Generalists And Specialists In The Credit Market
University of Oxford – Said Business School; Institute for New Economic Thinking at the Oxford Martin School; Kiel Institute for the World Economy
Université Libre de Bruxelles (ULB)
March 21, 2016
In this paper, we explore the cross-section of Japanese banks’ industrial loan port- folios over the last 34 years. We show that banks with diversified lending (generalists) and banks with focused lending (specialists) coexist. Similarly, industries with diversified borrowing (generalists) and industries with focused borrowing (specialists) also coexist. Interestingly, specialist banks and specialist industries rarely interact, suggesting significant overlap in banks’ loan portfolios. We introduce a model to describe these interaction patterns and identify a persistent and economically meaningful set of generalist banks/industries. We find that size is an important determinant for being a generalist. Finally, we show that generalist banks tend to be less vulnerable compared to specialist banks.
Generalists And Specialists In The Credit Market – Introduction
Whether banks should diversify their loan portfolios or focus on a small number of industries where they have special expertise remains an open research question. In fact, banks often face conflicting incentives which may encourage either diversification or specialization.1 On the one hand, banks that extend loans to firms from many different economic industries should be, through the benefits of diversification, less affected by firm- or industry-specific shocks. On the other hand, there is no doubt that gaining industry-specific expertise, e.g., via the screening and monitoring of a particular type of firm, is valuable but costly to banks (e.g., Stomper (2006)). By focusing on relatively few types of businesses, banks might therefore be able to improve their performance, but will be more vulnerable to industry-specific shocks. In what follows, we define generalist banks as those banks that diversify their loan portfolios across many different industries, thereby interacting with a very heterogeneous set of firms. We also define specialist banks as those banks that hold more concentrated portfolios and interact only with firms from a relatively small subset of industries.
Neither the theoretical nor the empirical literature offer a unanimous recommendation on whether it is optimal for banks to be generalist or specialist.2 Not surprisingly, existing empirical evidence indicates that banks’ levels of diversification can be quite heterogeneous (e.g., Acharya et al. (2006) for Italy, and Hayden et al. (2007) for Germany), but little is known about the prevalence of generalists and specialists in these systems. We seek to fill this gap.
In this paper, we explore the characteristics of generalist and specialist banks in more depth using detailed data on Japanese banks’ industrial loan portfolios over the period 1980 – 2013. By the same token, we are also interested in analyzing the interaction patterns for the borrowing side of the credit network. For this purpose, we introduce a generalist-specialist model that allows us to identify the different types of banks and industries based on the interactions between these actors in the credit market.
A major advantage of our generalist-specialist model is that we can explicitly study the heterogeneity of these groups’ lending and borrowing portfolios respectively. One interesting question in this regard is whether specialists tend to occupy niches, e.g., to what extent focused banks tend to invest in certain industries where few other banks are present. In other words: are specialist banks indeed special? We would expect this to be the case, since otherwise there is little room for gaining superior information relative to all competitors with the very same strategy. In fact, competition should lower the expected profits from gathering this particular information. Our major finding is that specialist banks are not special at all, since they tend to interact with the very same generalist industries.
Figure 1 provides a graphical network representation of the interactions between Japanese banks and industries from our data in the year 1980, and illustrates the main results of this paper. First, we find that a similar dichotomy of generalists and specialists can be applied to both the credit supply side, i.e., banks (nodes on the left), and the credit demand side, i.e., industries (nodes on the right). In Figure 1 nodes are sorted according to their number of connections, such that generalists are sorted above the specialists. It is apparent even to the naked eye that highly diversified banks coexist alongside much more specialized ones and a similar result holds for the industries’ borrowing partners. Somewhat surprisingly, however, when we explore the connection patterns in the credit network in Figure 1, it also becomes clear that specialist banks tend to interact with generalist industries, and similarly that specialist industries tend to interact with generalist banks. Hence, the set of specialist banks tends to concentrate their loan portfolios on the very same economic industries, which implies a significant overlap in their loan portfolios. The theoretical literature suggests different reasons for such a strong portfolio overlap (see, e.g., Rajan (2005); Acharya and Yorulmazer (2008); Wagner (2010)), but it is not clear which of these theories are most relevant in the case of banks’ loan portfolios. Again, a similar remark can be made about the borrowing patterns of specialist industries: they tend to borrow from the same set of generalist banks, suggesting that these industries might be rather vulnerable to the funding behavior of a small set of banks. Overall, there are very few interactions between specialist banks and specialist industries – nodes closer to the bottom of Figure 1 are almost unconnected.
See full PDF below.