Research Projects

Households with a Federal Housing Administration (FHA) loan who refinance to a new FHA loan may overpay on their mortgage, because FHA-to-FHA refinance does not reassess the property value, resulting in borrowers continuing to pay an avoidable mortgage insurance premium (MIP). Using loan level data, I find that about 40% of FHA-to-FHA refinancers during 2009-2020 had enough property value appreciation for a conventional Government Sponsored Enterprise (GSE) loan that would save about $110 (10%) monthly and $2,957 upfront on mortgage costs. I explore both demand and supply-side channels to explain this choice. On the demand side, I show that information about the value of the house is important to refinancing decision. Households who received a new tax bill are more likely to refinance to a GSE loan. Additionally, minority and low-income FHA borrowers are less likely to refinance to a GSE loan. On the supply side, I document that lenders make higher profit from FHA mortgage upfront fee and securitization revenue. After an exogenous shock in securitization revenue, FHA borrowers refinance with affected lenders are less likely to refinance to a GSE loan.

Households often incur substantial losses by refinancing their mortgages too early or failing to refinance altogether. We provide a unified framework that accounts for both types of mistakes. In the model, households with present bias decide both when to refinance and whether to use lender credits to reduce upfront refinancing costs. We characterize a closed-form solution for the refinancing threshold under present bias and show that present-biased households are more sensitive to changes in upfront costs. This combination of present bias and upfront cost choice gives rise to both procrastination (for households with sufficiently high present bias) and precrastination (for households with a moderate level of present bias that use lender credits). We then test our model using U.S. mortgage data from 2010-2019 and identify lender credits usage through FHA loan features. We find that borrowers using lender credits refinance significantly earlier in during periods when interest rate decreases and achieve less rate savings compared to those who pay closing costs out of pocket.

The efficiency of monetary policy pass through into household sector rely on households forming interest rate expectations in alignment with policy goals. However, evidence shows household are subject to significant bias. Using survey data, we document that professional forecasters often overreact to interest rate changes, while households tend to underreact. These frictions lead to mortgage refinance mistakes. Households with lower property values tend to refinance too early in declining rate environments, while households with higher property values often wait too long, missing optimal refinancing opportunities. These results contribute to the understanding of household interest rate expectation formation and household borrowing decisions.

I utilize an updated flood risk map to examine how lenders implement credit rationing in response to flood risk. In regions with elevated flood risk, conventional GSE loans demonstrate higher loan-to-value (LTV) ratios. In contrast, FHA loans do not adjust LTV based on flood risk at the intensive margin. Instead, FHA lenders adapt at the extensive margin by increasing rejection rates in areas with greater flood risk. This research underscores the complexities of risk assessment in the multidimensional mortgage contract space, carrying significant implications for public policy and financial stability.