Research

"Over-collateralization with multiple lenders" [link]

Abstract:

Over-collateralized lending is frequently used in practice. I study a framework in which collateral serves to protect against spillovers from multiple lenders. I find that collateral is particularly potent when lending is opaque. In my model, lenders use over-collateralized contracts to force competing lenders into offering contracts that are susceptible to dilution. Furthermore, over-collateralized contracts promote competition. When the borrower cannot accept arbitrarily many contracts due to over-collateralization, more lenders make offers. I use my framework to discuss the effect of collateral-related policies such as "minimum haircuts", "loan-to-value rules" and "close-out netting". 

Further work in progress:

"Cherry-picking in the Repo Market"

Abstract:

I study how the presence of informed and uninformed lenders affects outcomes in the repo market. Since this market is opaque, uninformed lenders might fear that informed lenders secretly finance more valuable collateral, thereby reducing the value of the collateral left to uninformed lenders. Using a difference-in-difference methodology, I show that in line with my model's prediction, uninformed lenders charge higher interest rates when they expect a higher chance and impact of "cherry-picking". Primary dealers, the borrowers in my analysis, have relationships with hedge funds, informed lenders, through their prime brokerage business (ADV data). An exogenous increase in the uncertainty of all collateral asset values increases interest rates charged by uninformed lenders more for borrowers with many hedge fund connections (N-MFP data). This effect holds true only for collateral that is prone to information asymmetries, such as risky corporate bonds or equities, but not for safe corporate bonds.

"Dynamic Liquidity Provision under Capital Constraints"