Elsevier

Journal of Banking & Finance

Volume 24, Issue 10, 1 October 2000, Pages 1605-1628
Journal of Banking & Finance

Efficiency and risk in Japanese banking

https://doi.org/10.1016/S0378-4266(99)00095-3Get rights and content

Abstract

This paper investigates the impact of risk and quality factors on banks’ cost by using the stochastic cost frontier methodology to evaluate scale and X-inefficiencies, as well as technical change for a sample of Japanese commercial banks between 1993 and 1996. Loan-loss provisions are included in the cost frontier model to control for output quality, with a financial capital and a liquidity ratio included to control risk. Following the approach suggested in Mester (1996) we show that if risk and quality factors are not taken into account optimal bank size tends to be overstated. That is, optimal bank size is considerably smaller when risk and quality factors are taken into account when modelling the cost characteristics of Japanese banks. We also find that the level of financial capital has the biggest influence on the scale efficiency estimates. X-inefficiency estimates, in contrast, appear less sensitive to risk and quality factors. Our results also suggest that scale inefficiencies dominate X-inefficiencies. These are important findings because they contrast with the results of previous studies on Japanese banking. In particular, the results indicate an alternative policy prescription, namely, that the largest banks should shrink to benefit from scale advantages. It also seems that financial capital has the largest influence on optimal bank size.

Introduction

Studies of the Japanese banking industry have typically found strong evidence of scale economies across a broad range of bank sizes (see for example, Kasuya, 1986; Yoshioka and Nakajima, 1987; Tachibanaki et al., 1991, Fukuyama, 1993; McKillop et al., 1996). The aforementioned studies provide an indication of the cost characteristics of the Japanese banking industry, however, they may be limited because they do not take into account the risks associated with banks' operations. This is a particularly relevant issue in Japanese banking given that the system has experienced substantial asset quality problems and low levels of capitalisation since the early 1990s Bank of Japan, 1995, Bank of Japan, 1996. Recent studies such as those Hughes and Mester, 1993, Hughes et al., 1995, McAllister and McManus, 1993, Mester, 1996 and Clark (1996) have drawn attention to the fact that bank efficiency studies typically ignore the impact of risk on banks' costs or profits and they suggest that risk characteristics need to be incorporated in the underlying industry cost or profit functions because, `unless quality and risk are controlled for, one might easily miscalculate a bank's level of inefficiency' (Mester, 1996, p. 1026). For instance, Hughes et al. (1995) in their study on 1989–90 US banks which exceed $1 billion in asset size, find that when they control for risk, inexhaustible economies of scale that increase with size are prevalent. When risk neutrality is imposed on the estimation, the large scale economies disappear and constant returns to scale are obtained. McAllister and McManus (1993) also show that for larger US banks estimates of scale economies are increased when they control for risk. (This tends to be more prevalent for banks up to $1 billion in asset size.) In contrast, Hughes and Mester (1993) find that estimates of increasing returns to scale for a wide range of bank sizes become constant returns when asset quality, financial capital and risk are taken into account. Clark (1996) also re-emphasises the influence of risk factors in the cost characteristics of US banks and shows how risk can statistically influence efficiency levels – in particular, he shows how X-inefficiency estimates tend to become smaller (about 3% from 9%) when risk factors are incorporated in frontier estimations.

In order to advance the aforementioned literature, this paper evaluates cost efficiency and technical change for Japanese commercial banks by comparing results obtained from two cost functions specifications. First we use the stochastic cost frontier methodology to estimate scale economies, scale efficiencies and X-inefficiency, as well as technical change, for a sample of Japanese commercial banks between 1993 and 1996 using a three input–three output Fourier-flexible cost function specification. The three outputs include total loans, total securities and total off-balance sheet items (nominal value). We then compare these results with those generated by a similar cost function which also includes variables controlling for risk and quality factors. In addition, we also evaluate the sensitivity of scale economy and X-inefficiency estimates to various risk and quality variables.

Our results show that when the underlying cost function specification does not control for risk and quality factors, scale economies are prevalent for all but the largest Japanese banks. Those larger than Yen 10,000 billion exhibit significant diseconomies. Overall, Japanese banks appear to be relatively scale efficient. When risk and quality factors are taken into account, however, only the very smallest banks exhibit significant economies with the majority of banks experiencing significant diseconomies of scale which appear to get larger with asset size. Estimates derived from the model including risk and quality variables suggested that there are widespread scale inefficiencies in the Japanese banking market. These results suggest that bank minimum efficient scale becomes smaller after controlling for risk, a finding similar to that of Hughes and Mester (1993). This contrasts with the results of Hughes et al. (1995) and McAllister and McManus (1993) who find either minimum efficient scale is increased or scale economies are never exhausted after controlling for risk. Our results also point to the limitations of previous studies on the cost characteristics of Japanese banking and also suggest different policy conclusions, namely, that the largest banks should reduce their size if they wish to benefit from scale economies. In contrast to the scale estimates and Clark’s (1996) findings on US banking, X-inefficiency scores appear to be less sensitive to the inclusion of risk and quality variables. The mean level of X-inefficiency for the largest banks range between 5% and 7%. This also indicates that the largest Japanese banks should focus on reducing managerial and other inefficiencies, as well as reducing their size if they are to generate cost savings. Finally, this study also finds that technical change has reduced bank cost, but at a declining rate, between 1993 and 1996.

Section snippets

Methodology

Following Mester, 1996, Cebenoyan et al., 1993 and Allen and Rai (1996) we use the stochastic cost frontier methodology. This approach labels a bank as inefficient if its costs are higher than those predicted for an efficient bank producing the same input/output configuration and the difference cannot be explained by statistical noise. The cost frontier is obtained by estimating a Fourier Flexible cost function with a composite error term, the sum of a two-sided error representing random

Data and results

Our data comprise the population of Japanese commercial banks listed in the London based IBCA bank credit rating agencies Bankscope (1997) database for the years 1993–1996, and consists of 139 banks for each year from 1993 to1995 and 136 in 1996.10 The median asset size of banks included was Yen 1.6 trillion and the average asset size was Yen 5 trillion (see Appendix A for more details). Table 1

Conclusion

Previous studies on the cost characteristics of Japanese banking have found strong evidence of scale economies across a wide range of bank sizes and even for the largest firms. This overall finding implies that substantial cost savings can be exploited through further expansion. These results, however, are limited because the approaches that have been taken to estimate cost economies do not take into account important asset quality and risk factors which influence bank inefficiency. This is

Acknowledgements

The authors wish to acknowledge the helpful comments of two anonymous referees.

References (44)

  • G. Lang et al.

    Efficiency and technical progress in banking: Empirical results for a panel of German cooperative banks

    Journal of Banking and Finance

    (1996)
  • P.H. McAllister et al.

    Resolving the scale efficiency puzzle in banking

    Journal of Banking and Finance

    (1993)
  • D.G. McKillop et al.

    The composite cost function and efficiency in giant Japanese banks

    Journal of Banking and Finance

    (1996)
  • L.J. Mester

    Efficiency in the savings and loan industry

    Journal of Banking and Finance

    (1993)
  • L.J. Mester

    A study of bank efficiency taking into account risk-preferences

    Journal of Banking and Finance

    (1996)
  • R.E. Stevenson

    Likelihood functions for generalised stochastic frontier estimation

    Journal of Econometrics

    (1980)
  • T. Tachibanaki et al.

    Economies of scope and shareholding of banks in Japan

    Journal of the Japanese and International Economies

    (1991)
  • Altunbas, Y., Molyneux, P., 1994. Sensitivity of stochastic frontier estimation to distributional assumptions: The case...
  • B.H. Baltagi

    Econometric Analysis of Panel Data

    (1995)
  • B.H. Baltagi et al.

    A general index of technical change

    Journal of Political Economy

    (1988)
  • Bank of Japan, 1995. Financial reports of Japanese financial institutions for fiscal year 1994. Quarterly Bulletin,...
  • Bank of Japan, 1996. Recent monetary and economic conditions and issues facing the Japanese financial institutions for...
  • Cited by (0)

    1

    The views expressed are those of the author and do not necessarily reflect those of the Federal Reserve Bank of New York or the Federal Reserve System.

    View full text