« Back to Results

Environmental Risk, Justice, and Amenities in Housing Markets

Paper Session

Friday, Jan. 4, 2019 8:00 AM - 10:00 AM

Atlanta Marriott Marquis, M202
Hosted By: Association of Environmental and Resource Economists
  • Chair: David Albouy, University of Illinois-Urbana-Champaign

Flood Risk Belief Heterogeneity and Coastal Home Price Dynamics: Going Under Water?

Laura Bakkensen
University Of Arizona
Lint Barrage
Brown University


How will climate risk beliefs affect coastal housing market dynamics? This paper
provides both theoretical and empirical evidence: First, we build a dynamic housing
market model with heterogeneity in home types, consumer preferences, and flood risk
beliefs. The model incorporates a Bayesian learning mechanism allowing agents to
update their beliefs depending on whether flood events occur. Second, to quantify these
elements, we implement a door-to-door survey campaign in Rhode Island. The results
confirm significant heterogeneity in flood risk beliefs, and that selection into coastal
homes is driven by both lower risk perceptions and higher coastal amenity values.
Third, we calibrate the model to simulate coastal home price trajectories given a future
flood risk increase and policy reform across different belief scenarios. Accounting for
heterogeneity increases the projected home price declines due to sea level rise by a
factor of four and increases market volatility by an order of magnitude. Studies
assuming homogeneous rational expectations may thus substantially underestimate the
home price implications of future climate risks. We conclude by highlighting potential
implications for welfare and flood policy.

Sorting or Steering: Experimental Evidence on the Economic Effects of Housing Discrimination and Its Consequences for Environmental Justice

Peter Christensen
University of Illinois-Urbana-Champaign
Christopher Timmins
Duke University


Housing discrimination is illegal. Despite this fact, paired-tester audit experi-
ments have revealed evidence of discrimination in the interactions between potential
buyers and realtors, although the most blatant forms of discrimination (e.g., refusal
to show units) has declined over the last 30 years. We explore a new dimension
of housing discrimination. Using data from HUD's most recent Housing Discrimi-
nation Study combined with novel spatial attribute data, we nd strong evidence
of discrimination in the characteristics of neighborhoods towards which individu-
als are directed. Conditional upon the characteristics of the unit suggested by the
audit tester, minorities are signi cantly more likely to be \steered" towards neigh-
borhoods with less economic opportunity and greater exposures to local pollutants.
These results have important implications for studies of neighborhood e ects. In
particular, they suggest an important and understudied mechanism underlying ob-
served correlations between race and pollution described in the environmental jus-
tice literature. Results also suggest that the basic utility maximization assumptions
underlying hedonic and residential sorting models may often be violated.

Moving to Floodplains: The Unintended Consequences of the National Flood Insurance Program on Population Flows

Abigail Peralta
Louisiana State University
Jonathan B. Scott
University of California, Berkeley


Despite the large costs of covering flood losses, little is known about whether flood insurance availability
affects the decision to move and stay in more flood-prone areas. In this paper, we present evidence that
suggests households in flood-prone areas would have otherwise moved to less risky areas, absent flood
insurance availability. This moral hazard imposes an additional cost to the National Flood Insurance
Program (NFIP), which insures these households who might otherwise expose themselves to less risk.
This adds to the many inefficiencies of the NFIP, and inhibits potential climate change adaptation efforts.
Since it is reasonable to think that communities would be more or less likely to join depending on their
particular circumstances, directly using NFIP adoption may incorrectly identify the causal direction of
flood insurance. Therefore, to identify the direct effects of flood insurance, we exploit the within- and
across-county variation in the various nudges that the Federal Emergency Management Agency (FEMA)
used to induce flood-prone areas to join the NFIP.
Results suggest that flood insurance availability caused population to increase by 4 to 5 percent in high
flood-risk counties. Furthermore, by considering the intensity of historical flood risk by area, we find that
NFIP causes a 4.4 percent increase in population per one standard deviation increase in risk. We attribute
most of this effect from the increased propensity of current residents to stay in flood-prone areas, rather
than move, though we do find some evidence of increased migration into these areas. Our pattern of
results suggest that, to the extent that shielding people from the full cost of flood insurance led them to
take on additional risk that they otherwise would have avoided, the current structure of the NFIP
contributes to its own increasing cost through its effect on increasing populations in risky areas.

Estimating the Willingness to Pay for Air Quality: New Evidence from Recent Movers

Adam Theising
University of Wisconsin-Madison
Daniel J. Phaneuf
University of Wisconsin-Madison


Estimating the marginal cost of air pollution has long been of interest to environmental
economists. Historically researchers relied on the first stage hedonic model to identify how
air quality capitalizes into home prices under the assumption of perfect mobility. More
recently, residential sorting models, which can accommodate household level heterogeneity
and moving costs, have been used to estimate the value of air quality. These and other papers
examining national-level sorting (e.g. 1, 2, 3, 4) rely on the Public Use Microdata Sample
(PUMS) from the US Census, which allows observation of the current place of residence and
demographic information. However, PUMS does not provide information on actual moves.
For this paper we build a new database drawn from a large sample of recent movers, which
allows us to exploit detailed information on both households and their specific moves. With
these data we revisit the cost of air pollution in a residential sorting model.
To construct the database, we obtained records on 200,000 households that moved in 2017-
18 from a private vendor. Each record includes demographic and income information, along
with current and previous addresses and an indicator for renter or home ownership. A subset
of these movers will be surveyed during summer 2018, with the goal of obtaining explicit
information on factors affecting their location decision, other household characteristics, and
data useful for identifying the salience of air quality and the marginal utility of income.
Using the analytical framework from (5), we use the movers data in a discrete choice
framework to estimate mean preferences conditional on origin, and the survey data to
identify important dimensions of preference heterogeneity. We estimate the marginal cost
of PM2.5 concentrations, while also testing assumptions about the marginal utility of income
and household choice sets that have become standard in the sorting literature.
David Albouy
University of Illinois-Urbana-Champaign
Ann Wolverton
U.S. Environmental Protection Agency
Justin Gallagher
Case Western Reserve University
Kelly Bishop
Arizona State University
JEL Classifications
  • Q5 - Environmental Economics
  • R3 - Real Estate Markets, Spatial Production Analysis, and Firm Location