Exceeding the Threshold: Analysis of Public Information Transfer using Instrumental Variables

The Duke Community Standard was upheld in the completion of this assignment. 2 Acknowledgements I wish to thank Professor Christopher Timmins for his extensive knowledge, support and advice throughout the creation of this paper. I am also especially grateful to Lala Xun Ma for her never-ending patience, encouragement and Stata expertise. I would finally like to thank Professor Michelle Connolly for her mentorship and insightful feedback over the past year and a half. Without all of you, this paper would not have been possible. Abstract This paper examines how information transfer about contamination levels found at brownfield sites capitalizes into nearby property values. More specifically, a hedonic model is used to test the impact on housing transaction prices when a binary measure (i.e. exceeding a threshold or not) or a continuous measure (i.e. chemical levels) is used. In the analysis, I exploit the variation in the contaminant thresholds, caused by regulatory conditions defined by the state of Massachusetts, holding the contaminant level constant. As thresholds are tied to neighborhood attributes in areas surrounding brownfields, threshold exceedance is potentially correlated to unobserved factors that impact housing values. An instrumental variables approach is used to create variation in threshold exceedance through the use of an instrument that measures the presence to underground aquifers. After instrumenting for threshold exceedance, my estimates indicate that a 10.8% decrease in housing values occurs when a contaminant threshold is exceeded, while the continuous measures of toxicity indicate a negative but insignificant effect. These findings suggest that policy makers should consider information conveyance when creating policies to inform homeowners of pollution presence, as improved information provision may increase public awareness about local environmental concerns.