The institutional memory hypothesis and the procyclicality of bank lending behaviour

Stylised facts suggest that bank lending behaviour is highly procyclical. We offer a new hypothesis that may help explain why this occurs. The institutional memory hypothesis is driven by deterioration in the ability of loan officers over the bank.s lending cycle that results in an easing of credit standards. This easing of standards may be compounded by simultaneous deterioration in the capacity of bank management to discipline its loan officers and reduction in the capacities of external stakeholders to discipline bank management. We test the empirical implications of this hypothesis using data from individual US banks over the period 1980-2000. We employ over 200,000 observations on commercial loan growth measured at the bank level, over 2,000,000 observations on interest rate premiums on individual loans, and over 2,000 observations on credit standards and bank-level loan spreads from bank management survey responses. The empirical analysis provides support for the hypothesis.

This paper was presented at the conference on "Changes in risk through time: measurement and policy responses" organised by the BIS on 6 March 2002 and, as such, is appearing in the BIS Working Papers series.