In a modified Diamond-Dybvig framework that incorporates an interbank asset market, we
examine how regional liquidity shocks influence asset prices and the stability of the banking
system. The framework posits banks in two distinct regions with offsetting depositor liquidity
needs, thereby eliminating aggregate risk. Bankers initially allocate deposits between cash and
assets. After observing the state, they can adjust their portfolio in the interbank market. Patient
depositors decide whether to run on a bank based on asset price realizations in the interbank
market and the implications for the liquidity and/or solvency of their bank. We advance a
theory based on the canonical banking model that captures these features and design laboratory
experiments to test the theory’s predictions. The theory predicts that bankers can achieve the
first-best outcome, where consumption is independent of the region or state, and bank runs
are avoided. However, our experiment reveals that bank runs occur with some frequency and
asset prices exhibit significant volatility. These findings are robust to changes in the nature of
the idiosyncratic liquidity shocks. Compared with the theory, bankers over-allocate deposits
to cash, inflating asset prices in the interbank market. We find that interbank asset markets
create heretofore neglected strategic uncertainty among banks leading to distorted allocation
decisions, asset mispricing, and destabilization of the banking sector.