Study on the county-level economic spatial-temporal disparity in Xinjiang

The issue of regional economic disparity has attracted considerable scholarly attention. Unfortunately the conventional measures of regional economic disparity mask geographical clustering. Based on the social-economic data from 1978 to 2004 in Xinjiang, this paper utilizes coefficient of variation (01), spatial statistical and GIS techniques to analyze changing county-level spatial-temporal patterns of regional economic disparity in Xinjiang. CVs for inter-county disparity do not exhibit an invert U-shaped pattern from 1978 to 2004 in Xinjiang. It declined in the 1980s, but has risen substantially since 1990. We also compared the temporal change of the spatial autocorrelation in Xinjiang, and found that there is an obviously temporal increase of Moran's I since 1978 to 2004. That is, there is a dramatic increase of the Xinjiang per capita GDP's spatial clustering in the last 20 years. This means that the disparity between rich clusters and poor is enlarging in the same periods. Moran's I scatter plots and LISA (Local Indicators of Spatial Association) cluster maps were used to test the local pattern of the Xinjiang economic development. The results of local statistic show that the two types of clusters (High-High and Low-Low) are increasing which means that the heterogeneous is increasing too.