[Abstract] In June 2020, the U.S. faced a surge in cases of COVID-19. Immediately prior to this wave of cases, the largest mass protests in U.S. history took place. We show that when holding other factors constant, COVID-19 cases increased most in places where more demonstrations occurred. We exploit variation in rainfall during the protest period as an exogenous source of variation in attendance. We find that good weather coincides with both more people protesting and more subsequent COVID-19 cases and deaths. Thus, mass gatherings during a pandemic can lead to more contraction and fatalities of COVID-19, and we quantify these effects.

[Abstract] Flooding remains the most destructive and costliest natural hazard. Among the 371 billion-dollar climate disasters in the US since 1980, 288 (78%) are directly related to flooding. And they cost 1.37 trillion in losses (in 2023 dollars adjusted by CPI) and 9,722 in fatalities. With climate change and sea level rising, floods are becoming more and more frequent and costly. The pressing issue is how we manage the flood risk. The current risk management framework in the US, designed and implemented by the government, has three fundamental pillars: Flood Risk Information Infrastructure, Flood Insurance, and Government Aid. The first pillar, the Flood Risk Information Infrastructure, guides decision-making using FEMA flood maps that visually depict flood risks. The second pillar, Flood Insurance, managed by the National Flood Insurance Program since 1968, aids in addressing ex ante flood-related financial risks. The third pillar, Government Aid, provides post-flood support through programs such as FEMA’s IHP and the Small Business Administration’s disaster loans. Though vital for flood risk mitigation and economic decision-making, each pillar confronts multiple challenges. FEMA flood maps, as the foundational pillar, face challenges in mapping accuracy, update regularity, and coverage, suggesting a need for ongoing innovation and cooperation among government agencies, research entities, and private sectors. The second pillar, FEMA flood insurance, has its issues too. Its pricing, marked by a multifaceted risk rating system, often doesn't mirror actual loss potential, leading to criticisms. Despite its role as a financial buffer, flood insurance's limited coverage and low adoption rates highlight public reluctance toward it. Government aid, the third pillar, while meant for post-disaster recovery, carries unintended side effects such as spatial discrepancies, increased inequalities, and potential deterrence from buying insurance, underscoring the delicate balance between policies and societal impacts. Furthermore, the interplay between the three pillars also creates enormous complexity in practice and might create unintended consequences. One example is the free-riding problem due to the coexistence of ex ante flood insurance and ex post government aid. Therefore, it is crucial to provide a comprehensive overview of the flood risk management system for both academics and policymakers.

[Abstract]   This paper explores how immigration policy affects residential construction and home prices. Using a staggered-rollout design, we show that increased immigration enforcement reduces homebuilding and drives up new construction prices. Existing stock prices reflect some downward pressure from reduced demand, however upward supply-side pressure dominates in most neighborhoods---except those with the highest baseline share of undocumented residents. Exploring mechanisms, we find a significant loss of undocumented workers in construction, and that domestic workers do not provide one-to-one replacement. Instead, undocumented labor appears to complement domestic labor resulting in net job losses for US-born workers, especially in higher-skilled occupations.

[Abstract]  Credit access is central to homeownership, yet racial disparities in mortgage lending persist. While lender have increasingly adopted diversity policies, it is unknown whether and how such policies impact mortgage lending disparities. Using a robust difference-in-differences design, we find that diversity policies widen approval disparities, reducing origination rates for minority borrowers. Results from an instrumental variables approach and event-study design support a causal relation. Ex-post loan performance suggests the widened disparities cannot be fully explained by application risks. Mechanism analysis indicates that a stigma effect in approval decisions, triggered by an increase in risky minority applications, likely plays a role.

[Abstract] Several wholesale funding markets are dominated by government agencies such as the Federal Home Loan Bank (FHLB), which collectively channel hundreds of billions of dollars into the banking sector every year. Proponents of this intervention argue that it lowers retail borrowing costs significantly. This paper exploits quasi-experimental variation in access to low-cost wholesale funding from the FHLB arising from banks mergers, and shows that access to this funding source is associated with an 18-basis-point reduction in a bank's mortgage rates and a 16.3% increase in mortgage lending. This effect is 25% stronger for small community banks. At the market level, a census tract experiences an increase in local competition after a local bank joins the FHLB, with the market concentration index (HHI) falling by 1.5 percentage points. This intensified local competition pushes other lenders to lower their mortgage rates by 7.4 basis points, and overall market lending grows by 5%. Estimates of a structural model of the US mortgage market imply that the FHLB increases annual mortgage lending in the US by $50 billion, and saves borrowers $4.7 billion in interest payments every year, mainly through changing the competitive landscape of the mortgage market.

[Abstract] This paper examines how negative retail demand shocks are shared between commercial landlords and tenants, and how eviction protections shift this burden. Exploiting plausibly exogenous regional variation in the summer 2020 spread of COVID-19, instrumented by rainfall during Black Lives Matter protests, we show that adverse retail demand shocks transmit to landlords and their lenders through higher delinquency and foreclosure rates on retail mortgages. In response, distressed landlords reshuffle their tenant base by evicting in-line stores while retaining anchor tenants, consistent with cutting weaker renters to restore cash flow. Where eviction protections are in place, small retailers are shielded and losses remain on leveraged landlords’ balance sheets.

[Abstract] Flood risk poses a growing threat to economic activities and real estate in the US. Intuitively, both residential households and commercial businesses should gradually move out of flood zones, especially when those areas face increasing flood risk. However, we find a puzzling pattern that shows businesses in US grew 10% faster in certain central business districts in the past two decades when those districts were designated as flood zones. We build a spatial model that explains this puzzle: Business misplacement in response to rising flood risk is due to a free riding problem, in which flood insurance is imperfectly enforced while governments provide financial aid to uninsured properties ex post. Our model predicts lower commercial rent after an area becomes a flood zone, matching price results we obtain from our empirical analysis. Finally, we quantify welfare losses of imperfect flood insurance enforcement using our model.

[Abstract] This paper examines realtor-loan officer referral networks as a key source of mortgage market power. Despite the high level of competition in mortgage lending, significant price dispersion persists. We argue that realtors steer homebuyers toward a limited set of loan officers, restricting borrower choice even in competitive markets. Using a unique dataset that maps the entire realtor-loan officer network across 17 states and Washington, D.C., we document substantial concentration within these networks, with 85% of realtors likely referring their clients to a limited number of loan officers. Borrowers who work with high-concentration realtors pay 12 basis points higher mortgage rates, even after controlling for borrower and mortgage characteristics. Instrumental variable (IV) estimates confirm that referral-driven constraints impose a premium of 19.7 basis points (equivalent to $2,722 in upfront costs) on homebuyers who choose referred loan officers. This premium primarily results from suboptimal lender selection and is particularly severe for Black, Hispanic, and financially constrained borrowers. While referred loan officers might improve the likelihood of mortgage approval and expedite mortgage processing (by 0.45 days), these benefits do not fully justify the higher borrowing costs. Our findings suggest that realtor referral networks reinforce mortgage market power, imposing significant financial burdens and raising equity concerns for borrowers.

[Abstract] This paper provides evidence that weakening power separation in the government reduces private investment. We study a wave of municipal leadership consolidation in China, during which the chief of the police department is appointed to be the adjunct supervisor of the courthouse. This concentration of executive and judicial powers leads to the aggressive increase of the police's law enforcement against private businesses for rent-seeking. Concerned about property right infringement in the future, businesses reduced their capital expenditure by 14.2% and R&D investment by 21.6%. They also splash out on developing relationships with government officials.

[Abstract] Credit default swaps (CDSs) create side bets on the underlying assets, where CDS traders have an incentive to manipulate the fundamental assets' performance. Focusing on the private mortgage securitization market, we empirically find that mortgage pools have 4% more mortgage refinance and 2% less default if they are covered by CDSs. We further explore the local randomization due to the discontinuous sale of mortgages by their originators, to establish a causal relationship between CDS coverage and mortgage refinance and default performance. The difference is largely driven by the cases in which the CDS seller and mortgage servicer are in the same financial holding company.  These empirical patterns suggest that CDS seller is unloading their credit risk by helping borrowers refinance their mortgages through the affiliated servicers. Our paper provides direct evidence that credit derivatives can affect fundamental assets through ex-post actions of derivative traders, which contaminates the fairness of financial transactions. Furthermore, this match-fixing behavior through refinancing is more severe when credit derivatives are allowed to be sold to parties without insurable interests.