Vintage Growth and Population Density: An Empirical Investigation

This paper tests a central prediction of vintage growth models of urban structure: that there are discontinuities in the population density function. The data set covers quarter sections in Chicago in 1980. Using such highly disaggregated data is critical because discontinuities are less likely to be found the larger is the unit of observation. Both a switching regression model and a nonparametric estimator reveal discontinuities and upward-sloping segments in the function, which supports the vintage growth model.