Purpose of the present thesis is to improve the shortcomings of the empirical literature on regional social capital and investigate the following three less addressed research questions; i. Do regional social capital and economic growth have an endogenous circular relationship? ii. What are the determinants behind cross-regional variation in social capital? iii. Are there any significant spatial spillovers of social capital across neighboring regions? Empirically, the social capital (for 81 Turkish provinces in 2015) was measured by using three indicators; social trust, norms, and participation. In terms of methodology, descriptive statistics, illustrative maps, Kernel density estimates, Jarque- Bera Normality tests, cross-sectional OLS, 3SLS and Spatial Error Model (SEM) models were adopted. As an outcome, the results of the thesis can be summarized into four groups. First, regional social capital is heterogeneously distributed across regions, particularly in social participation. Second, having estimated the models, it is understood that social capital has no significant impact on economic growth but the growth induces significantly the generation of social capital (only in types of social norms). This represents the first result of reserve causality that has not yet been considered by the literature. Third, crossregional variation in social capital is best explained by robustly significant economic and demographic determinants. Hence, a typical province that has high social capital can be defined as an Anatolian province with a relatively high-income level, low unemployment and poverty rate, big households and older age profile. Fourth, social capital is shown to emerge in spatially correlated clusters.