Spatial autocorrelation and neighborhood quality

Although it is intuitively obvious that neighborhood quality and accessibility should affect housing prices, the empirical evidence is weak: Most hedonic estimations show few significant coefficients on the neighborhood and accessibility variables. The lack of empirical support for the capitalization of neighborhood and accessibility effects may stem from the multicentric nature of the city as well as measurement problems with regard to neighborhood quality. In this paper, an alternative approach is taken: Omit all neighborhood and accessibility measures from the set of explanatory variables and instead model the resulting autocorrelation in the error term. Data from Baltimore are used in an empirical example; the results show that this approach provides a very plausible pattern of housing price variation.