Abstract: This paper examines inefficiencies arising from a lack of long-term contracting in small business lending in China. I develop and estimate a dynamic model where firms repeatedly interact with the same lender. All loans are short-term. Collateral can be used to deter a strategic default by a firm, but the lender cannot recover the full value of the collateral in the case of a default. The endogenous contract terms—including interest rates, loan size and collateral—reflect a firm’s probability of default in equilibrium. Learning drives the dynamics of contract terms because a firm’s profitability type is unknown. Long-term contracts improve welfare mainly by mitigating the incentives for a firm to default.  

Abstract: We study the unintended effects of the Home Affordable Refinance Program (HARP) on mortgage borrowers. Originally designed to help underwater borrowers refinance after 2008, HARP inadvertently amplified the market power of incumbent lenders by introducing a cost differential between incumbents and their competitors. To assess the welfare implications of this cost advantage, we develop and estimate a structural model of dynamic refinancing decisions with lenders’ offers arising from a search and negotiation process. Our findings reveal that although the cost asymmetry was rectified by a 2013 policy, it still resulted in a welfare loss exceeding the impact of search frictions.

Abstract: Using detailed data from the introduction of a new high value payment (HVPS) system in Canada, we show how participants learn to use the new gridlock resolution arrangement and coordinate to create gridlocks to save liquidity. These observed behaviors are consistent with the equilibrium of a ``gridlock game'' that captures the key incentives participants face in the system. The findings have important implications for the design of HVPSs and shed light on financial institutions' liquidity preference.

Abstract: This paper studies how the liquidation mechanism in DeFi loans affects the price volatility of cryptocurrencies. We consider the two commonly used mechanisms: fixed spread, and auctions. We show that results depend on liquidators' cost of participating in liquidations. If this cost is low, auctions amplify negative shocks less than the fixed spread mechanism. If this cost is high, the reverse is true. If the cost is intermediate, auctions amplify small negative shocks less and big shocks more than the fixed-spread mechanism.