Research

Abstract: This paper studies the role of insurance payment timing in mortgage performance after Hurricane Irma. Using Freddie Mac loan-level data matched to insurer-level claims data, I find that faster claim settlements significantly reduce post-disaster mortgage delinquency. A 1% increase in the loss-year payment share reduces delinquency by 0.34% over the year following the hurricane. Payment timing matters most for financially constrained borrowers with limited access to alternative credit. The effects of delayed insurance payments are concentrated among borrowers and Freddie Mac, with little transmission to capital-market investors. The costs borne by borrowers and Freddie Mac are roughly equal to the amount of delayed settlements. Since insurers ultimately pay these claims, delays primarily shift liquidity costs onto other stakeholders rather than reducing insurers' obligations. A model of insurer payment incentives, supported by time-series and cross-sectional evidence on insurer liquidity and asset composition, rationalizes why some insurers delay settlements. 

Abstract: In this paper, we show that investor composition shapes both the pricing of mortgage credit risk and the extent to which prices reflect underlying credit fundamentals. Using a novel dataset of Credit Risk Transfer (CRT) securities issued by the Governmentsponsored Enterprises (GSEs), we find that mutual funds, which account for more than 55% of CRT purchases at issuance, lower CRT spreads and weaken the relationship between underlying default risk and prices. A one-standard-deviation increase in mutual fund share reduces spreads by 8.2%, translating into a 3% reduction in the credit-risk component of guarantee fees. More importantly, mutual fund participation attenuates the pass-through of expected default risk into prices, showing that investor composition affects the cost of risk transfer and the mapping between fundamentals and market prices. As policymakers weigh GSE privatization, our results show that who bears mortgage credit risk can be as important as how much risk is transferred.

Abstract: Many subnational governments worldwide heavily rely on commercial real estate (CRE) taxes, a fiscal architecture currently threatened by the confluence of hybrid work, high inter- est rates, and asset stranding. While the financial stability risks of falling CRE valuations are well-understood, the fiscal transmission channel remains underexplored. We theoretically and empirically establish an “Urban Doom Loop”, a self-reinforcing dynamic where eroding CRE values constrain municipal budgets and force public service cuts, further depressing CRE demand. Using a novel, AI-assisted dataset of U.S. municipal bond yields as an empirical laboratory, we estimate dynamic panel models for three major metropolitan areas. We document a severe, bidirectional transmission mechanism: declines in CRE performance drive up municipal borrowing costs and force real-economy public service cuts with an 18-to-24-month delay. In turn, rising municipal credit risk exerts a severe, contempora-  neous downward drag on CRE valuations. Because the mechanisms driving this loop exist across many advanced and emerging economies, our findings reveal a structural vulnerability inherent to CRE tax-reliant fiscal architectures globally, underscoring the critical need for subnational fiscal diversification.