A Space-Time Model of Fertility and Development in China, 1982-2000

Katherine King, University of Michigan

Although China is extremely regionally diverse, and fertility policy is implemented at a local level, research often implicitly assumes spatial homogeneity. This paper applies exploratory spatial data analysis (ESDA) and spatial panel regression models to examine county-level variation in fertility rates in China. Spatial regressions reduce bias resulting from spatial and temporal autocorrelation in regression models, while spatial analysis allows detailed examination of relationships between pairs of predictors. Using county-level data from China’s 1982, 1990, and 2000 censuses, I model general fertility rates and changes in rates using variables related to social and economic development characteristics of the counties.