dc.description.abstract |
<p>In the first chapter of this dissertation, I outline a hedonic equilibrium model
that explicitly controls for moving costs and forward-looking behavior. Hedonic equilibrium
models allow researchers to recover willingness to pay for spatially delineated amenities
by using the notion that individuals "vote with their feet." However, the hedonic
literature and, more recently, the estimable Tiebout sorting model literature, has
largely ignored both the costs associated with migration (financial and psychological),
as well as the forward-looking behavior that individuals exercise in making location
decisions. Each of these omissions could lead to biased estimates of willingness
to pay. Building upon dynamic migration models from the labor literature, I estimate
a fully dynamic model of individual migration at the national level. By employing
a two-step estimation routine, I avoid the computational burden associated with the
full recursive solution and can then include a richly-specified, realistic state space.
With this model, I am able to perform non-market valuation exercises and learn about
the spatial determinants of labor market outcomes in a dynamic setting. Including
dynamics has a significant positive impact on the estimates of willingness to pay
for air quality. In addition, I find that location-specific amenity values can explain
important trends in observed migration patterns in the United States.</p><p>The second
chapter of this dissertation describes a model which estimates willingness to pay
for air quality using property value hedonics techniques. Since Rosen's seminal 1974
paper, property value hedonics has become commonplace in the non-market valuation
of environmental amenities, despite a number of well-known methodological problems.
In particular, recovery of the marginal willingness to pay function suffers from important
endogeneity biases that are difficult to correct with instrumental variables procedures
[Epple (1987)]. Bajari and Benkard (2005) propose a "preference inversion" procedure
for recovering heterogeneous measures of marginal willingness to pay that avoids these
problems. However, using cross-sectional data, their approach imposes unrealistic
constraints on the elasticity of marginal willingness to pay. Following Bajari and
Benkard's suggestion, I show how data describing repeat purchase decisions by individual
home buyers can be used to relax these constraints. Using data on ozone pollution
in the Bay Area of California, I find that endogeneity bias and flexibility in the
shape of the marginal willingness to pay function are both important.</p><p>Finally,
in the third chapter of this dissertation, I combine the insights of the Bajari-Benkard
inversion approach employed in second chapter with more standard estimation techniques
(i.e., Rosen (1974)) to arrive at a new hedonic methodology that allows for flexible
and heterogeneous preferences while avoiding the endogeneity problems that plague
the traditional Rosen two-stage model. Implementing this estimator using the Bay
Area ozone data, I again find evidence of considerable heterogeneity and of endogeneity
bias. In particular, I find that a one unit deterioration in air quality (measured
in days in which ozone levels exceed the state standards) raises marginal willingness
to pay by $145.18 per year. The canonical two-stage Rosen model finds, counter-intuitively,
that this same change would reduce marginal willingness to pay by $94.24.</p>
|
|