly», Journal of Banking et Finance 18, 1063- 1082. Stigler, G.J. (1964), «A Theory of Oligopoly», Journal of Political Economy 72, 44-61. SUDIN HARON, 1996,« Competition and other external determinants of the profitability of Islamic banks », Islamic Economic Studies, Vol. 4, No. 1, December 1996. Turk Ariss, R, (2009),« Competitive behavior in Middle East and North Africa banking systems», Quarterly Review of Economics and Finance, 49, 693−710. Turk Ariss, R, 2010, «Competitive conditions in Islamic and conventional banking: A global perspective», Review of Financial Economics 19 (2010) 101–108 Weiss, L.W. (1989), «Concentration and Price», MIT Press, Cambridge (Mass.) and London. White, A.P. (1982), «A Note on Market Structure Measures and the Characteristics of the Markets that they ‘Measure’»,Southern Economic Journal, 542-549. 22 Appendix 1 3-Banks Index Concentration 2005 2006 2007 2008 2009 2010 C3 (deposits) Islamic Conventional 0,10258194 0,19192997 0,09902329 0,15251665 0,18211511 0,31249836 0,26664465 0,31559097 0,01120973 0,71636996 0,07455837 0,44766648 C3 (Loan) Islamic Conventional 0,15059951 0,18424247 0,13278345 0,19352794 0,13828704 0,18844455 0,15025185 0,20114203 0,15677682 0,22061411 0,16050524 0,23326031 C3 (Assets) Islamic Conventional 0,1129183 0,19099889 0,11281041 0,18501362 0,11541345 0,1824176 0,12688453 0,18631007 0,13536846 0,20708316 0,14034673 0,21486276 5-Banks concentration Index 2005 2006 2007 2008 2009 2010 C5 (deposits) Islamic Conventional 0,10258194 0,28200397 0,09902329 0,24019802 0,18211511 0,37802686 0,26664465 0,27738147 0,01120973 0,72340958 0,07455837 0,51289813 C5 (Loans) Islamic Conventional 0,18520836 0,29032098 0,17199042 0,30438236 0,18338152 0,29861883 0,1894472 0,30440873 0,19516895 0,31905024 0,20069041 0,34210945 C5 (Assets) Islamic Conventional 0,1129183 0,29837198 0,11281041 0,28725493 0,11541345 0,28918337 0,12688453 0,30138002 0,13536846 0,32943033 0,14034673 0,33597143 Herfindal-Hischman Index 2005 2006 2007 2008 2009 2010 IHH (deposits) Islamic Conventional 0,00592987 0,03667525 0,02202562 0,04521244 0,01904352 0,06841192 0,04310182 0,06793099 6,0942E-05 0,12540462 0,00304982 0,14253137 IHH (Loans) Islamic Conventional 0,01273151 0,04567499 0,00825766 0,04249717 0,00850456 0,04199886 0,00919863 0,04364158 0,00998077 0,04568467 0,01055759 0,035667 IHH (Assets) Islamic Conventional 0,00643596 0,03951757 0,00601749 0,03908138 0,00609025 0,03897188 0,00707457 0,04021875 0,00766132 0,04238889 0,00836749 0,04282191 Entropy Measure 2005 2006 2007 2008 2009 2010 E (deposits) Islamic Conventional 1,12230001 3,73357113 1,45107695 3,51219762 1,25170961 3,35993166 1,40087111 3,27868161 0,19846123 1,17597851 0,95000593 3,25244806 E (Loans) Islamic Conventional 1,05215042 3,79248679 1,02303977 4,173252 1,07760014 3,80787182 1,0961167 3,46781421 1,12081822 3,67130495 1,16018733 3,61773464 23 E (Assets) Islamic Conventional 0,98284437 4,05240982 0,99781594 4,03056934 1,02611664 3,99218103 1,06087546 3,92137074 1,08676396 3,84675464 1,17558971 3,76905624 Appendix 2 Appendix 2-1Panel regression results of competitive condition for total revenue ( Model Agregate Banking industry -GCC countries) Panel Regression - Estimation by Random Effects Dependent Variable Y Panel(6) of Annual Data From 1//2005:01 To 62//2010:01 Usable Observations 371 Degrees of Freedom 365 Total Observations 372 Skipped/Missing 1 Centered R**2 0.999716 R Bar **2 0.999712 Uncentered R**2 0.999931 T x R**2 370.974 Mean of Dependent Variable 31.417789757 Std Error of Dependent Variable 17.873447208 Standard Error of Estimate 0.303191936 Sum of Squared Residuals 33.552752677 Variable Coeff Std Error T-Stat Signif ******************************************************************************* 1. WL 0.1526253 0.1073310 1.42201 0.15502441 2. WF 1.7361353 5.3914985 0.32201 0.74744246 3. WK 0.0270870 0.7364978 0.03678 0.97066189 4. Y1 0.0047361 0.0276241 0.17145 0.86387287 5. Y2 0.9061875 12.4614713 0.07272 0.94202962 6. Constant -275.2022614 215.6067826 -1.27641 0.20181127 Appendix 2-2 Panel regression results of competitive condition for total revenue ( Conventional banking Industry -GCC countries) Panel Regression - Estimation by Random Effects Dependent Variable Y Panel(6) of Annual Data From 1//2005:01 To 47//2010:01 Usable Observations 281 Degrees of Freedom 275 Total Observations 282 Skipped/Missing 1 Centered R**2 0.999283 R Bar **2 0.999270 Uncentered R**2 0.999849 T x R**2 280.958 Mean of Dependent Variable 33.451957295 Std Error of Dependent Variable 17.299628549 Standard Error of Estimate 0.467487948 Sum of Squared Residuals 60.099869913 Variable Coeff Std Error T-Stat Signif ******************************************************************************* 1. WL -0.1801023 0.1363672 -1.32072 0.18659603 2. WF -0.1745021 7.2434657 -0.02409 0.98078005 3. WK -0.2890106 0.7638188 -0.37838 0.70515143 4. Y1 -0.0000157 0.0304842 -5.15536e-04 0.99958866 5. Y2 6.0046954 29.2948438 0.20497 0.83759206 6. Constant 394.9140193 273.9479347 1.44157 0.14942486 24 Appendix 2-3 Panel regression results of competitive condition for total revenue (Islamic Banking Industry- GCC countries) Panel Regression - Estimation by Random Effects Dependent Variable Y Panel(6) of Annual Data From 1//2005:01 To 15//2010:01 Usable Observations 89 Degrees of Freedom 85 Total Observations 90 Skipped/Missing 1 Centered R**2 0.990460 R Bar **2 0.990124 Uncentered R**2 0.996701 T x R**2 88.706 Mean of Dependent Variable 24.752808989 Std Error of Dependent Variable 18.097137455 Standard Error of Estimate 1.798487169 Sum of Squared Residuals 274.93726828 Variable Coeff Std Error T-Stat Signif ******************************************************************************* 1. WL -0.463053 0.290946 -1.59155 0.11148692 2. WF 0.000000 0.000000 0.00000 0.00000000 3. WK 0.000000 0.000000 0.36653 0.71396837 4. Y1 0.000000 0.000000 0.00000 0.00000000 5. Y2 -0.000000 0.000000 -0.07296 0.94183429 6. Constant 953.764967 584.408137 1.63202 0.10267556 Appendix 3 The H statistic is calculated as the sum of the input price elasticities given by: H=𝜷𝟏 +𝜷𝟐 +𝜷𝟑 +𝜸𝟏 +𝜸𝟐 Similarly the long-run equilibrium equation (2) was modified in the same way as above and described below. The long-run equilibrium test is given as: E=𝜷𝟏 +𝜷𝟐 +𝜷𝟑 +𝜸𝟏 +𝜸𝟐 25