Abstract

Investments in transportation infrastructures in urban areas have significant impacts on land use and real property values. However, the (huge) literature on residential land and property values shows a great deal of variability in the estimated effects of proximity to transportation infrastructures such as train stations, bus stops and light rail stations. This paper attempts to explain the variation in the findings by using tools from meta-analysis. More than 980 published papers, along with 60 unpublished working papers, are extensively covered to assess the use of transportation infrastructures as explanatory variables within a hedonic pricing framework. A meta-regression analysis examines the variability in the valuation that might be due to country, time-span, sample size, mean property value, data aggregation or the addition of other important explanatory variables. The results indicates that model specification has important effects on the measurement of the hedonic valuation of transportation infrastructures. The way distance and/or accessibility to transportation infrastructures is measured, as well as the way nonlinearity is dealt with, are especially important. The net effect of transportation infrastructures on property values is indeed a combination of a positive effect (due to improvement in accessibility) and a nuisance effect (due to disamenities such as traffic and noise), which operate at different scales. The paper also investigates the issue of publication bias for the valuation of transportation infrastructures within hedonic pricing models. It appears that researchers report both positive and negative results, but they tend to be biased towards statistically significant estimates.