摘要：Poverty is still being one of big problems in Indonesia. Any efforts are done to find a solution for this problem. Poverty itself can be caused of the high unemployment that occurs. With a number of unemployment, it will be lower income thus reducing also purchasing power and the ability to meet the needs of life thus causing poverty. This study analyzed the impact of unemployment to the poverty as involving spatial factors, using spatial regression analysis. Used data on poverty and unemployment in each regency in the central java, the analysis shows that based on likelihood ratio test, obtained LR test value 6,038 or p-value 0,014001 which means there is a spatial correlation. By testing model simultaneously nor individually using Breusch-Pagan test and Wald test, it show that both are significant, with BP = 6,7094; df = 1; p-value = 0,009591 and Wald statistic = 7,0238; p-value = 0,0080434. The results means there are spatial element in the relations between unemployment and poverty in central java so that SEM is more proper used than ordinary linear regression. Keywords : Spatial Error Model (SEM), Spatial Autocorrelation, Spatial Heterogeneity
其他摘要：Poverty is still being one of big problems in Indonesia. Any efforts are done to find a solution for this problem. Poverty itself can be caused of the high unemployment that occurs. With a number of unemployment, it will be lower income thus reducing also purchasing power and the ability to meet the needs of life thus causing poverty. This study analyzed the impact of unemployment to the poverty as involving spatial factors, using spatial regression analysis. Used data on poverty and unemployment in each regency in the central java, the analysis shows that based on likelihood ratio test, obtained LR test value 6,038 or p-value 0,014001 which means there is a spatial correlation. By testing model simultaneously nor individually using Breusch-Pagan test and Wald test, it show that both are significant, with BP = 6,7094; df = 1; p-value = 0,009591 and Wald statistic = 7,0238; p-value = 0,0080434. The results means there are spatial element in the relations between unemployment and poverty in central java so that SEM is more proper used than ordinary linear regression. Keywords : Spatial Error Model (SEM), Spatial Autocorrelation, Spatial Heterogeneity
