| Autor: |
JAIN, ARVIND K., GUPTA, SATYADEV |
| Předmět: |
|
| Zdroj: |
Journal of Money, Credit & Banking (Ohio State University Press); Feb87, Vol. 19 Issue 1, p78-89, 12p, 2 Charts |
| Abstrakt: |
An analysis of the causal links between the lending behavior of the banks of different sizes points to less herding than many observers believe existed. If the banks were herding, there were many leaders and the nine largest banks were not the only leaders. The ambiguous evidence of imitative behavior between the top nine and the next fifteen banks on one hand, and between the regional and both the top nine and the next fifteen banks on the other hand, indicates that one cannot draw the line between the leader and the follower banks after the top nine banks. Clearly some of the fifteen banks were among the leaders. The results that show regional banks followed both the top nine and the next fifteen banks strengthens this conclusion. This research was based on the assumption that one can divide U.S. commercial banks into homogeneous groups according to size. This assumption was dictated more by the availability of data than by any economic criterion. It is possible that a different classification of banks might lend stronger support to the herding hypothesis. The model specification also assumes that the domestic assets of the three groups of banks are similar. Differences in the risk-return characteristics of the domestic portfolios of the three groups of banks could lead them to acquire different international portfolios. It is conceivable that banks in different groups were combining domestic and international portfolios in different proportions to achieve a certain risk-return combination for their total portfolios. In that case, one could not characterize bank behavior as herding and there would be no a priori reason for the research done here. The nature of the available data also created some problems for the application of Granger-Sims causality tests. The tests, using semiannual observations, assume that the relevant time lag for herding is half a year. Although the frequency with which the data are collected dictated the use of semiannual data, the time... [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
