Spatial econometric models for simultaneous systems: Application to rural community growth in France

In this article, comparisons are made of several spatial econometric approaches to estimation of multiequation models of small region development applied to rural community growth. Spatial extensions of the Carlino and Mills's 1987 and Boarnet's 1994 models are estimated to analyze the spread of population and employment into 3,500 rural communes in six French regions. Results are compared for the Henry Barkley, and Bao's 1997 extension of the Boarnet model, the Carlino-Mills and Boarnet models, and three spatial autoregressive models suggested by Rey and Boarnet in 2001. Tests for spread and backwash effects in the spatial autoregressive model, and the Carlino-Mills and Boarnet models, with spatial autoregressive terms added, indicate that population growth spreads to rural communities from nearby areas but that evidence on employment spread is less robust. The Henry et al. modification to Boarnet adds insight into how urban growth affects proximate rural areas by decomposing the spatial cross-regressive term into rural area, urban core, and urban fringe effects.