Benefits of hazardous waste cleanup: new evidence from survey- and market-based property value approaches

Contemporary Economic Policy, July, 2005 by Sudip Chattopadhyay, John B. Braden, Arianto Patunru

I. INTRODUCTION

Studies have shown that community knowledge of the presence of a potentially hazardous toxic site results in significant value depreciation of nearby real estate. A majority of these studies involve ex post hedonic analysis of the real estate market with data from before and after community knowledge of the change in site conditions. (1) In the absence of any change in community perceptions, studying the welfare implications of cleaning up a toxic site would involve investigating the differential impact of the site on the properties across nearby locations. Unfortunately, very few studies have systematically pursued this line of research, despite its policy relevance, especially at the level of local governments. In the United States, the vast majority of Superfund sites await cleanup, while the local economies lag behind. Significant costs of cleaning up often come in the way of an early remedial action (Hamilton and Viscusi, 1999). (2) This article focuses on one such hazardous site and investigates its differential impact on the properties across nearby locations to estimate the benefits of a proposed Superfund cleanup.

With a view to improve reliability, this article presents applications of two different models: the survey-based random utility model (RUM) and the market-based hedonic model. The welfare estimates are first compared between the two models and then with a number of market-based hedonic studies in the past. Implications of financing the cleanup, through increased property tax revenues are also investigated.

The survey-based conjoint choice approach is particularly suited to modeling housing choice. In this approach, respondents are asked to select the best alternative from all the available options in the choice set. (3) Each choice option in a choice set may consist of attributes of a differentiated product, such as housing. The appeal of the conjoint choice approach lies in its ability to let the analyst design choice options that include plausible site remediation scenarios along with structural and locational features of the homes that are located near the toxic site that is proposed to undergo remediation. This article employs the conjoint choice approach to survey homeowners and models housing choice as a discrete choice process under the RUM framework. One aim of the present study is to empirically investigate how effective the conjoint choice approach is in revealing homeowners' willingness to pay (WTP) for different levels of site remediation.

Like any choice-based study, the present study includes not only an elaborate and carefully made survey design, the choice experiment designed for the purpose--called orthogonal fractional factorial design in the statistical literature--has a special appeal, as it leads to efficient estimation of the utility parameters under the RUM framework. Unfortunately, econometric literature does not provide much guidance to judge the empirical validity of the results of the conjoint-based RUM estimation. To address the effectiveness of the survey-based approach, we present a hedonic price regression model to explore how welfare estimates obtained in the market-based hedonic model compare with those based on the survey-based RUM model. Comparing the two econometric methodologies is not only an important empirical exercise, it may also provide better understanding as to what needs to be done to bridge the gap between the welfare results from the survey-based and market-based approaches.

To achieve the above stated objectives, our application is based on market transactions of single-family homes by a sample of homeowners and a subsequent survey of the same sample of homeowners in Lake County, Illinois. Contaminated sediments in Waukegan Harbor, a Superfund site on the western shore of Lake Michigan, have been a major concern of the residents of the City of Waukegan and other nearby cities in Lake County. The site has been partially remediated, but a major condition for further investigations in cleanup is that the benefits are sufficient to justify the costs. This study builds on a recent exploratory study by Braden et al. (2004) and uses a more exhaustive dataset to shed light on the comparability of survey- and market-based methods with regards to the benefits to the residents of Waukegan, as well as on the potential for tax increment financing to fund cleanup.

In the next section we discuss issues relating to Superfund cleanup, with a particular focus on the Waukegan Harbor Superfund site. Section III discusses the data and presents the fractional factorial experimental design developed for use in the conjoint choice-based survey questionnaires. Section IV is devoted to theoretical specification of the RUM appropriate in the present context. The empirical results obtained from the RUM and the hedonic models are presented in Section V. Section VI focuses on computations of the welfare impacts, and comparison of those impacts between the RUM and the hedonic models, and between our study and some recent hedonic studies. Section VII concludes.