Incorporating neighborhood effects into spatial allocation models

In this paper the authors are concerned with the explicit introduction of endogenous spatial externalities into gravity-type allocation models. The discussion focuses on the locational choice by households, which is taken to be affected by the spatial distribution of all other actors. A stochastic linear function of utility is assumed, in which the arguments include, as endogenous variables, land rents and the level of externalities. A multinomial logit model of demand for land is defined for each zone and each population group and is imbedded in a long-run equilibrium model of the land market. The equation system which results has multiple solutions which can be found by use of a simple iterative algorithm. By means of numerical examples it is shown that positive spatial externalities result in the spatial concentration of sensitive households, whereas negative externalities lead to their dispersal to the periphery. Combined effects may result in different asymmetric local equilibria depending upon the initial solution used in the algorithm.