This study try to identify the â-convergence process among regions in East Java using cross-section data of 30 regions for period 1983-2001, taking into account the presence of spatial heterogeneity and spatial dependence. Detection of spatial regimes using G-I* statistics on regional per capita GDP values in 1983 found cluster of high income regions (group of “rich”) in central & eastern part of East Java, and cluster of low income regions (group of “poor”) in western part. The result of OLS & GLS regression on absolute convergence model does not found any convergence process in East Java regional income. The convergence process is only found in spatial cross regressive absolute â-convergence model estimated for spatial club A (group of “rich”), but there is no evident for the same convergence process is happening in spatial club B (group of “poor”). Using the spatial cross-regressive model for absolute â-convergence this study founds that the coefficient of spatial lag of initial income (ô) is positive and significant in every equations. This shows how the spatial dependence has a significant contribution in explaining regional income growth in East Java. The positive and significant sign of (ô), means that the growth of a region is affected by initial income of its neighbors. The region which surrounded by wealthy neighbors will grow faster than the region surrounded by poor neighbors. The effect of neighbor’s initial income level to the growth of a region can be a result of technological or pecuniary spillovers. This will be the situation when technology or cost of production in a region depends not just on factors within the region but also on the level of technology in the neighbors (technology is embodied in in factors of production). These effects can be consider as supply-side externalities
