WEAI/AERE 2012 - Individual Paper Abstract
Title: What Can We Learn From Hedonic Models Where Markets Are Dominated By Foreclosures?
Author(s): Edward Coulson, Pennsylvania State University; Jeffrey ZABEL, Department of Economics, Tufts University, 8 Upper Campus Road, Medford MA 02155, USA, 617-627-2318, 617-627-3917, jeff.zabel at tufts dot edu [Photo credit: based on Wikimedia Commons]
Abstract:
Hedonic property value models have been frequently used to value environmental amenities (or dis-amenities) since markets for these goods (bads) do not usually exist. Typically, researchers cite Rosen's (1974) seminal work that allows one to interpret the coefficients as the marginal willingness to pay for the environmental good. There are a number of assumptions needed for the Rosen result to hold. One is that the market is in equilibrium. This assumption is highly questionable in today's housing markets given the large number of foreclosures that have resulted in an excess supply of housing units. In this paper, we address the following question "How can we interpret the coefficient estimates for environmental goods in hedonic property value models where markets are dominated by foreclosures?"
The issue of how to interpret the coefficient estimates for environmental goods where markets are dominated by foreclosures can be cast in a more general framework of sample selection bias. As is well known, observed sales are a subset of all housing units in a market. Hence there is an underlying question as to whether the hedonic results based on this subset of units can be generalized to the market as a whole. This is particularly true in a slow market or in other circumstances where transactions are low. Such an example is when a hazardous waste site is discovered and this results is a drop in sales around the site. How do we interpret the impact of the hazardous waste site on house prices in this case?
In the case of a large number of foreclosures, one might be tempted to throw out foreclosed sales. We investigate how this affects the interpretation of hedonic regression coefficients. This is related to the above issue of selection. Another related issue is that coefficient estimates can be affected by the housing cycle; that is estimated implicit prices of housing characteristics can vary across the cycle. We investigate this issue as well.
In order to evaluate the impact of foreclosures and other related issues, we start with basic model as outlined in Rosen (1974), making clear all the assumptions underlying the model. We then next this is a general sample selection model. We then look at specific issues related to selection such as foreclosures and show how this affects the outcome and hence the interpretation of the coefficients in the hedonic model. Some of the important papers in this area that we will draw upon in developing our model are given in the references.
The results of this analysis will have important implications for the evaluation of the benefits of environmental goods and hence the economic efficiency of policies that affect the provision of environmental goods. For example, there is a large literature that uses hedonic models to estimate the benefits from cleaning up Superfund sites (e.g. Kiel and Williams (2007), Greenstone and Gallagher (2008), and Gamper-Rabindran and Timmons (2011)). Given our findings, we will provide a re-interpretation of these results. Finally, we will come back to the initial issue of foreclosures and review the latest literature on estimating the MWTP for environmental goods and provide a critical interpretation of these results.