REGIONAL INTERACTION AND THE RESERVE ADJUSTMENT LAG WITHIN THE COMMERCIAL BANKING SECTOR.

The results of most recent econometric studies of the banking sector which employ quarterly data suggest that commercial banks close less than half the gap between their actual and "desired" levels of excess and/or borrowed reserves per quarter. At this rate, the banking system would take a full year to close nine-tenths of the gap. The length of this lag has been regarded as implausible by many researchers. Among the skeptics, many have expressed suspicion of the statistical techniques underlying the estimates. This paper investigates two commonly-proposed statistical sources of bias. The investigation reveals that one of them, temporal data aggregation (averaging over time), does exaggerate the lag. Nonetheless, even after correction for this bias, the estimates indicate a full adjustment period of a quarter or more. The primary purpose this paper is to explain the remaining lag. The major hypothesis of this study is that the lag in the aggregate is caused by lagged interactions among banks in different regions. As this paper demonstrates, when Federal Reserve Member Banks are disaggregated by region, estimated adjustment speeds rise dramatically. Moreover, the lag in the aggregate is well explained statistically by the impact of some regions' reserve adjustments on other regions. The more important of these interregional effects are identifiable and statistically significant, and they are shown to be useful in analyzing the timing of the regional impacts of monetary policy. [ABSTRACT FROM AUTHOR]