A meta‐analysis of the effect of environmental contamination on non‐residential real estate values

Purpose – This paper seeks to reduce the lack of quantitative research by addressing diminution in value to non‐residential property resulting from environmental contamination. Design/methodology/approach – This meta‐analysis extracts data from approximately a dozen peer‐reviewed articles and 100 case studies from real estate appraisers in the USA. A dataset containing 106 contaminated non‐residential observations is examined using Regression (OLS). Forward (stepwise) and backward selection was performed. The dependent variable included percentage loss and dollar amount. The independent variables were contamination type, US region, land use type, distance from the source (mostly contaminated subjects), passage of time, year, urban or rural, market conditions, litigation, and indemnification. Findings – The model adjusted R squares range from 37 percent to 66 percent. Approximately a third of cases had no loss. This research used petroleum case studies as the reference category for comparison with other types of contamination. The following variables were statistically significant in all four models: Creosote/PCB and Other contamination. The following were significant in two models: Other land use, 30‐year mortgage rate, Rural location, TPH, Multiple contamination, TCE, Under‐remediation, and Mineral extraction region. Finally, the following variables were significant in one model at least at a 90 percent level of confidence: Heavy metals, Industrial Midwest region, and pre‐1995 sale. Practical implications – Properties in the remediation phase show less of a loss in value. Selective case studies within the same period of the clean‐up cycle make the best comparables. The US regional location was less important. Originality/value – This is the first empirical research using a meta‐analysis to study damage effects for non‐residential property affected by contamination.