دانلود رایگان مقاله انگلیسی جامع عموم و متخصصین بازار اعتباری - الزویر 2018

abstract

In this paper, we propose a method to analyze the structure of the credit market. Using historical data from Japan, we explore banks’ lending patterns to the real economy. We find that generalist banks (with diversified lending) and specialist banks (with focused lending) coexist, and tend to stick to their strategies over time. Similarly, we also document the coexistence of generalist and specialist industries (based on their borrowing patterns). The observed interaction patterns in the credit market indicate a strong overlap in banks’ loan portfolios, mainly due to specialist banks focusing their investments on the very same generalist industries. A stylized model matches these patterns and allows us to identify economically meaningful sets of generalist banks/industries. Lastly, we find that generalist banks are not necessarily less vulnerable to shocks compared to specialists. In fact, high leverage levels can undo the benefits of diversification.

7. Conclusions

In this paper, we propose a method to analyze the structure of the credit market and apply it to the case of Japanese banks’ loan portfolios. We find that a stylized model of generalists and specialists captures the main features of the credit network. This indicates that the interactions between banks and the real economy can be described by a reasonably simple structure, where generalist banks interact with firms from all kinds of industries and specialist banks tend to focus their lending on generalist industries. Quite remarkably, despite the fact that the Japanese banking system underwent substantial changes in its institutional features over the sample period, there is no obvious structural break in the fit of the model. Rather, we observe a very slow deterioration of the fit.

Our findings suggest several interesting avenues for future research. First and foremost, it is important to ask whether other banking systems display a similar generalist–specialist architecture. In line with the empirical finding of a general core-periphery structure of interbank networks, we expect that the generalist– specialist architecture is not unique to the Japanese banking system. Given that our dataset exhibits a sample bias (i.e., we only observe borrowing of listed companies), we look forward to applications of the model to more complete datasets. Another important question is how the structure of the credit network affects systemic risk as a whole. Is the observed generalist–specialist architecture more or less vulnerable from a systemic perspective? Finally, the characteristics of generalist banks match those of moneycenter banks in interbank networks. It is of utmost importance to explore interactions between these networks, both theoretically and empirically, and incorporate them in broader macroeconomic models.