Historical path-dependence of the urban population density gradient

Recent research has emphasized path-dependence and the effect of vintage factors in urban development. Jerusalem and Tel Aviv are cities where the modern CBD is not in its historic location. Distance from the historic center is taken as a proxy of housing and infrastructure vintage. A polycentric urban population density function is used to assess the relative importance of distance from the CBD and from the historic center in explaining population density. Additionally, I examine econometrically the validity of a number of population density functions, and analyze the structure of the error variance and the design of appropriate measures of goodness of fit and hypothesis tests for regression models of population density. The results show that distance from the historical center is an important factor in explaining population density but its explanatory power has declined over time. It is more significant than distance from the CBD in explaining residential density in Tel Aviv and gross density in Jerusalem. It explains a larger proportion of the variance in gross density in the inner city than does distance from the CBD, but vice versa in the suburbs.