This paper summarize about fifteen years of empirical research on various spatial levels on the impact of social capital on economic growth. At the lowest level, the firm level, the results are unambiguous: there is strong evidence on the impact of social capital on firms’ performance. However, when the unit of research is moved from single actors to spatial units with a large number of anonymous actors, the results become less clear. The conclusion of this paper is that the contradictory results of studies on national and regional levels at least partly can be explained by insufficient measures of social capital’s two main component parts, i.e. social networks and the norms and values that are being distributed in them. “Trust in other persons” is only one of several values that according to the literature would influence economic variables. Associations of the civil society are just a small fraction of all the social networks of human interaction. To develop measures for values like creativity, entrepreneurship and tolerance, and to find better measures for social networks is accordingly the main challenges for future research on social capital’s impact on the economy.