Determinants of land use changes: a spatial multinomial probit approach

Changes in land use patterns impact significantly environmental conditions as well as economic and social welfare. These changes are influenced by socio-economic as well as pedo-climatic factors. A good understanding of how these factors influence land use patterns would provide new dimensions to policy making and public policy evaluation. To this end, we propose a Spatial Multinomial Probit model to examine the determinants of land use change, at the parcel level, in the French D´epartement du Rhones from 1992 to 2003. It is based on an economic model that assumes that landowners have a choice between 4 land use categories for a given parcel at a given date: (1) agricultural, (2) forest, (3) urban and (4) no use. Each landowner compares costs and benefits associated with each parcel and each class and chooses the optimal land use to maximize his/her profit. We propose a Spatial Multinomial Probit model that allows for covariates and spatial dependence, and we use these features to explore the relative importance of factors that drive landowners to choose a specific land use category. Our findings suggest the presence of spatial patterns, implying that the decision to choose land use for a given parcel of land is strongly influenced by nonobserved factors in neighboring parcels. Moreover, results confirm that interdependencies among land use alternatives, accounted for in the spatial multinomial probit model by the correlation structure, are required for the estimation of land use decision.