Provincial Migration Dynamics in China: Borders, centripetal forces and trade (?)

This paper studies the degree of mobility and estimates the motion law of workers across China between 1985 and 1995 in light of the new economic geography. We rely on intra and inter-provincial migration data for 30 Chinese provinces to measure the importance of provincial borders as determinants of migration patterns. Densities of internal and inter-provincial migration flows are compared in order to derive estimates of border effects for migration in China. They provide a benchmark against which to evaluate the border results for goods over the same period. The study pays particular attention to the influence of access to market on workers? location choices. New economic geography models suggest that agglomeration process occurs through a self-reinforcing process because firms and workers are driven by a comparable determinant summarized by the regional market potential. In presence of scale economies and high transport costs, a province with high market potential attracts firms ensuring them a large demand and draws migrants through high real wages. We rely on a multi-regional model of economic geography based on Krugman (1992) to investigate the migration dynamics. Results confirm the overall small degree of mobility of workers between Chinese provinces. We find large negative effects of provincial borders and distance on migration. They vary greatly between provinces. These impediments appear to have shrink over the 1990s, in light with the easing of migration restrictions. The hierarchy of border effects by province is coherent with the literature on provincial economic dynamics. Economic size of the destination region turns out as a fundamental determinant of its attractiveness, confirming that migration patterns reveal centripetal dynamics. Estimations underline a self-reinforcing dynamic of industrial agglomeration. Workers move in order to get greater real wage. They leave small and peripheral regions (with low market potentials) to more central ones.