Abstract
Major steps were taken in Indian banking sector in the early 1990s to liberalise commercial bank functioning in order to improve their efficiency. The measurement of bank efficiency provides the foundation for consequent inquiry into reasons producing efficiency differences. This article measures technical efficiency of balanced panel of 54 commercial banks operating in India during 1991–92 to 2006–07 using Data Envelopment Analysis (DEA) with an aim to study effects of reforms and ownership on bank efficiency. Since existing literature lacks uniformity in methodology used for identifying determinants of efficiency in banking, this study employs a blend of tests including profitability analysis (suggested by Spong et al., 1995) and found evidence that financial reform, ownership and listing of bank shares have influenced bank efficiency in India. However, no conclusive evidence was found regarding a relationship between size and efficiency. In addition, ‘most’ efficient banks were found to be characterised by higher Net Profit as percentage of Total Assets (NPTA) and higher Profits per Employee (PPE) while ‘least’ efficient banks reported higher levels of Non-Performing Assets (NPAs).
Keywords
Introduction
Importance of banks stems from their role as main channels of savings and allocators of credit in an economy. For an emerging economy like India, faster growth essentially depends on an efficient banking sector, its efficiency having far-reaching effects on the financial system and other parts of the economy. Moreover, size and complexities of today’s globalised economies place heavy demands on their banking systems. Banking systems that are efficient are better equipped to meet these demands and to grow while inefficient banking systems have the inherent capacity to weaken and endanger the entire economic system of which they are an integral part. Thus, identification of determinants of efficient banks can help in taking appropriate steps by policymakers and bankers in strengthening the banking sector.
However, it is not easy to define efficiency, let alone measure it. The concept of efficiency and its measurement varies from one system to another. Economists define efficiency with respect to the relationship between scarce factor inputs and outputs of goods and services. Theoretically, efficiency from management point of view refers to the achievement of ends with least amount of resources (Koontz and O’Donnell 1978). In practice, managers define efficiency in terms of performance and as an index of deviation of ‘observed’ performance and ‘desired’ performance. Simply stated, the term ‘efficiency’ may be used to refer to the quality or degree of producing a set of desired effects.
In addition, efficiency can be studied from different perspectives. Efficiency may be ‘productive’ efficiency, ‘cost’ efficiency, ‘revenue’ efficiency or ‘profit’ efficiency. ‘Productive’ efficiency is characterised by minimum input bundles required to produce various outputs, or the maximum output producible with various input bundles and a given technology. ‘Cost’ efficiency refers to the existence of minimum expenditure to produce a given bundle of outputs, given the prices of the inputs used in its production and the given technology. ‘Revenue’ efficiency refers to maximum revenue obtainable from a given bundle of inputs, given the prices of the outputs produced and given the technology in place. ‘Profit’ efficiency, on the other hand, refers to the maximum profit obtainable from production activity, given the prices of the outputs produced and given the technology in place. Cost, revenue and profit efficiency are economic concepts whose measurement requires both price information and the imposition of an appropriate behavioural objective on producers (Kumbhakar and Lovell 2000).
In banking, price determination is not easy since banking operations involve the use of multiple inputs to produce multiple heterogeneous outputs. Moreover, imposing a behavioural objective such as cost minimisation or profit (revenue) maximisation on banks is difficult as banks pursue different objectives simultaneously that may change from time to time and make such efficiency measurement unsuitable. ‘Productive’ efficiency or ‘Technical’ efficiency, on the other hand, is a purely physical notion that can be measured without recourse to price information and without having to impose a behavioural objective on the producer (bank). This makes the concept of ‘technical’ efficiency apt for studying the efficiency of a service industry like banking. Under it, failure to achieve the best possible output levels and/or usage of an excessive amount of inputs is considered as technical inefficiency.
Indian Banking: A Historical Perspective
Post-independence, India went for a mixed economy, giving public sector the dominant role in the economy with the aim of becoming a self-sufficient welfare state (Kumar and Arora, 2010). Since the first dose of nationalisation in 1969, the domestic state-owned banks have been dominating the banking industry in India. They have played a crucial role in the development of the Indian economy and have financed industrialisation. During the pre-reforms era, the domestic state-owned banks enjoyed protection and had funding advantages in the form of subsidies. Also, the government determined the volume, allocation and price of credit resulting in financial repression. As consequence, until early nineties, Indian banking sector suffered from low capital base, high intermediation cost, low productivity, low profitability, low productivity and low competition. The beginning of 1990s brought the realisation in India that the efficiency of the banking system was not to be measured only by quantitative growth in terms of branch expansion and growth in deposits/advances or merely by fulfilment of social obligations of economic development; in fact, the operational efficiency of banks is equally important. It was also felt that the domestic banks working in a highly protected and regulated environment were not measuring up to international standards as far as efficiency was concerned. Consequently, a process of financial sector reforms was set in motion.
Against this backdrop, beginning early nineties, India embarked on the road to economic (including banking) reforms based on the recommendations of Narasimham Committee Report (Narasimham, 1991). Since then, Indian banks have been gradually exposed to the rigours of domestic (new Indian private banks) and international (foreign) competition in both deposit and credit markets. In 1998, Narasimham Committee gave its second report particularly on banking sector reforms. Based on these recommendations, the government adopted preventive measures aimed at further strengthening the reforms in the sector.
The objective of the reforms was to create a common operating environment for all banks operating in India (the banking sector in India consists of three types of banks, namely, domestic public sector banks, domestic private sector banks and foreign banks). It involved overcoming the external constraints related to the administered interest rates, high levels of pre-emption in the form of reserve requirements and credit allocation to certain sectors. The reforms also intended to enhance efficiency through infusing competition. Guidelines were laid down for establishment of new banks in the private sector and the foreign banks were allowed more liberal entry. In a step towards consolidation, initiatives were taken towards universal banking. Also, measures were introduced to ensure that the banking regulatory framework and supervisory practices in India converged with the best practices in the world (Y.V. Reddy 2005).
Reform measures had far-reaching impact on bank functioning. New requirements of maintaining capital adequacy and compliance with asset classification and provisioning norms were introduced as part of the banking reforms putting pressure on state-owned banks. Government regulations such as deregulation of interest rates on deposits and advances were relaxed leading to increased competition for public sector banks as they had to now contend with competition not only from other public sector banks but also from old/new private sector banks and foreign banks. Bank Nationalization Act of 1949 was amended and state-owned banks were allowed to access the market to raise funds from the public. Public involvement in bank ownership made the public sector banks more conscious of the need to run the institution efficiently. While the enabling policy framework and the operating environment provided a platform for improving the efficiency of the banking sector, its ultimate success depended on how the individual banks responded to the competitive environment. 1
The financial sector reforms, implemented in a phased manner over the last two decades, aimed at creating an efficient banking system. This article estimates efficiency of commercial banks operating in the Indian banking sector during the post-reform era with an aim to examine important determinants of bank efficiency. This exercise can prove helpful in providing important information which may be useful to bankers, regulators and the policymakers. The present study contributes to the existing literature on bank efficiency by explaining why some banks were more efficient as compared to others. Bankers can use the results of such an analysis to determine where their respective banks stand in relation to other banks in terms of efficiency and to make strategies to improve their bank efficiency. The regulators and policymakers can understand the behaviour of efficiency levels in banks and incorporate necessary dimensions of efficiency in the banking policies.
This section gave an overview of the reforms introduced in the Indian banking sector besides highlighting the importance of studying efficiency differences among banks. The rest of the article proceeds as follows. The second section gives a brief review of existing (though few) studies on determinants of bank efficiency. The third section deals with the methodology (theory and modelling) for calculating the efficiency estimates and identifying determinants of bank efficiency. The fourth section details the data set construction from Reserve Bank of India (RBI) regarding inputs and outputs over the period 1991–92 to 2006–07. The fifth section presents the results of the empirical analysis and the sixth section summarises and concludes.
Literature Review
Early efforts in the investigation of efficiency and its measurement were made by Koopmans (1951) and Debreu (1951). Both studied technical efficiency. Koopmans defined a feasible input–output vector to be technically efficient if it is technologically impossible to increase any output/or to reduce any input without simultaneously reducing at least one other output and/or increasing at least one other input. Using this definition, he was able to prove that an input–output vector is efficient if and only if it possesses a positive normal to the production possibilities set, and he provided an economic interpretation to this normal as a set of shadow prices thus, laying the foundation for measurement of efficiency by offering a definition and characterisation of technical efficiency. Debreu on the other hand, provided a measure or an index of the degree of technical efficiency with his coefficient of resource utilisation.
Over the years, numerous studies have emerged studying bank efficiency. The most popular technique used to measure efficiency in banking is Data Envelopment Analysis (DEA). It is a non-parametric technique developed by Farrell in 1957. He related one output to more than one input under the assumption of Constant Returns to Scale (CRS). His analysis was extended by Charnes, Cooper and Rhodes (CCR) (1978). They applied DEA to the case of more than one output while maintaining the assumption of CRS. The limitation of DEA being used only under CRS was removed by Banker et al. (1984) by making it applicable to increasing, constant or decreasing returns to scale. Sherman and Gold (1985) in their pioneering work applied DEA to banking and they evaluated the efficiency of a single cross-section of 14 branches of a US savings bank. Since then, DEA technique has been employed by various researchers to study bank efficiency. Recent studies that employed DEA include those by Berg et al. (1992, 1993), Chien et al. (2003), Drake and Howcraft (1994), Elyasiani and Mehdian (1995), Favero and Papi (1995), Sherman and Ladino (1995), Fukuyama (1995), Haag and Jaska (1995), Jemric and Vujcic (2002), Jackson and Fethi (2000), Miller and Noulas (1996), Portela et al. (2004), Sherman and Gold (1985), Lovell (1996) and Chen (2001) to name a few.
There are, however, very limited number of studies which examine the determinants of efficiency. Bhattacharya et al. (1997), Shanmugam and Lakshmanasamy (2002) and Mahesh and Rajeev (2006) are some of the studies that explored the impact of deregulation in the Indian banking sector on bank efficiency. The impact of ownership styles has been studied by Chatterjee (2006), Sarkar et al. (1998), Sathye (2003), Bhaumik and Dimova (2004), Ram Mohan and Ray (2004) and Bhaumik and Piesse (2004) to name a few. Das et al. (2004) reported that a bank listed on the stock exchange had higher efficiency level. They found that size has been a positive factor in determining the efficiency in state-owned banks. They also found evidence supporting the proposition that competition rather than foreign ownership per se is more likely to be a panacea to Indian banking sector.
Another study by Das and Ghosh (2006) found evidence supporting the impact of non-performing asset levels, size and asset quality on efficiency besides ownership style. International evidence on the relationship between bank efficiency and size is not very clear with different studies giving different results, prominent among them are studies by Attaullah et al. (2004), Miller and Noulas (1996), Girardone et al. (2004), Berger and Mester (1997) and Pi and Timme (1993). A few studies such as Spong et al. (1995), Ataullah et al. (2004) and Casu and Girardone (2004) attempted to discover financial characteristics of ‘more efficient’ and ‘less efficient’ banks.
An attempt was made by Amarender Reddy (2004) to determine factors influencing bank efficiency only 2001 and 2002 years data. It was found that banks with more number of branches were having significant negative association with both technical efficiency and scale efficiency. Priority sector advances to total advances were having significant positive association with scale efficiency. Share of non-approved investment to total investment was having significant negative association with all efficiency measures. This might have been due to the banks having acquired high risk assets that not only increase credit risk, but also increase capital requirement for provisioning for high risk assets leading to lower efficiency. High capital adequacy ratio (CAR) was having positive influence on pure technical efficiency and negative association with scale efficiency. As against share of branches, share of total assets was having significant positive association with all efficiency measures. Old private banks were having significant positive influence on scale efficiency, while significant negative influence on pure technical efficiency. New private banks were having positive sign for all efficiency measures, but significant in overall technical efficiency and scale efficiency. Almost similarly, foreign banks were having significant positive influence on overall technical efficiency and scale efficiency.
Arun and Turner (2002) compared 27 public sector banks and 34 private sector banks in terms of five indicators of efficiency, namely, business per employee, cost of intermediation as proportion of total assets, non-interest income as a proportion of total assets, staff costs as a proportion of total assets and staff costs as a proportion of non-interest income. All indicators uniformly suggested that private sector banks operate substantially more efficiently than public sector banks. The authors also carried out regressions that control for the bank size, and found that privately-owned banks perform more efficiently than the public sector banks along all five dimensions.
The measurement of bank efficiency provides the groundwork for consequent investigation into causes that produce efficiency differences. It has been observed that studies examining bank efficiency reported conflicting results. This has been generally attributed to differences in the methodologies used. ‘The reasons for numerous conflicting results of banking studies may lie not only in their methodological differences, but also in the nature of the banking industry and its environment’ (Jackson 1972). Differences in bank efficiency scores across years may be attributed to common macro-economic and exogenous forces influencing all banks together. Examples of such forces are structure of the banking industry, banking regulations, swings in monetary and business cycles, etc. It is not possible to incorporate the macro-economic factors while building a model for individual banks of one country since the same macro-economic factors influence all the banks. However, when a cross-country comparison of bank performance is studied, the inclusion of these factors is possible. Or in countries like the US, where state branching laws exist, there is a possibility of incorporating these factors at the state level.
The reasons for efficiency differences may be attributed to two types of forces, namely (a) external macro-economic forces influencing all banks and (b) internal bank-specific forces. The present study identifies a number of bank-specific determinants of bank efficiency. While, differences in efficiency across banks may be due to forces internal to the bank itself such as bank objectives, bank conduct, size, ownership styles and managerial capabilities of banks. It may, however, not be possible to account for all bank-specific determinants. One such major determinant of bank activity is the set of goals to be achieved by each bank. The goals of the banks determine how the bank conducts its business. In India, the goals of domestic banks are definitely different from those of the foreign banks. Similarly, the goals of nationalised domestic banks are different from those of their private counterparts. Domestic nationalised banks are required to meet both commercial as well as social objectives. Private banks, on the other hand, operate primarily to earn maximum returns. Foreign banks are also solely run on commercial lines. Public sector banks are considered more as a public servant rather than a commercial entity. These different conflicting goals enter into the conduct of business of these banks. Another factor influencing bank behaviour is government regulation. For example, there is no explicit definition of public sector nature of nationalised banks in India. In the Budget (2000) speech, it was emphasised that ‘public sector nature of the nationalised banks will continue even if the government stake drops to 33 per cent’. They will not be able to enjoy ‘full’ operational autonomy and economic independence as long as there is no change in this policy of government. The goals together with public interest and bureaucratic considerations, determine government regulation and this in turn limit the range of bank behaviour. Government regulation influence entry and exit of banks and bank mergers are heavily regulated. For example, a loss making bank, New Bank of India, was made to merge with a profit-making bank Punjab National Bank (PNB). This resulted in dip in profits of PNB for the next couple of years. Government also resorts to recapitalisation of banks resulting in artificial protection given to banks (UCO bank) and huge burden on the National Exchequer. Keeping aside minimum Statutory Liquidity Reserves and Cash Reserve Ratios (SLRs and CRRs) as per the government regulation, the banks make financial investments. Domestic nationalised banks are not allowed to invest in more return–more risk avenues. They are required to follow conservatism. Private banks and foreign banks are relatively more independent regarding their investments.
Earlier studies found that foreign banks operating in India were more efficient than Indian banks (Sarkar et al. 1998). More recent studies report that competition in the Indian banking industry induced public sector banks to eliminate the performance gap that existed between them and both domestic and foreign private sector banks (Bhaumik and Dimova 2004). A noticeable presence of foreign banks in the Indian credit market coincided with a steady improvement in the technical efficiency of domestic banks resulting in domestic banks being more efficient than the foreign banks which had an initial advantage with respect to technical efficiency (Bhaumik and Piesse 2004). It may be argued that while comparing the performance of public sector banks with that of private and foreign banks, attention also needs to be focused on the differences between the objectives 2 of the two types of banks. The public sector banks have been expected to fulfil social objectives as well and on occasions specific advises are received from the central government in this regard. However, attainment of such social objectives may not have any implication on the operating efficiency of such banks.
Methodology: Theory and Model
The period of study has been divided into two sub-periods. The first sub-period spans from 1992–93 to 1997–98 (Phase I) while the second spans from 1998–99 to 2006–07 (Phase II). DEA is applied to measure (technical) efficiency of banks operating in India from 1991–92 to 2006–07. 3 DEA uses linear programming for the construction of non-parametric piece-wise surface or frontiers and the measurement of efficiency relative to the constructed frontiers. The production frontier under DEA is not determined by any specific functional form but is generated from the actual data for the evaluated firms (banks). In other words, the DEA frontier is formed as the piece-wise linear combination that connects the set of ‘best-practice observations’ in the data set under analysis, yielding a convex Production Possibility Set (PPS). Farell (1957) proposed the piece-wise convex hull approach to frontier estimation. The approach to frontier estimation proposed by Farrell was not given much detailed empirical attention for about two decades, until a paper by Charnes et al. (1978), in which the term Data Envelopment Analysis 4 was first used.
This article follows a two-step procedure. In step 1, technical efficiency scores are calculated for the period 1991–92 to 2006–07 using output oriented DEA for each of the 54 sample banks (27 domestic nationalised banks, 15 domestic private banks and 12 foreign banks) assuming Variable Returns to Scale (VRS). Second, various tests are applied on the estimated efficiency scores to identify determinants of bank efficiency. Tests were conducted to study the relationship between reforms, ownership, listing of bank shares, size and profitability. Relationship between ownership and efficiency has been analysed by investigating whether different types of banks differ significantly in terms of efficiency. To achieve this objective, the present study divides the banks into three mutually exhaustive bank types, namely, domestic public sector banks, domestic private sector banks and foreign banks. Estimated efficiency scores showed that sharp variations were witnessed by Indian banking sector in terms of efficiency scores in the post-reforms period and bank efficiency seemed to stabilise only in the latter three years of the period of study. Therefore, profitability test suggested by Spong et al. (2005) is applied over the period 2004–05 to 2006–07.
Data Structure
Not only is it difficult to define outputs and inputs to be included in a bank model, but their measurement is also fraught with difficulties. Though early attempts were made by Sealy and Lindley (1977), Colwell and Davis (1992) and Berger and Humphrey (1997) to define inputs and outputs in the banking industry, there is still no single definition of inputs and outputs of the banking firm. To get over the problem of measurement of inputs and outputs in banks, researchers take the help of different approaches available on banking.
Generally, two approaches are followed on modelling bank behaviour namely, (a) intermediation approach and (b) production approach. Intermediation approach (also known as the asset approach) assesses the performance of banks using as inputs the total operating and interest costs and outputs as the amount of money intermediated. In this approach, deposits are regarded as being converted into loans. As a result, three inputs namely, total operating costs, total interest costs and deposits are considered. Production approach (also known as service-provision approach) uses the value of loans and investments as measures of output while labour and capital as the inputs. This approach emphasises the services provided and banks are regarded as using labour and capital to generate deposits and loans. Banks as intermediaries render a variety of services in the intermediation process. These services include payments services, advisory services and portfolio management services. The potential outputs of banks include loans, deposits, cheques, stocks and bonds, investments, branches, interest and fees on loans, interest and fees on deposits, guarantees, etc. An accurate identification and measurement of inputs is as necessary as that of outputs for measuring efficiency. The banks consume different types of inputs to produce the necessary outputs. Potential inputs of banks may be labour, capital, deposits, interest and fees on deposits and material and machines. Table 1 lists the definition of outputs and inputs on the basis of the approach followed.
The treatment of bank deposits also poses a problem in the measurement of bank efficiency. Considerable debate in the literature surrounds the input–output status of deposits. Traditionally, deposits are regarded as the main ingredients for loan production and the acquisition of other earning assets.
Definition of Outputs and Inputs on the Basis of Approach Used
On the other hand, high value-added deposit products, such as integrated savings and checking accounts, investment trusts and foreign currency deposit accounts tend to highlight the output characteristics of deposits. There are different opinions among authors with regard to the definition of deposits. Some authors maintain that they should be regarded as input while others take them to be output. Deposits are thus, ‘simultaneously an input into the loan process and an output, in the sense that they are purchased as a final product providing financial services’. The debate regarding the treatment of deposits was pointed put by Wykoff (1992) by questioning when are deposits outputs, why they are so cheap and when are they inputs, why do people provide them to banks. The production approach treats deposits as output while the intermediation approach treats it as an input (refer Table 1). Table 2 provides a brief summary of the inputs and outputs used by various researchers in their endeavour at measuring efficiency in banks using DEA. Recognising that the primary function of banks is to intermediate inputs xn into outputs ym, this article follows the intermediation approach to define input and output variables of the DEA model facilitating taking the input and output variables in money terms rather than in physical terms. Further, deposits are taken as input in the bank intermediation process.
Post reforms, the Indian banking sector has three broad types of ownership forms, namely, domestic state-owned banks, domestic privately-owned banks and foreign-owned banks. These ownership forms have various advantages and disadvantages in relation to each other. Foreign banks have state of the art technology, access to capital, superior managerial expertise and risk diversification abilities. They have the disadvantage of inability to have access and interpret correctly the local information. Since the first dose of nationalisation in 1969, the domestic state-owned banks have dominated the banking industry in India. They have played a crucial role in the development of the Indian economy and have financed the large scale industrialisation. During the pre-reforms era, these state-owned banks did not have to compete with the private sector banks, both domestic and foreign banks. The domestic state-owned banks enjoy protection and have the advantage of funding advantages in the form of subsidies. They, however, suffer from the mandates of priority sector directed lending and lack of market discipline. The efficiency estimates of each ownership type are a reflection of the net effects of the comparative advantages and disadvantages.
The sample for the present study comprises of 27 state-owned Indian banks, 15 Indian private banks and 12 foreign banks aggregating to 54 banks. For the purpose of this study, all commercial banks that have continued their existence in India since 1991–92 to 31 March 2006–07 are included in the sample. The number of such banks is 54. Merged banks, new banks and those banks that ceased to exist during the period of study are dropped for the purpose of uniformity. The study relates to the period beginning with the initiation of banking sector reforms in India starting from the year 1991–92 and covers 16-year period till 2006–07. Data are balanced panel data and the number of observations is 864. The secondary data used in the study are obtained from the official database archived and accessible from the RBI’s (central bank of India) website (
Summary of Banking Inputs and Outputs Applied by Previous Authors
Input and Output Variables
Results
Efficiency Measurement
Figure 1 shows the efficiency movements experienced by the Indian banking sector in the post-reform era. As can be seen from the figure, on implementing the reforms, there was a sharp fall in the efficiency levels in the Indian banking sector. This was the result of the implementation of stringent regulatory and supervisory banking regulations. The sector took 2 years to assimilate the new reforms and in 1994, the efficiency levels experienced a gigantic jump. Ever since then, though there have been efficiency variations, the levels have been high. It can also be observed that the efficiency levels have become stable since 2004–05 making the period suitable for investigating the determinants of efficiency in banks. Since the reforms were introduced in 1991–92, it was considered appropriate to drop the year 1991–92 for further analyses to recognise the lag involved in the effect of reforms to materialise. The efficiency scores for all the 54 sample banks for the period of Phase I of banking reforms (1992–93 to 1997–98) are presented in Table 4.
Ownership Effects on Efficiency
It may be observed from Table 5 that 13.33 per cent of domestic private banks had low efficiency level banks.
This figure for domestic public banks and foreign banks was 22.22 per cent and 16.67 per cent, respectively. Further, the mean efficiency scores were studied for an in-depth analysis of efficiency scores by bank type. Table 6 gives the mean efficiency scores by bank type.
Figure 2 shows the trend of mean efficiency of banks in different ownership groups. It may be observed that on an average the efficiency of the private banks was more than the other bank groups. Table 6 divides the overall mean efficiency scores by bank types into that for Phase I and Phase II respectively in order to trace the trend of efficiency scores in the two phases of reforms.
It may be observed that there was not much difference in the average efficiency levels of domestic banks and foreign banks during the post-reform period although the banks in each category enjoyed higher efficiency in Phase II as compared to that in Phase I. During Phase I, foreign banks were the most efficient group followed by domestic private banks. During Phase II, the domestic banks (both public and private taken together) became the most efficient group. Within the domestic banks, private banks emerged the leaders in terms of efficiency. It may be concluded that reforms have proved beneficial for the domestic banks and have helped them attain higher levels of efficiency comparable with their foreign counterparts.
Efficiency Movements in the Post-reform Period 1991–92 to 2006–07
Mean Bank Efficiency Scores During the Period 1991–92 to 2006–07
The last row in Table 6 shows the average efficiency of bank groups during the entire period of study consisting of 16 years. During the entire period of the study, private sector banks turned out to be the most efficient group of banks. The second position was occupied by the public sector banks. Foreign banks occupied the third position with the lowest average efficiency score of 0.9270. An analysis of the overall efficiency levels of domestic and foreign banks based on trend patterns of efficiency scores during 1991–92 to 2006–07 shows that the overall efficiency levels of public sector banks appears comparable with foreign and private domestic banks. In general, foreign banks had higher efficiency levels than domestic banks (public-sector and private domestic banks) in terms of efficiency in Phase I. However, domestic banks overtook foreign banks in terms of efficiency in Phase II. The efficiency of foreign banks has somewhat deteriorated in Phase II. In fact, domestic and public-sector banks improved their efficiency over the sample period. Private banks had generally better efficiency levels over the period of study. Sathye (2003) also found evidence that fully public sector banks are catching up with the banks in the private sector. The inferior efficiency levels of public sector banks relative to private sector banks have been attributed by earlier researchers to (a) the larger share of credit extended to the public-sector undertakings, (b) more stringent requirements imposed on direct lending, (c) a lesser degree of diversification baking business and (d) lower interest rate margins across the banking sector.
Efficiency Levels by Bank Type
Efficiency Levels by Bank Type and Phase of Reforms
Public-sector banks appear to have improved their efficiency levels reasonably well in the post-reform period. The results are consistent with those reported by Das and Ghosh (2006) who also uncovered evidence that public sector banks recorded higher efficiency gains in the post-reform period. Bhaumik and Piesse (2004) attributed this phenomenon to the greater learning of the public sector banks with respect to best practices in the credit market than the learning of the foreign banks. As a result, the domestic banks which were not as efficient as the foreign banks on an average in phase I caught up with and indeed outperformed the foreign banks in Phase II. They stated that it was possible that the ability of the foreign banks to augment their efficiency was affected not by their slower rate of learning but by a paucity of high quality borrowers. Bhaumik and Piesse (2004) found that there was no perceptible difference between the inefficiency of foreign banks and public sector banks during the period 1996–98. Independent samples t-test is calculated to find if there exists any significant difference between efficiency of (a) domestic banks and foreign banks; and (b) domestic private banks and domestic public banks.
Trend of Mean Efficiency Scores by Bank Type (1991–92 to 2006–07)
Table 7 gives the results of Levene’s t-test for Equality of Means. The table contains two sets of analysis—the first one between domestic banks and foreign banks and the second between domestic public and domestic private banks. The t-test associated with equal variance not assumed for the first analysis is 2.850 with 26.253 degrees of freedom. The corresponding two tailed p-value is 0.008, which is less than 0.05 and 0.01. Therefore, the null hypothesis was rejected both at 5 per cent and 1 per cent significance levels.
It implies that the average efficiency of the two bank groups was significantly different from each other, that is, the domestic banks and foreign banks did not have the same efficiency. There was thus, a significant difference in the average efficiency of domestic banks and foreign banks operating in India during the period of study, that is, 1991–92 to 2006–07.
Significance of Difference in Efficiency Scores in Domestic Banks–Foreign Banks
The t-test result with equal variances not assumed for the second analysis showed t-statistic of -0.056 with 26.253 degrees of freedom. The corresponding two-tailed p-value was 0.955 which is higher than 0.05 and 0.01. Thus, no significant difference was observed between the efficiency of the two domestic bank groups. Table 8 reports the t-statistics for comparing the mean efficiency scores of different types of banks between Period I and Period II. The negative sign on a t-statistic indicates that Phase I had a lower mean efficiency score than in Period II. Statistical significance at 1 per cent and 5 per cent is represented by a bold p-value.
Significance of Difference in Efficiency Scores in Phase I – Phase II
The test results for all banks taken together showed a t-statistic of -8.511 with 53 degrees of freedom. The two-tailed p-value was 0.000 which is less than the conventional 1 per cent level of significance. This means that the average change in efficiency of all the 54 sample banks in Period II as compared to Period I was statistically significant.
The test results for domestic private banks showed a t-statistic of -4.405 with 14 degrees of freedom. The two-tailed p-value is 0.0012 which is less than the conventional 1 per cent level of significance. Therefore, the null hypothesis at 1 per cent significance level was rejected. This means that the average efficiency of the domestic private banks has increased significantly. The test results for domestic public banks show a t-statistic of -7.226 with 26 degrees of freedom. The two-tailed p-value is 0.001 which is less than the conventional 5 per cent or 1 per cent levels of significance. Therefore, we can again reject the null hypothesis at 5 per cent (or 1 per cent) significance level. This means that the average efficiency of the domestic public banks has improved significantly. The test results for foreign banks showed a t-statistic of -3.081 with 11 degrees of freedom. The two-tailed p-value is 0.012 which is more than the conventional 1 per cent levels of significance. This means that the average change in efficiency of the domestic foreign banks in Phase II as compared to Phase I was not statistically significant.
The results of the above analysis support the hypothesis that ownership of banks matters in so far as efficiency improvements are concerned. The presence of foreign banks led to more efficient domestic banks. The state-owned banks were found to be not significantly less efficient than their domestic private counterparts. The entry of relatively more efficient foreign banks 5 in Phase I together with improved prudential norms might have resulted in an environment that forced the entire banking sector to become more efficient. Moreover, within the bank groups, despite common ownership, bank specific efficiency levels vary and there have been wide variations in their respective abilities to maintain competitive efficiency. Banks supported by strong non-financial factors such as leadership, skills, market awareness, good Management Information Systems (MIS) and internal controls and effective strategy for countering competition have weathered the storm and exploited the opportunities provided by reforms and deregulation. 6 It can also be concluded that for public sector banks, fulfilment of social objectives and efficiency can and, in fact, need to go hand in hand.
Size and Efficiency
The need to study the relationship between bank size and efficiency becomes pertinent as there is an emphasis on consolidation in the banking sector all over the world. The relationship between size and efficiency in banks has been examined in this section. An attempt has also been made to find out whether the expansion of banking activity during post-reforms period had any implications for the efficiency of these banks.
To study the relationship between size and efficiency, the sample banks were classified into large and small banks on the basis of average deposits (2004–05 to 2006–07). Those banks that fell in the lower most quartile were taken as small banks and those banks that fell in the upper most quartile were taken as large banks. Panels A, B and C present the relationship between size and the ‘most’ and ‘least’ efficient banks, for domestic private sector banks, domestic public sector banks and foreign banks, respectively.
As may be observed from Panel A of Table 9, in the case of domestic private sector banks, four banks were categorised as ‘most efficient’ banks. They are listed in Panel A. Interestingly, all the four banks were small in size. On the other hand, none except one of the least efficient banks was small bank. This leads to the conclusion that ‘small is efficient’ in the case of private domestic banks. On the contrary, domestic public sector banks (Panel B) exhibit a different tendency. None of the most efficient banks in this category was small-sized whereas least efficient banks were small- or medium-sized with the exception of one bank (Canara Bank). Panel C on foreign banks offer no clue regarding relationship between efficiency and size. It may thus be concluded that relationship between size and efficiency is dependent upon ownership. On aggregate basis, the results do not offer any conclusive evidence of relationship between size and efficiency.
Relationship between Size and Efficiency
Listing on Stock Exchange and Efficiency
During the reform period, many public sector banks began to get listed on the stock exchange. They were allowed to diversify their activities and float issues in capital market to strengthen their capital base. There was a belief that broad basing of the ownership of public sector banks through listing leads to greater market discipline and improvement in their governance (Rakesh Mohan, 2007). The diversification of ownership may also lead to a qualitative difference in their functioning as there is induction of private shareholding as well as attendant issues of shareholder’s value as reflected by the market capitalisation, board representation and interests of minority shareholders (A. Reddy 2004).
Listing may also result in efficiency gains for domestic banks. An attempt has also been made in this section to study the impact of listing on the efficiency of these banks. Those public sector banks and private sector banks are identified that have gone in for listing during the period 1991–92 to 2006–07. Average efficiency of these banks before listing and after listing were calculated and compared. Analysis of impact of listing on efficiency scores shows that the average level of efficiency of both private banks and public sector banks have increased since the listing of the banks’ shares on the stock exchange.
The average efficiency for public banks increased from 90.5 per cent to 96 per cent after listing of their shares on the stock exchange and for private banks the two figures were 88.9 per cent and 95.4 per cent (Tables 10 and 11).
Tables 10 and 11 show that there was an improvement in the average efficiency of all banks after listing except for three banks in public sector and two banks in private sector. To determine whether this improvement in efficiency was statistically significant, the dependent samples t-test was calculated for the two series, before listing and after listing for both the bank groups namely, public sector and private sector. The dependent samples t-tests gave the following results.
Efficiency of Public Sector Banks before and after the Listing on Stock Exchange
For public sector banks
There was a significant difference between the average annual efficiency scores of private sector banks before listing and after listing on the stock exchange (t-statistic is -5.577, df = 20 and p value is 0.00) both at 5 per cent and 10 per cent level of significance.
For private sector banks
There was a significant difference between the average annual efficiency scores of private sector banks before listing and after listing on the stock exchange (t-statistic is -2.598, df = 9 and p value is 0.029) at 5 per cent level of significance. It may be concluded that the domestic Indian banks have experienced statistically significant listing gains as a consequence of the reform process.
Efficiency of Private Banks before and after the Listing on Stock Exchange
Profitability Test
While efficiency may not always translate into profitability due to factors such as asset quality and the impact of competitive pricing, it would be expected a priori that ‘most’ efficient banks are generally more profitable (Sturm and Williams 2005). As stated earlier, the sample banks belong to three different bank groups namely, domestic public sector banks (state-owned), domestic private sector banks and foreign anks. These bank groups have different ownership structures and hence are bound to have different characteristics. The present study could not obtain any conclusive result of the profitability test by taking all the banks together. Thus, further analyses were carried out according to bank group. Table 12 shows that the ‘most efficient’ group of public sector banks has an advantage over the ‘least efficient’ public sector banks in terms of higher Net Profit as percentage of Total Assets (NPTA) and Profit per Employee (PPE) in all the 3 years under study. As regards the expenses side, the ‘most efficient’ and ‘least efficient’ banks show similar average Operating Expenses as percentage of Total Expenses though it is marginally higher in the case of ‘least efficient’ banks in all the 3 years.
Profitability Test (2004–06): Public Sector Banks
This means that the ‘most efficient’ banks do not enjoy advantages in funding costs and therefore they are achieving efficiency by other means. Further, the average CAR of the ‘most efficient’ and ‘least efficient’ public sector banks is almost similar in all the 3 years and ranges from 12.07 to 12.55. This may be attributed the mandatory capital adequacy norms to be followed by all banks as stipulated by RBI. Another advantage that the ‘most efficient’ public sector banks seem to be having is lower Net NPA as percentage to Net Advances as compared to the ‘least efficient’ public sector banks.
Profitability Test (2004–06): Private Sector Banks
It may be observed from Table 13 that the ‘most efficient’ private sector banks enjoy the advantages of higher NPTA, Business per Employee and PPE in all the 3 years. The ‘least efficient’ private sector banks seem to be maintaining higher CARs as compared to the ‘most efficient’ private sector banks. The ‘least efficient’ banks in this category suffered from extremely high ratio of Operating Expenses as percentage of Total Expenses in the year 2004. In 2005 and 2006 however, the ratio of Operating Expenses as percentage of Total Expenses of the ‘least efficient’ private sector banks became almost similar to that of the ‘most efficient’ private sector banks. The ratio of Net NPA as percentage of Net Advances of the ‘least efficient’ private sector banks was higher than that of the ‘most efficient’ private sector banks in all the 3 years.
Table 14 shows that in the case of foreign banks, the ‘most efficient’ banks clearly have the characteristic of higher ratios of NPTA, Business per Employee and PPE as compared to the ‘least efficient’ banks in all the 3 years, 2004–06. The ‘least efficient’ foreign banks are characterised by higher ratios of Operating Expenses as percentage of Total Expenses and Net NPA as percentage to Net Advances in all the 3 years. The ‘most efficient’ foreign banks are characterised by higher CARs than the ‘least efficient’ foreign banks.
Profitability Test (2004–06): Foreign Banks
It may be concluded from the above discussed profitability test that the main differences between the ‘most efficient’ and ‘least efficient’ bank seems to be mainly related to profitability as represented by NPTA and PPE. ‘Most efficient’ banks are on an average characterised by higher profitability. Moreover, ‘least efficient’ banks always appear in the 3 years studied to have higher levels of non-performing assets. As there is not very clear and wide difference between the efficient banks and inefficient banks in any group in terms of Operating Expenses as percentage of Total Expenses, the efficient banks do not seem to be characterised by lower operational cost.
Summary and Conclusion
Banks in India differ widely in terms of ownership, regulation, goals, size and functioning. This article focuses on bank ownership and efficiency in commercial banks in post-reforms India. The article aims at answering the oft-repeated question, ‘why are some banks more efficient?’ The answer has its base in determinants of bank efficiency. Thus, this article investigated technical efficiency of Indian banking sector using DEA for the period 1991–92 to 2006–07 on a sample of 54 banks in terms of reforms, ownership type, size, listing on stock exchanges and profitability.
Overall, results seem to suggest that Indian banking experienced a clear improvement in technical efficiency over the period 1991–92 to 2006–07. A number of tests were employed to discover the determinants of bank efficiency in India. The investigation of the determinants of bank efficiency led to some very important results. First, over the period of study, it was found that average efficiency of domestic as well as foreign banks operating in India improved. Second, the study found evidence that reforms proved beneficial for the domestic banks and have helped them attain higher levels of efficiency comparable with their foreign counterparts. Thus, ownership of banks matters in so far as efficiency improvements are concerned. Third, the study also found evidence that relationship between size and efficiency is dependent upon ownership. On aggregate basis, however, results do not offer any conclusive evidence of relationship between size and efficiency. Fourth, domestic Indian banks have experienced statistically significant listing gains as a consequence of the reform process. Fifth, ‘most efficient’ banks were on an average characterised by higher profitability. Moreover, ‘least efficient’ banks always appear in the 3 years studied to have higher levels of non-performing assets. These results have several implications for bankers and policymakers such as, deregulation results in improved bank efficiency, state-owned banks can also pursue economic objectives besides social objectives efficiently, etc.
A few issues remain for further study. One such issue is cross-country comparison of efficiency in Indian banking sector, results of such a study would contribute considerably to the existing literature in this field. Liberalisation brought with it freedom for state-owned banks to raise funds from public. Although, this article studied the impact of listing on efficiency, it would be interesting to analyse the relationship between bank stock prices and bank efficiency. Another area which requires attention is the development of a framework to not only define but also assess appropriate prices of bank outputs. Existing literature is inadequate. The author intends to return to these issues in her future work.
Footnotes
Acknowledgements
The author wishes to express gratitude to Shubhashis Gangopadhyay (Editor-in-Chief, JoEMF), Bappaditya Mukhopadhyay (Managing Editor, JoEMF) and anonymous reviewer(s) for their analytical and insightful comments and suggestions. They identified gaps in the analysis and gave a new perspective to the entire study. My sincere thanks are also due to Prof. Tim Coelli, CEPA for development of DEAP software for efficiency measurements.