RIETI Policy Symposium

Japan's Financial System: Revisiting the Relationship between Corporations and Financial Institutions

Information

  • Time and Date:
    Thursday, February 16, 2006, 13:30-17:10
    Friday, February 17, 2006, 9:00-17:00
  • Venue:
    Hall, Shinsei Bank (1st Floor, Head Office, 2-1-8 Uchisaiwaicho, Chiyoda-ku, Tokyo)
  • Language:
    Japanese / English (with simultaneous interpretation)

Summary of Proceedings

Session 1: The Workings of Relationship Banking - impact on capital procurement and corporate activity

Session Chair: UESUGI Iichiro (Fellow, RIETI)

Presentation 1: WATANABE Tsutomu (Faculty Fellow, RIETI / Professor, Institute of Economic Research, Hitotsubashi University)
"Relationship between interest rate and firm age"

Presentation 2: UCHIDA Hirofumi (Associate Professor of Banking and Finance, Faculty of Economics, Wakayama University) and Gregory F. UDELL (Bank One Chair of Banking and Finance, Kelley School of Business, Indiana University)
"Factors that influence the relationship between banks and SMEs"

Presentation 3: OGAWA Kazuo (Professor, Institute of Social and Economic Research, Osaka University)
"Does main bank's health matter in corporate customers' behavior?"

Comments: HORIUCHI Akiyoshi (Professor, Faculty of Policy Studies, Chuo University)

Session Summary

Three presentations were made during the session from the critical perspectives set out below. Each of the presentations examined comprehensively, based on microdata, the question of the impact of relationships with financial institutions on the management environment for companies and their corporate behavior, and of whether the manner in which that impact is made has changed since the 1990s.

  • In corporate life-cycles, how does the environment for borrowing from financial institutions change, and is that change deemed rational?
  • In what cases and what environments do firms benefit from relationship banking?
  • To what degree does the health of financial institutions, through the relationships between firms and financial institutions, affect corporate activity?

Outline: Professor WATANABE Tsutomu presentation

Prof. Watanabe's presentation was on the subject of the "Relationship between interest rate and firm age," setting out the following points.

The changes in the interest rates charged on loans to firms in relation to their age can be divided into two channels: "selection," in which changes in interest rates are caused by firms of lower quality exiting from markets, and "adaptation," in which changes in interest rates are caused by changes in the behavior of surviving firms or their surrounding agents. For changes in interest rates accompanying the aging of Japanese SMEs, the impact of adaptation is greater than that of selection.

With respect to selection, it has been pointed out that among large Japanese firms there is "unnatural selection" in which the pricing is such that loans at low interest rates are extended to firms of low quality, and their resultant continued clogging of markets reduces the overall efficiency of the economy. Nevertheless, in our analysis we found that among SMEs, loan pricing sets high interest rates for low-quality firms, leading to natural selection in which they exit from markets.

In regard to adaptation, the reason for declines in the interest rates for surviving firms in markets is not because the firms become larger, but because their reputation is enhanced by a history of having continued in existence without going bankrupt.

Prof. Horiuchi put forward the following arguments concerning the Prof. Watanabe's presentation.

In the paper it was indicated that, with respect to selection, low-quality firms were charged high interest rates and exited from markets. However, it may be that clear consideration was not given to the existence of mistakes in financial intermediation in which promising firms have high interest rates imposed on them and have no alternative but to exit from markets.

The existence of collateral and guarantees is regarded as having a major impact on changes in interest rates, but this does not appear to have been considered explicitly in this paper.

The following response was made to the above arguments.

The performance of firms is poor prior to their exit from markets. Therefore I believe that, on average, mistakes in interest-rate pricing of the kind you have mentioned did not exist.

In the CRD data that we used there is no information on collateral and guarantees. However, by taking differences in time series directions through the use of panel data over a period of six years, we endeavored to exclude effects characteristic of individual firms, such as collateral and guarantees.

Outline: Professors UCHIDA Hirofumi and Gregory F. UDELL presentation

Profs. Uchida and Udell's presentation was on the subject of "Factors that influence the relationship between banks and SMEs," and set out the following points.

The longer business relationships with banks are, the more interest rates decline, and the longer and wider the business relationships with banks, the more the probability of loan turndown falls. However, this advantage was seen only in cases of unaudited SMEs borrowing from shinkin banks in uncompetitive markets. For other SMEs, no such benefits from bank-borrower relationships were seen.

Owing to data limitations in empirical analyses of relationship banking conducted up until now, it has not been possible to distinguish between relationship lending and financial statement lending. In this study we were able to distinguish between relationship lending and financial statement lending, and our biggest contribution has been to demonstrate that financial statement lending does in reality differ from relationship lending.

There are various methods for lending to SMEs in addition to relationship lending. Whether, among these, relationship lending is an effective technology or not depends on the market's competitive environment and the characteristics of financial institutions.

Prof. Horiuchi put forward the following argument concerning Profs. Uchida and Udell's presentation.

I believe there is a mutual relationship between interest rates and collateral/guarantees. In this paper, interest rates and collateral are explained separately, but it should examine how interest rates and collateral relate to each other.

The following response was made to the above argument.

With respect to the simultaneous determinacy of interest rates and collateral, by using instrumental variables we intend to estimate this in a way that takes endogeneity into consideration. In estimates in which interest rates are the explained variables we are making estimates that take the simultaneous determinacy of collateral into consideration, but basically there has been little difference in the results. We are now studying the making of estimates in which collateral is the explained variable.

The following question was posed from the floor.

This debate is based on the premise that relationships characterized by the length and width of business transactions with banks influence the degree of benefit that can be obtained for interest rates and collateral, but is it really true to say that this shows a causal relationship?

The following response was made to this question.

In view of the predetermined nature of the variables, I believe that there is no problem. As regards simultaneous determinacy and the converse causal relationship, those should perhaps be examined, though in terms of economic theory it is difficult to imagine that those possibilities exist. I would add that the methods employed in our analysis are well-established in this field, so I believe that their use does not present any particular problem.

Outline: Professor OGAWA Kazuo presentation

Prof. Ogawa's presentation was on the subject of "Does main bank's health matter in corporate customers' behavior," setting out the following points.

If a main bank's bad loan ratio rises and its health deteriorates, it becomes very difficult to borrow money from the bank through its lending channel. This reduces the firm's capital investment and employment, and it also leads to a decline in the liquidity in hand as the firm uses such funds as a substitute for bank borrowing.

A rise in a main bank's bad loan ratio and the deterioration of its health also makes it impossible to expect long-term commitments from it through channels in which main-bank functions other than lending are conducted, such as payment settlement, the supply of information, and insurance. This reduces the firm's capital investment and employment and leads to an increase in liquidity for precautionary purposes.

Prof. Horiuchi put forward the following argument concerning Prof. Ogawa's presentation.

Doubt remains as to whether settlement functions and other financial services other than lending are services that necessarily depend on bank-firm relationships.

The following response was made to the above argument.

It is certainly the case that in this analysis such non-lending functions as settlement and information-provision are not made full use of, but long-run decision-making such as those on capital investment and employment almost certainly depend on long-term commitments with main banks. We believe that the fact that the deterioration of the health of a main bank affects long-run decision-making of this kind is because of the disappearance of expectations with respect to main bank functions other than lending, such as insurance functions.

Dr. Berger posed the following questions concerning all of the presentations.

In Profs. Uchida and Udell's presentation it was shown that relationship banking was beneficial only in the very limited circumstances of small banks not exposed to competition. In Prof. Ogawa's presentation, however, it was shown that Japanese SMEs establish very strong relationships with main banks, and also engage in wide-ranging transactions with major banks. If the assertion in Profs. Uchida and Udell's presentation is correct, if companies that engage in wide-ranging transactions with those major banks do not benefit, why do they build bank relationships? Is it because the data used in the two analyses differ?

The following response was made by Prof. Ogawa to the above question.

Our analysis was limited to private-sector financial institutions (i.e. those excluding public financial institutions) in the case of main banks, and in the case of firms to those for which it was possible to use three years of balance sheet information. In this sense there may have been a slight element of sample selection bias.

A supplementary remark was made by the session chair concerning the response.

The benefits of business relationships with large banks include the possibility of borrowing in large lots, being able to borrow at lower interest rates than from other types of institution, and enhancing creditworthiness by doing business with large banks. Benefits such as these may not depend on the number of years a relationship has existed.

(This report was compiled by SAKAI Koji)