Research

Agency MBS issuers can repurchase delinquent loans from securitization pools at par. I develop a framework in which loss mitigation becomes privately optimal only once the issuer owns the loan, and I study how the timing of buyout eligibility affects loan performance. Exploiting Ginnie Mae's suspension of an early buyout option for FHA-insured mortgages, I find that earlier buyout eligibility reduces foreclosure by 40 percent and increases loan cures by about 45 percent, primarily through forbearance rather than loan modification. For a subset of loans, delayed eligibility postpones effective intervention and leads to irreversible deterioration in loan performance.

This paper examines how market structure shapes the pass-through of secondary-market funding costs to mortgage borrowers. We exploit the Federal Housing Finance Agency's unexpected announcement of the Adverse Market Refinance Fee in August 2020, which imposed a 50 basis point surcharge on conforming refinance loans sold to the GSEs. Using high-frequency mortgage rate-lock data and a difference-in-differences design, we find that lenders fully passed the fee through to borrowers, with 58 percent reflected in higher interest rates and 46 percent in upfront charges via discount points. Pass-through is uniform across counties regardless of local market concentration, indicating that geographic competition plays little role in shaping pass-through. In contrast, we document systematic heterogeneity by lender size—large lenders consistently exhibit full pass-through, while smaller lenders transmit significantly less, even after conditioning on time-varying local economic conditions. These findings suggest that the incidence of funding cost shocks is not mediated by local competition but instead reflects systematic differences across lender size.

I show that when repossessing collateral becomes costly, creditors choose to sell their delinquent debt on the secondary market rather than renegotiate with borrowers. Only when repossession becomes prohibitively expensive, thus impeding sale, do creditors offer forbearance. I exploit quasi-experimental variation from an increase in foreclosure costs due to Maine's Greenleaf judgment. To foreclose on loans associated with an electronic registry, the judgment required affected creditors to request reassignment of their mortgage from the initial lender. For treated loans, I estimate that foreclosures fell by 26% (0.09 pp) and debt sales increased by 57% (0.05 pp). I find no evidence of an increase in formal modifications on the part of creditors or default on the part of borrowers. In cases where the initial lender filed for bankruptcy, treated loans experienced no increase in sales. Instead, these loans benefited from reduced delinquency in a manner resembling creditor forbearance.

This paper uses a shock to downside risk in contract enforcement to study lender behavior when collateral enforcement becomes uncertain. I exploit quasi-experimental variation from Maine's 2014 Greenleaf judgment that increased enforcement costs on existing creditors but mechanically left new lenders unaffected. I estimate that the most exposed banks restricted lending by 21%, almost exclusively for portfolio loans. I provide evidence that this contraction in credit was not driven by balance sheet losses or capital flight on the part of investors. Exploring the mechanism, I find that banks likely tightened lending standards following the judgment. Consequently, exposed banks issued safer loans and denied high debt-to-income loan applications above the conforming loan limit. Furthermore, the salience of the shock increased in proximity to the judgment, size of the bank, and bank portfolio concentration. Lastly, following the judgment, I identify spillovers to other bank operations and the local economy. Small business lending increased in bank branch localities, house prices fell in the most exposed zip codes, and exposed counties experienced an increase in unemployment.