جنبه های خرد سیاست های پولی: آخرین راهکار وام دهنده و انتخاب بانک ها قبل از جنگ ژاپن

This paper explores how the Bank of Japan (BOJ) dealt with the trade-off between stability of the financial system and the moral hazard of banks in pre-war Japan. The BOJ concentrated Lender of Last Resort (LLR) loans with those banks that had an established transaction relationship with the BOJ. At the same time, the BOJ carefully selected its transaction counterparts, and did not hesitate to end the relationship if the performance of a counterpart declined. Further, the BOJ was selective in providing LLR loans. Through this policy, the BOJ could avoid the moral hazard that the LLR policy might otherwise have incurred.

Since Bagehot (1873), the central banks of many countries have come to adopt the role of Lender of Last Resort (LLR), and we have a rich store of theoretical and empirical literature on LLR (Goodhart, 1985, Miron, 1986, Bordo, 1990 and Goodhart and Huang, 2005 , among others). According to the “classical view” of the LLR, the Central Bank should prevent illiquid but solvent banks from failing by lending money at a penalty rate (Bordo, 1990, p. 19). That LLR lending has been effective in preventing bank panics is well established (Bordo, 1990, Butkiewicz, 1995 and Miron, 1986). However, as Goodhart (1985) argues, it is difficult for central banks to distinguish between solvent and insolvent banks. Therefore, the bank as the LLR is faced with a trade-off between the stability of the financial system and moral hazard (Cordella and Yayati, 2003). Drawing on the experience of the United States and Europe, Bordo (1990, p. 9) states that “[a]ssistance to insolvent banks was the exception rather than the rule until the 1970s… [t]he monetary authority in earlier times erred on the side of deficiency rather than excess.” However, not so much is known about how the central banks have dealt with the trade-off. In this paper, we address this issue focusing on the Bank of Japan (BOJ) in the period before the Second World War. The Japanese financial system became unstable in the 1920s and the BOJ actively played the role of LLR. We explore how the BOJ selected the banks to be bailed out and what implications the BOJ’s policy had. In the literature on Japanese financial history, Ehiro, 1920 and Ito, 2003 review the role of the BOJ as the LLR during this period. As they point out, the LLR loan by the BOJ was a major policy tool for stabilizing the financial system of the 1920s. In another strand of the literature, Yabushita and Inoue (1993) found that the probability of bank closure during the financial crisis of 1927 was negatively correlated with the profitability and the ratio of risky assets of a bank. Okazaki, 2002 and Okazaki et al., 2005 confirmed this result using data regarding a wider range of bank exits. Yabushita and Inoue concluded that bank closures during the financial crisis of 1927 were not contagious.1 In the context of this paper, this interpretation suggests that the LLR loans by the BOJ successfully prevented financial crises from becoming contagious. Further, the negative correlation between bank performance and bank closure suggests that the LLR loans did not impair the selection mechanism of the market, by bailing out insolvent banks. In other words, it seems that the BOJ could deal with the above trade-off reasonably well. In order to understand how this occurred, it should be noted that the BOJ was selective in its provision of LLR loans, and that LLR loans were crucial for banks (Ishii, 1980). Ishii (1980) indicated that those banks which already had transaction relationships with the BOJ were the main recipients of LLR loans, and that those banks, for the most part, were large-sized ones.2 Referring to this fact, this paper will examine how the BOJ selected its transaction counterparts, using internal documents from the BOJ and bank-level quantitative data. The BOJ archives hold the original documents on the individual openings and closings of transaction relationships with private banks in the pre-war period. I look at how the BOJ evaluated banks that applied to open a transaction relationship, and how it made the decision to approve or reject such applications. Nihon Ginko Enkakushi (The History of the BOJ) also contains comprehensive records of the individual transaction relationships between the BOJ and private banks. 3 Based on these materials, a database of the transaction relationships was constructed and matched with another containing financial data of individual banks. Using the matched dataset, I econometrically analyze the determinants of the transaction relationships between the BOJ and private banks. Adding another dataset to the above data, I directly examine how the BOJ selected the recipient banks of LLR loans. Lastly, I investigate how the transaction relationships with the BOJ affected bank management. Specifically, I examine the effects on the portfolio management and risk-taking of a bank to see whether moral hazard, which the literature indicates is a possible consequence, was incurred or not.

Under the unstable financial system of the 1920s, the BOJ actively intervened in the market as the LLR, which is reflected in the spikes in BOJ lending during periods of bank panic. The BOJ concentrated LLR loans with those banks which already had a transaction relationship with the BOJ, and it selected transaction counterparts based on the applications made by private banks. From case studies regarding the opening of transaction relationships, we found that the BOJ used the following criteria in selecting counterparts: (a) the financial condition of the bank (i.e. profitability and soundness of the portfolio), (b) the composition of the directors and large shareholders, and their private assets, (c) the scale of the bank and its position in the local financial market, and (d) the availability of funds other than BOJ loans. This finding is confirmed by econometric analysis of the determinants of the transaction relationship. That is, the probability of having a transaction relationship with the BOJ was high for those banks whose ROA was high and whose scale was large. And, for banks whose ROA was low and whose reserve ratio was low, the probability of a transaction relationship with the BOJ being closed was high. It is noteworthy that banks whose profitability was low could not maintain a transaction relationship with the BOJ. This policy of the BOJ in selecting transaction counterparts was consistent with the policy for selecting Special Loan recipients. In selecting recipients from the banks which needed rescue, the BOJ focused on their profitability. The probability of receiving an LLR loan was higher for a bank with a higher ROA. Further, a transaction relationship with the BOJ did not have the overall effect of increasing the probability for transaction counterparts to receive LLR loans, but it did increase the probability for profitable counterparts to receive LLR loans. As a result, a transaction relationship with the BOJ had no significant overall effect on a bank’s survivability, but it did enhance the effect of a high ROA on the survivability of a bank. These policies of the BOJ were effective in preventing moral hazard in banks, which could be incurred by LLR loans. There is no evidence that a transaction relationship with the BOJ had a significant impact on a bank’s risk-taking. The 1920s was an epoch of structural changes in Japanese financial history. Due to the harsh competition and poor macro-economic environment, a number of banks exited through mergers and failures. At the same time, during this wave of mergers and failures, an important feature of the Japanese financial system, namely, close ties between banks and industrial firms, which was one of the basic sources of bad loans in this period, declined (Teranishi, 2003 and Okazaki et al., 2006). The LLR loans by the BOJ were a measure to cope with financial instability accompanying the structural changes. However, it is possible that the LLR loans themselves could have incurred a moral hazard and impeded structural changes. The BOJ’s LLR policy successfully avoided this problem, while mitigating instability in the financial system.