Urban size (dis)amenities revisited

An article by Kelley in this Journal has argued that most studies of urban-size amenities built around the analysis of wage differentials have ignored the interaction of supply and demand forces. The result, it was argued, has been serious simultaneous equation bias which suggested that larger cities had lower-valued non-pecuniary amenities. If the bias is eliminated, the opposite is shown to be correct: urban amenities increase in value with city size. This note argues that Kelley's conclusions are themselves based on serious theoretical and econometric errors: (i) He ignores the equilibrium conditions of the model. (ii) He treats variables as exogenous which are not. (iii) He ignores almost perfect collinearity between the variables he uses.