تاثیر رقابت در بازار سرمایه بر بانکداری رابطه ای: شواهدی از تجربه ژاپن

Disagreement exists about the potential effects of changes in competition on relationship lending. Boot and Thakor (2000) predict that an increase in capital market competition should lead to a reduction in relationship lending; however, Dinç (2000) predicts that greater capital market competition should increase relationship lending. Thus far, data limitations have precluded empirical tests of these competing hypotheses. In this study, we use a unique data set drawn from the deregulation of the Japanese financial system. Our findings show that increased capital market competition is associated with reduced relationship lending. However, the effect differs according to the maturity of the loans; increased capital market competition is associated with reduced long-term, but greater short-term, relationship lending.

Relationship banking is frequently the focus of recent literature on financial intermediation (Boot, 2000). This literature often discusses how banks can develop close relationships with their borrowers to overcome the fundamental problem of asymmetric information in the credit market. In recent years, however, the role of banks in the credit process has diminished due to the reduction in the marginal cost of information and in the ratio of private to public information. At the same time, dramatic increases in competition from the capital market have altered the relationship between banks and their borrowers. These changes raise a number of important questions about financial intermediation. This paper focuses on two primary questions: (i) do the bank's commitments to relationship lending diminish when competition from the capital market intensifies? and (ii) does the maturity of relationship lending (short-term versus long-term) matter? These two questions become increasingly important since significant changes have occurred in capital market competition ( Boot and Thakor, 2000). Moreover, existing literature offers conflicting predictions on the effects of this increased competition. Boot and Thakor (2000) predict that increases in capital market competition for bank loans will lead to a reduction in relationship lending. Their underlying premises are: (i) commercial banks will engage in both relationship and transaction lending; and (ii) these banks will substitute one for the other based upon a variety of factors including technology, competition, regulation, and relative profitability. While relationship lending is more valuable to the borrower than transaction lending, it is more expensive for the bank, since the bank must develop the sector-specialized expertise that is necessary to become a relationship lender. Following this argument, commercial banks may be expected to reduce their fund allocation to relationship lending and shift to transaction lending as they face increased competition from the capital market. Boot and Thakor (2000) also suggest that an increase in interbank competition will diminish the profitability of transaction lending more than relationship lending, thereby inducing banks to shift from transaction to relationship lending. In contrast, Dinç (2000) argues that a bank's commitment to relationship lending depends upon how such lending affects its reputation. Increased competition from the capital market will strengthen the bank's incentive to maintain a good reputation, thus leading to an increase in relationship lending. But Dinç agrees with Boot and Thakor (2000) in that increased interbank competition will not lead to a decline in relationship lending because it reduces the profitability of transaction lending more than that of relationship lending. Several studies also focus on the link between the degree of competition and the importance of relationship lending. For example, Yafeh and Yosha (2001) argue that an increase in competition leads to more relationship loans for multi-product banks. Dell'Ariccia and Marquez (2003) argue that banks may substitute relationship loans for transaction loans as a response to more interbank competition. Degryse and Ongena (2005) use Belgian data to examine the effects of greater interbank competition on relationship loans, even though the effects of capital market competition is not the focus their study. While our paper focuses primarily on the effects of changes in capital market competition, we also consider the effects of change in interbank competition. The deregulation of the Japanese financial system provides an excellent laboratory to test the competing predictions of Boot and Thakor (2000) and Dinç (2000); both predictions focus on the effects of increased capital market competition on relationship lending. During the 25-year time frame of our analysis, the extent of the capital market competition has increased dramatically. These changes allow us to measure the impact of increased competition on relationship lending and provide empirical data by which we can reasonably isolate the differential effects of increased capital market competition from the effects of increased interbank competition. Our 25-year study period is divided into four subperiods based upon the degree of capital market competition. These subperiods are primarily based upon the evolving deregulation of the Japanese financial markets. Period 1 (1978–1984) signifies a period of bank domination; Period 2 (1985–1990) is characterized by expanding capital market financing; Period 3 (1991–1997) represents a period of a more openly competitive financial system that has restrictions; and Period 4 (1998–2002) characterizes an almost complete abolishment of the barriers to the capital market. Our findings show that increased capital market competition is associated with reduced relationship lending. However, the effect differs according to the maturity of the loans. In general, increased capital market competition is associated with reduced long-term, but greater short-term, relationship lending. This suggests that commercial banks tend to switch from long- to short-term lending in response to greater competition from the capital market. This finding is not surprising given that the deregulation of the Japanese financial sector provides corporate borrowers with increased access to bond market financing. We also find that increases in interbank competition, following intensifying capital market competition, are generally associated with increases in short-term relationship lending. The remainder of this study is organized in the following manner. Section 2 discusses relationship banking in greater detail. Section 3 describes how the Japanese financial system evolved over the four subperiods and introduces our hypothesis. This section also details the empirical model, the sample selection, and other variables of interest. Section 4 reports our empirical findings and the results of robustness checks. Section 5 concludes the paper.

The predictions made by Boot and Thakor (2000) and Dinç (2000) on the effects of greater capital market competition on relationship lending have not been thoroughly investigated in the past because of data limitations and the lack of an appropriate empirical setting. However, the deregulation of the Japanese capital market in the 1980s and 1990s provides a unique empirical setting to test these predictions. Furthermore, the information dominance of and implicit insurance rendered by the main banks in Japan matches the profile of relationship lending as documented by previous studies; this allows us to relatively easily identify the loan exposure associated with relationship lending. Consistent with the prediction by Boot and Thakor (2000), but contrary to Dinç (2000), we have observed that increased capital market competition is associated with reduced relationship lending. However, the effect varies depending on the loan maturities. Increased capital market competition is, in general, associated with reduced long-term relationship loans but greater short-term relationship lending. These results suggest that commercial banks tend to switch from long- to short-term lending in response to greater competition from the capital market. This finding is not surprising given that the deregulation of the Japanese financial sector provides corporate borrowers with increased access to bond market financing. We also find that increases in interbank competition following intensifying capital market competition lead to an increase in short-term relationship lending. We attribute this result to the nature of relationship lending and to the difficulty in detecting sector specialization, which is the critical link between relationship lending and interbank competition. We suspect that even though the main banks possess proprietary client firm information and enjoy the benefits of other business-related banking arrangements, they still require a certain level of monitoring activities to deter the possible opportunistic behavior of the client firms.