Author:

Abstract:

We develop a theoretical interbank market (IBM) model which explains a number of facts observed during financial crises: a sudden and significant increase in IBM rates, reduced participation by lending banks, liquidity shortages (i.e. the lenders' supply being smaller than the borrowing banks' demand), and ultimately an IBM freeze. Here a crucial element is a bank's ability to assess risk correctly. The model's core feature is then the distinction between ambiguity averse and risk averse lending banks. While the latter have unique priors, ambiguity averse lenders do not and consider a set of possible priors. Our key insight is that ambiguity and risk differentially affect a bank's decision to participate as a lender in the IBM. Moreover, at a given rate, the ambiguity averse lenders supply less than the risk averse ones. Deciding on the optimal amount results in a unique IBM equilibrium where either all lenders participate (full participating equilibrium) or only the risk averse ones do (partial participating one). This distinction allows us to show how a shock with ambiguity (as at the start of a crisis) leads to a sudden spike in IBM rates, liquidity shortages, and a market freeze. This ambiguity effect is more direct and stronger compared to a rise in risk. Finally, we discuss the (in)effectiveness of different government policies in supporting the IBM.