Revisiting Old and Investigating New Empirical Evidence.
By George S. Ford and R. Alan Seals
A key NIMBY complaint in the siting and expansion of quarries is the effect of the operations on nearby home values. According to Census data, housing amounts to about 70 percent of the average American’s net wealth, so naturally homeowners are sensitive to any adverse effect, real or imagined, on property values.
Despite NIMBY opposition, nearly all the evidence on quarry operations finds no price effect. Frequently mentioned studies include Rabianski and Carn (1987) and Dorrian and Cook (1996), both of which find no relationship between appreciation rates of property values near to and far from quarries.
An even earlier study conducted for the U.S. Bureau of Mines in 1981 also found no consistent relationship between quarry operations and the prices of nearby homes. There are a number of consulting reports on the question, and none report price attenuation attributable to a quarry.
Opposition to quarries based on home valuations relies universally on a report by Professor Patricia Hite (2006). This brief, 250-word study analyzes data from a few thousand homes sales (apparently in the mid-to-late 1990s) around a single quarry in Delaware, Ohio.
Using an unconventional regression model and data on transactions occurring decades after the quarry opened, the Hite report finds a positive relationship between home prices and distance from the quarry. Based on that evidence, the report concludes that quarries reduce home values. Yet, the report’s methods and data do not support a causal interpretation.
Despite replicating both the location and methods of the Hite report, our regression analysis finds that prices fall – not rise –as distance from the quarry increases. This result conflicts with that appearing in the Hite report, so we look for more evidence by analyzing data on homes sales near a quarry outside of Murfreesboro, Tenn., over the same time interval. Again, we find prices fall as distance from the quarry increases.
We are reluctant, however, to claim this evidence implies quarries raise home prices. Rather, we conclude, based on the method of randomized inference and other tests, that the Hite report’s method is unreliable. Using a simulation of pseudo-treatments, we find that the null hypothesis that home prices rise or fall in distance from a randomly selected location is rejected in no less than 67 percent of cases at the 10 percent nominal significance level.
Estimating price-distance relationships, especially without explicitly considering selection bias, is a highly unreliable statistical procedure. The nature of real estate markets do not permit the effect of quarries to be identified with such naïve empirical tests.
Second, using data on home sales near a relatively new quarry in Gurley, Ala., we augment the Hite-style analysis with a difference-in-differences estimator, which quantifies the price-distance relationship both before-and-after operations begin. By exploiting the timing of the quarry buildout and the location of home sales with respect to the quarry, we can credibly identify a causal relationship, at least in theory.
We find no evidence the quarry reduced housing prices. If anything, relative home prices rose near the quarry site. While our evidence suggests that quarries do not reduce, but may increase, home prices, our analysis suggests more than anything that the identification of the effect of quarries on prices is a very difficult problem, facing many conceptual and practical obstacles.
We do not resolve all these difficulties. That said, we can conclude the evidence strongly implies the Hite report and its methods are unreliable. Further analysis is, as usual, encouraged.
To read the full report go to Public Policy Studies, www.phoenix-center.org.