The Risk Odium of Banks in Emerging Credit Markets: Reference to India

Dr. Vijay Pithadia  

Using bank-level data from India, for six years (1995-96 to 2000-01), we examine banks’ behavior in the context of emerging credit markets. Our results indicate that the credit market behavior of banks in emerging markets is largely determined by past trends. We also find evidence to support the hypothesis that prudential regulations have a significant impact on the banks’ behavior with respect to credit disbursal. Finally, we find evidence that suggest that credit expansion by banks in emerging markets maybe significantly constrained by the inability of the banks to reduce the liquidity risk associated with non-securities loans.

Keywords: Indian banking, Development, Credit-to-deposit ratio, Risk Odium

1. Introduction

Much of the literature on banking in emerging markets focuses on either the broad relationship between ownership and financial performance (e.g., Sarkar, Sarkar and Bhaumik, 1998) or the agency aspect of ownership, i.e., the impact of separation between management and ownership on the performance of banks (e.g., Gorton and Schmid, 1999; Hirshey, 1999). The focus on the relationship between ownership and financial performance of banks in emerging markets stems from concern about both the possibility of inefficient allocation of scarce financial resources in the presence of dominant public sector banks that often manifest McKinnon-Shaw type financial repression, and also from the concern about the possible fiscal impact of banking sector fragility in an environment where directed credit, political patronage, and severe moral hazard on the part of public sector bank officials can lead to significant accumulation of non-performing assets (NPAs).

While the focus on ownership is not completely unjustified in the context of banks in emerging markets, it has drawn attention away from the fact that, unlike a manufacturing or services sector firm, a bank helps mobilize domestic savings for subsequent investment in various on-going and new projects, and thereby is also the conduit for the transmission of monetary policy, and the facilitator of economic growth. Indeed, it is now stylized in the literature that the intermediary role of banks plays an important role in encouraging growth, even though in some countries a well-functioning credit market has added the unwelcome effect of increasing debt accumulation rather than improving total factor productivity (Gertler and Gilchrist, 1993; Ketkar, 1993; Ma and Smith, 1996; Bulir, 1998; Acemoglu, 2001; Bell and Rousseau, 2001; Da Rin and Hellman, 2002; Jeong, Kymn and Kymn, 2003).

Thus, not only are allocating efficiency and financial performance of banks important, but so also is the amount of credit disbursed by them. The fallacy of analysis that emphasizes bank ownership, with the prior that private ownership is better than public ownership, is evident. While private ownership may improve allocate efficiency in the credit market, at least so long as the market is not subjected to financial repression, there is evidence to suggest that it may be detrimental to credit disbursal, if the risks associated with this are significantly high.

The main problem in extending the Banerjee and Duflo approach to a larger proportion of the banking sector in India, or indeed any other developing country, is that it requires data on all credit related transactions of the banks. Indeed, the data used by Banerjee and Duflo itself is very limited, restricted to the activities of one Indian public sector bank that accounted for about 5 percent of banking sector assets when the analysis was undertaken. Therefore, in this paper, we propose to address this important issue Centre at London Business Schoolwith respect to collection of data. The authors remain responsible for all remaining errors.

Banks in India have the choice of investing resources in safe government bonds, or risky credit instruments. Ceteris paribus, a bank has to choose the allocation of resources between the risks less and risky assets, and this choice is manifested in the credit-to-deposit ratio (CDR). We then define and estimate a model that expresses CDR as a function of the credit risk associated with the banks’ potential borrower pool, and the risk averseness of the banks. As in the literature (Banerjee and Duflo, 2001), we also use a lagged dependent variable in the specification, to allow for persistence in the CDR.

The estimation uses a random effects model that allows us to capture the impact of ownership on the CDR. The robustness of our results is verified by measuring CDR in two different ways, and by using two different samples, firstly the domestic banks and second domestic and foreign. The choice of India is justified as the Indian banking sector has a multiple ownership structure, comprising public sector banks, incumbent and de novo private domestic sector banks, and foreign banks. In addition, significant reforms and liberalization has taken place since the early 1990s (Sarkar, Sarkar and Bhaumik, 1998, Shirai and Rajsekaran, 2001; Bhaumik and Mukherjee, 2002), thereby granting all banks effective operational autonomy.

Recent literature has found evidence of convergence among these different types of Indian banks in terms of financial performance (Bhaumik and Dimova, 2004), indicating that they have taken advantage of the reforms to compete with each other, and learn from each other sufficiently to be able to invade each other’s market niches. However, the existing literature does not indicate whether there has also been a convergence in the credit market behavior of the different types of banks, with respect to credit disbursal. At the same time, despite a large market capitalization by developing country standards, banks remain the main source of capital for most micro, small and medium enterprises. Hence, Indian banking provides an ideal setting for further analysis.

Our analysis confirms that of Banerjee and Duflo (2001), suggesting there is a strong persistence in the CDR, whether due to the relationship nature of the business, or the inability of a large proportion of the banks to assess credit risk associated with individual loan applications effectively. The descriptive statistics and regression results illustrate that the de novo private banks, which have been the key driver of competition in the Indian banking industry since liberalization in the early 1990s, lend less in the form of non-securities debt than the other domestic banks. However, if securities debt is taken into consideration, there are no observable differences in the credit market activity of public sector banks, old private banks and the de novo private banks.

2. Modeling bank behavior in credit markets

A bank is a multi-product firm, with a portfolio consisting of non-securities loans, as well as securities issued by non-government entities and federal, state and local governments. In addition, a bank generates revenues from fee-based contracts and speculation/participation in the market for off-balance sheet items. In developing countries, the choices facing the banks are usually fairly limited, partly because of government regulations, but also because of missing or underdeveloped markets for assets and instruments such as equity and financial derivatives. For example, in India, equities accounted for less that 1 percent of the bank assets in both 1996-97 and 2000-01. At the same time, while states, regions and local bodies in developing countries have different degrees of credit worthiness, the political economy of most of these countries ensure that all government securities carry the implicit or explicit guarantee by the federal government. That is, it is possible to think of banks in developing countries having two broad choices; they can either invest their resources, net of the cash reserve ratio and other regulatory caveats, into safe government securities, or disburse them as credit to the non-government sector, where all such credit is inherently more risky.

The risk averseness of a bank can arise from two different sources. Firstly, a bank may be innately risk averse, but may also be reluctant to take risk on account of factors such as, the impact of past behavior with respect to credit decisions. In India, the degree of innate risk averseness bank is not difficult to measure, and initially, it can be argued that banks with different ownership patterns (OWNERSHIP) have different levels of innate risk averseness. However, it is difficult to predict a priori the exact relationship between ownership and risk averseness. For example, in principle, it can be argued that a foreign bank may be more risk averse than a domestic bank due to less knowledge of local credit markets and fewer informal options with respect to enforcing contracts. On the other hand, it can also be argued that the Indian assets account for a very small proportion of the overall asset base and therefore a foreign bank would be willing to take risk to capture market share.

The second measure of innate risk averseness in banks is likely to have a predictable relationship with the choice of CDR. All banks in India are required by the Reserve Bank of India (RBI) to maintain 25 percent of deposits in the form of safe and liquid assets, mostly in the form of government securities. However, since the mid 1990s, most banks have voluntarily invested much more than 25 percent of their assets in government securities, behavior that in Indian policy circles as “lazy banking”. The rationale for lazy banking is the risk associated with credit disbursal in a developing country with attendant economic cycles and underdeveloped legal institutions to enforce contracts, and also awareness of the responsible banks that they may not have the necessary expertise to screen potential borrowers. Thus, lazy banking is a manifestation of risk averseness. Therefore the ratio of banks’ exposure to government securities, as a percentage of deposits, in excess of the required 25 percent, to the median exposure of all the banks in the sample, is used as a measure of risk averseness (ExGov securities).

The RBI also requires banks to reserve a stipulated minimum share of disbursed credit for the priority sector, which is comprised largely of agriculture and small firms. Banerjee, Cole and Duflo (2003) have noted that the average risk associated with priority sector lending is high. Data suggests that in any suppose that a bank has invested 32 percent of its deposits into government securities. In that case, its excess holding of such securities over and above that required by the RBI is 7 percentage points. In the former case, the risk averseness of the bank in period t is likely to decline while, in the latter case, its risk averseness in that period is likely to increase. As with investment in government securities, a proxy for this risk aspect of aversion is the ratio of a bank’s distance from the RBI mandated lower limit for priority sector exposure to the median distance of all the banks in the sample (PRIORITY). Risk averseness would increase or decrease with this measure depending upon the effectiveness with which the RBI enforces priority sector lending requirements. As above, possible endogenously is avoided by using a lagged value in the estimation.

3. Data

The model has been estimated using data obtained from the Indian Banks’ Association. The empirical analysis involves the use of data from six financial years: 1995-96 through 2000-01. However, the use of lagged values in the specification results in the use of data from only years 1996-97 through 2000-01 for the regression analysis. The data suggests that although there were 36 foreign banks for all banks is 10 percent, then our measure of risk averseness for this bank is 0.7. Foreign banks with less than two branches were removed as these were considered to locate in Indiato trade credit and services related to cross-border transactions and were not involved in the credit market. The final sample is comprised of 27 public sector banks (PUBLIC), 24 incumbent domestic private sector banks (OLDPRIVATE) which had been in operation prior to liberalization of the banking sector, 8 de novo domestic private sector banks which started operation after liberalization (NEWPRIVATE), and 12 foreign banks (FOREIGN). Together, they account for approximately 98 percent of the deposits and assets of the Indian banking sector.

The ratio of non-securities advances to deposits (CDR1), and the ratio of the sum of non-securities and securities loans to deposits (CDR2). The descriptive statistics indicate the following: (a) the CDR of the foreign banks are noticeably higher than those of the domestic banks, and (b) over time, there has been an increase in the share of securities credit in overall loan commitment of all types of banks, although the increase is most Noticeable for the de novo private banks and the foreign banks an explanation for the former is that foreign banks often make loans within India, using deposits raised abroad, which means that the credit disbursed in India is high as a proportion of deposits collected in India.

The data on the exposure of the banks to government securities suggests that all types of banks, and especially the public sector banks, buy government securities over and above the Statutory Liquidity Ratio (SLR) requirement. Further, even after correcting for inflation of an average of between 6 and 7 Percent per year, there was a significant growth in the additional exposure in all types of Indian banks to these sovereign securities. The data also indicates that the average Indian bank does not meet the regulatory obligation with respect to priority sector lending during the period of analysis. The descriptive statistics suggest that this deviation rose over time for all types of banks, even after accounting for inflation. While the rise is not nearly as dramatic in percentage terms as the rise in bank exposure to government securities, nevertheless it suggests that the banks do not face the threat of punitive action from the regulators if they do not meet the priority sector target.

Finally, four different measures of NPA are reported, reflecting those disclosed by the RBI; the ratio of gross NPA to total assets (NPA1), the ratio of net NPA to total assets (NPA2), the ratio of gross NPA to gross assets (NPA3) and the ratio of net NPA to net assets (NPA4). Not surprisingly, the public sector banks had more NPA on their balance sheets in 1996-97, but they were able to reduce this considerably over time, even though there was an increase in NPAs in all other ownership groups.

4. Results

The regression results, based on specification [3b] are in Tables 2 and 3. Both tables report the coefficient estimates, with columns 1-4 showing the sample of domestic banks, and columns 5-8 all banks. Both specifications were estimated because unlike the domestic banks, foreign banks can use deposits from outside India. In addition, they raise a substantial part of their resources from the money market. Hence it is not obvious that they should be considered in the same sample. Further, given that the lagged dependent variable and NPA are strongly significant at the 1% level for almost all specifications, and given that ownership is appears to be important, the high R are unlikely to be due to multi collinearity.

The results indicate that, firstly, there is a strong persistence in the CDR of Indian banks on average, which is consistent with the findings of Banerjee and Duflo (2001). As noted above, this could simply be a consequence of the nature of banking which is a relationship-based activity but could also indicate that Indian banks are not skilled at evaluating the credit worthiness of potential debtors, thereby rewarding the moribund yet stable businesses at the expense of new or dynamic enterprises that have an expected flow of income that is more volatile. Secondly, NPAs have a significant impact on the ability of a bank to take risk by opting for more loans and less investment in government securities. Specifically, rather than take more risk in order to become too-big-to-fail, they become more risk averse and restrict lending. This is consistent with the ownership pattern of banks in Indiawhere about 85 percent of bank assets are either with public sector banks that cannot fail by definition, or with incumbent private sector banks that are closely held, and likely to have risk adverse managers making active decisions on credit assessment.

This result which indicates that the de novo banks are disbursing an increasingly larger share of credit in the form of securities loans, and this is robust across choice of samples and specifications. As noted above, this implies that profit seeking dynamic banks view interest rate risk associated with securitization of their loan based assets as a relatively smaller problem than the liquidity risk associated with corporate assets that are used as collateral for non-securities loans, and this implication is completely consistent with the evidence on the high cost associated with bankruptcy and liquidation in India (e.g., Enron/Dabhol Power Project). The move to securitization of debt has been further facilitated by the fact that all corporate securities that are purchased by the bank are scrutinized by credit rating agencies, thereby reducing the need for the banks to develop in-house capabilities to assess credit risk.

The implications for this result are twofold. Firstly, it suggests that competition, greater accountability in the sale and enlistment of shares on the stock exchanges, and hard budget constraints drive incumbent domestic banks towards greater profit orientation. In addition, as there is a change in the ownership profile of Indian banks, due to mergers and acquisitions and a gradual program of privatization, disbursal of credit by banks to businesses would significantly depend on the state of the market for corporate bonds. However, recent studies indicate that despite some measures by the government, the Securities and Exchange Board of India (SEBI) and the RBI, the Indian bond market is characterized by high liquidity risk, prices that are inconsistent with the term structure of interest rates and corporate credit ratings, and generally a slow rate of spread of information (see Bhaumik, Bose and Coondoo, 2003).

5. Concluding Remarks

This paper focuses on the behavior of banks operating in emerging markets where they have the choice of disbursing resources collected through deposits either as credit to commercial borrowers or as investment into sovereign securities. Commercial credit in these countries carry credit risk as well as liquidity risk, given the high cost of liquidating collateral for such credit, and the absence of markets for hedging these risks. Sovereign securities, on the other hand, carry near zero default risk, and a relatively low level of liquidity risk.

It is now established in the literature that the Indian banking sector has witnessed a rapid increase in competition, and that even public sector banks in India are significantly driven by the profit maximizing motive. This trend is likely to be exacerbated by changes in ownership of public sector banks and thus far closely held private sector banks, by disinvestment by government in the former and acquisition of the latter by foreign and de novo private banks, such as, ING’s takeover of Vaishya Bank. The results reported in this paper suggest that profit conscious banks are likely to restrict disbursal of commercial credit in the face of prudential regulations, and in the absence of institutions that allow minimization of credit and liquidity risk associated with commercial loans. Clearly, competition and privatization are important aspects of policy for banks in emerging markets. But, in the absence of supporting institutions, such as, a bankruptcy-liquidity code, well functioning markets for corporate debt instruments, markets for hedging credit risk, and personnel with adequate credit risk assessment skills, these policies may also hinder deepening and widening of the credit market, thereby adversely affecting economic growth.


Acemoglu, Daron, 2001. Credit market imperfections and persistent unemployment, European Economic Review 45(4-6), 665-79
Banerjee, Abhijit. V. and Esther Duflo, 2001 The nature of credit constraints: Evidence from an Indian bank, Mimeo, Massachusetts Institute of Technology, Cambridge, Massachusetts
Banerjee, Abhijit, Shawn Cole and Esther Duflo, 2003 Bank financing in India, Mimeo, Massachusetts Institute of Technology, Cambridge, Massachusetts
Bell, Clive and Peter L. Rousseau, 2001 Post-independence India: a case of finance-led industrialization? Journal of Development Economics 65(1), 153-75
Bhaumik, Sumon K and Paramita Mukherjee 2002 The Indian banking sector: A commentary, In: Parthasarathi Banerjee and Frank-Jurgen Richter (eds.) Economic Institutions in India: Sustainability under Liberalization and Globalization, Pal grave Macmillan, London, pp. 71-98
Bhaumik, Sumon K, Suchismita Bose and Dipankor Coondoo, 2003 the emerging Indian bond market: Evolution, problems and prospects. Journal of Global financial Markets, spring, 33-43
Bhaumik, Sumon K and Ralitza D Dimova, 2004 how important is ownership in a market with a level playing field? The Indian banking sector revisited, Journal of Comparative Economics 32(1), 165-180
Bulir, Ales, 1998. Business cycle in Czechoslovakiaunder central planning: Were credit shocks causing it? Journal of Comparative Economics 26(2), 226-45
Clarke, George, Robert Cull, Laura D’Amato and Andrea Molinari, 1999 The effect of foreign entry on Argentina’s domestic banking sector, Policy Research  Working Paper 2158, The World Bank, Washington, D.C.
Clarke, George, Robert Cull and Maria S. M. Peria, 2001 Does foreign bank penetration reduce access to credit in developing countries? Evidence from asking borrowers, Mimeo, Development Research Group, The World Bank, Washington, D.C
Clarke, George, Robert Cull and Maria S. M. Peria and Susana M. Sanchez, 2002 Bank lending to small businesses in Latin America: Does bank origin matter? Mimeo, the World Bank, Washington, D.C
Gertler Mark and Simon Gilchrist, 1993 the cyclical behavior of short-term business lending: Implications for financial propagation mechanisms, European Economic Review 37(2-3), 623-31
Gorton, Gary and Frank Schmid, 1999 corporate governance, ownership dispersion and efficiency: Empirical evidence from Austrian cooperative banking, Journal of Corporate Finance 5, 119-40
Hirshey, Mark, 1999. Managerial equity ownership and bank performance, Economic Letters 64, 209-13
Indian Banks’ Association, 1995 Indian Banking Yearbook
Indian Banks’ Association, 1995-2001. Performance Highlights of Banks International Monetary Fund, 2000, Bulgaria: Selection issues and statistical appendix. IMF Staff Country Report, 00/54
Jeong, Woocheon, Kem O. Kymn and Christine J. Kymn, 2003 the long-run interdependence of bank- health, investment-oriented bank loans, and economic performance: A time-series analysis, The Quarterly Review of Economics and Finance 43(1), 11-30
Ketkar, Kusum, 1993. Public sector banking, efficiency and economic growth in India, World Development 21(10), 1685-97
Ma, Chien-Hui and Bruce D Smith, 1996 Credit market imperfections and economic development: Theory and evidence, Journal of Development Economics 48(2), 351-87
Randall, Richard E. (Ed.), 1993. Safeguarding the Banking System in an Environment of financial Cycles, Federal Reserve Bank of Boston
Reserve Bank of India, Trends and Progress of Banking in India, various issues
Sarkar, Jayati, Subrata Sarkar and Sumon K. Bhaumik, 1998 Does ownership always matter? Evidence from the Indian banking industry Journal of Comparative Economics 26, 262-281
Shirai, Sayuri and Prithipal Rajsekaran, 2001 Is India’s banking sector reform successful from the perspective of the governance of the banking system? Mimeo, Economic and Social Commission for Asia and the Pacific