Abstract: We quantify the impact of carcinogenic risk exposure on property values and neighborhood composition. Following a plant’s first reported carcinogen emission, we observe a 0.8% to 2% decline in property values for houses closer to the plant relative to houses farther away. Combining our estimates with cancer hazard ratios from epidemiological studies implies a value of statistical life ranging from $2.6M to $6.4M. Our analysis reveals a shift towards a higher presence of minority and higher credit-risk households in houses closer to the plant. This evidence of changes in neighborhood composition in the wake of a change in perceived environmental health risk informs the debate on environmental justice and health inequities.
Discussant: Lucas Davis, University of California-Berkeley
Soon Hyeok Choi, Rochester Institute of Technology
Adam Nowak, West Virginia University
Patrick Smith, University of North Carolina-Charlotte
Alexei Tchistyi, Cornell University
Abstract: Homebuyers who participate in bidding wars are susceptible to a winner’s curse. We theoretically quantify the winner’s curse in housing markets, showing that the presence and intensity of a bidding war exacerbates the winner’s curse. We empirically test our theoretical hypotheses by examining the subsequent performance of bidding war transactions in four large US cities. We find that homeowners who purchase their property via a bidding war are more likely to default and earn lower annualized returns than those who did not purchase their property via a bidding war. We highlight the far-reaching implications of these findings by showing that the winner’s curse undermines housing affordability.
Discussant: Alina Arefeva, University of Wisconsin-Madison
Abstract: We quantify racial differences in the total rate of return on owner-occupied housing from 1975-2021.
The total rate of return of buying a house equals the price appreciation plus the rental value of its
housing services, minus taxes and maintenance. To measure the total return, we develop a new method
to estimate the rental value of each owner-occupied house, using houses that switch between the rental
and owner-occupied market. We then use this method to predict the rental value of the entire owner-
occupied housing stock and find this prediction out-performs standard hedonic techniques. We document
across multiple datasets that Black homeowners earn a .5% percentage point higher rental yield on
housing than white homeowners, consistent with our method’s estimates. This gap largely explains why
minority homeowners earn .6% percentage point higher total returns on housing. Minority homeowners’
total returns are also more volatile and sensitive to the business cycle. These racial differences can be
fully explained by other observables, with household income differences playing the largest role. Our
findings are broadly consistent with a model with a more severe credit constraint for minorities, which
bids up rents, lowers house prices, and makes house prices sensitive to credi
Discussant: Bronson Argyle, Brigham Young University
Abstract: The 1990s rollout of mortgage automated underwriting systems allowed for complex underwriting rules, cut processing time, and increased house prices. By comparing early systems with different characteristics, we separately estimate the effect of automation and the effect of new lending standards. We show that locations exposed to initial adopters of Freddie Mac’s Loan Prospector system experienced an early housing boom due to a switch to statistically-informed underwriting rules. Loan Prospector adoption increased the share of lending at high loan-to-income ratios by around 30 per cent. Using exposure to initial adopters of Fannie Mae’s Desktop Underwriter system (which initially did not apply new lending rules) we show that automation reduces loan processing time by up to one week but does not have a significant effect on house prices. House prices only grew differentially in these locations after Desktop Underwriter also started to apply new underwriting rules. Usage of both GSEs’ systems accelerated during the late 1990s refinancing boom, dramatically changing the underwriting rules U.S. lenders used. Applying our estimated price response to later adopters, we suggest that the rollout of new lending standards with the GSEs’ systems can explain a large share of U.S. house price growth between 1993 and 2002.
