Endogenous Credit Cycles

Abstract:
We study a dynamic continuous-time model of a credit market, where borrower’s type is unknown to the lenders, and a credit bureau reports information about borrowers’ past. Credit bureau has limited memory: records older than a certain limit are deleted. When credit memory is long the model generates endogenous, deterministic credit cycles in the absence of any aggregate shocks. Aggregate lending booms and busts happen sequentially and stem from the endogenous amount of information about borrowers produced by the market and reported by the credit bureau. Shortening credit bureau’s memory makes cycles shorter.