Economic globalization since 1970'es have increased the importance of world cities in the global economy, and a global urban system has emerged. Economic development is clustered in metropolitan regions and their periphery, which have certain advantages for international trade, such as proximity and historical connections; leading to sustaining regional inequalities. In this context, core, semiperipheral and peripheral areas form a global economic system. Transnational companies and transformation of production systems into more flexible forms play key roles in this spatial restructuring processes. Turkey's integration to the global economy which gained impetus in the post 1970 period has likely lead to similar spatial developments. The Neoclassical regional growth theory foresees a decrease in regional inequalities among countries or regions in the long run, due to decreasing returns to capital.. According to Solow's growth model, initially poorer regions grow faster and catchup with richer regions. Endogenous growth theories reject the Neoclassical theory which assumes technology as a purely exogenous process, and propose that regional inequalities in productivity rates could be persistent due to local knowledge growth and diffusion, or, growth of human capital. Postneoclassical empirical findings furthermore show that productivity growth may not be limited to a region and spatial spillovers may lead to growth of productivity in neighbor regions. Empirical studies about Turkey provide evidence of productivity convergence and first order neighbor region effects on productivity growth in the pre-1995 period. However, it should not be thought that spillovers are effective only for first order neighbors in a country like Turkey, where sharp differences between Eastern and Western provinces are observed. Spatial spillover effects may span longer distances, as evident elsewhere in the world. Apart from such spillovers empirical studies on Turkey have also demonstrated that human capital has an effect on per capita productivity growth in Turkey in the pre1995 period. It is thought that this is still an important cause of productivity growth at the regional level, in the period 19902000. Three questions are asked regarding theoretical approaches and empirical findings on Turkey. Has regional inequalities in productivity rates decreased, or, in other words, is there beta convergence in Turkey during 1990-2000? Do long distance spatial spill-overs have an impact on regional growth? And last, but not the least: Did growth of human capital have an effect on regional growth? In order to answer these questions, three variables are identified and a conditional beta convergence model is specified. The model is in the form of a log-linear cross-sectional model where the dependent variable is the productivity growth per worker in a region. Furthermore, it incorporates neighborhood effects within 500 km. Particularly, the model incorporates the influence of initial productivity levels of neighbor regions on growth of productivity within a region. A preliminary Moran's I test demonstrates that there is significant but weak spatial autocorrelation between the initial per worker productivity levels of neighbor regions and the growth rate of per worker productivity in a region. The model is estimated by using an OLS estimation procedure. The result is an adjusted R2 of 0.39. Usual tests like Breusch-Pagan and White tests show that hetero-skedasticity is not a problem and Jarque Bera test is not significant. Therefore, the model is thought to be thoroughly specified. Findings suggest that initially poorer regions grow faster, pointing to convergence. The speed of convergence is about 1.89% per annum, which indicates a half-life of approximately 37 years. The result suggests that the inequalities between productive and less productive regions are decreasing slowly. As expected, initial productivity levels of neighbor regions have a positive impact on growth of productivity in a region. Therefore, it could be concluded that spatial spillovers reach far distances, leading to growth in semi-peripheral areas. The strongest impact on productivity growth belongs to the human capital variable. Excessive growth of per worker scientific and technical professionals is an important determinant of productivity growth. Although the results suggest conditional convergence, they also point that regional growth is stronger in regions around developed metropolitan regions. Regional inequalities are likely to persist between East and West for a long time. [ABSTRACT FROM AUTHOR]