Lynne Kiesling
Many thanks to Andrew Morriss over at The Commons for the link to this Chay and Greenstone working paper on the effects of air quality on housing prices.
The typical econometric method for performing such an analysis is called hedonic estimation, where your dependent variable is the house price and your independent variables are all of the things that you can measure that you hypothesize can affect that price. As the authors note, cross-section hedonic analyses have not yielded robust results on the value benefits that clean air provides. So they are enhancing the hedonic technique by using instrumental variables to control for some of the omitted variables (which are likely to abound in such a case). In instrumental variables, you use a variable that is highly correlated with one of your explanatory variables (TSP, or total suspended particulate, concentration in this case), but is not correlated with the error term in your single-equation regression.
Chay and Greenstone find that air quality does affect housing prices, and their econometric technique allows them to be more confident in saying that because they have been able to control for many confounding effects that are difficult or impossible to measure.
Very interesting ...
It ahould be noted that a number of regional/urban economists have been using spatial econometric techniques to estimate hedonic pricing model. Housing prices tend to be spatially autocorrelated which leads to biased and inconsistent results. I haven't read the paper but this is something that needs to be addressed.