A spatial economic model and spatial econometric analysis of population dynamics in Brazilian MCAs

To extend existing population growth models, this article proposes a theoretical setting including spatial interaction effects. Using data pertaining to 3659 Brazilian Minimum Comparable Areas (MCA) over the period 1970-2010, this extension is tested by estimating a dynamic spatial panel model. The authors also compare the performance of a wide range of potential neighborhood matrices using Bayesian posterior model probabilities. Six of the thirteen determinants of population growth considered produce significant spatial interaction effects. Moreover, five produce significant long-term spatial spillover effects; a mathematical analysis reveals the strength of this result, which requires more than a single parameter. Treating areas as independent entities, as many previous population growth studies have done, underestimates the impact of various policy measures.