Using several different samples, model specifications, and variable proxies, this study revisits the role of social capital in generating life satisfaction (happiness). The main outcome of the exercise is that the parameter for social capital (generalized trust) is extremely fragile, and most estimates show little significant role of social capital in generating happiness. Six additional points are noted. First, the role of income seems generally positive and significant. Second, there are marked parametric differences between high-income and low-income subgroups, but it is difficult to say whether social capital or income is more important in either group. Third, significance of income inequality and of inflation vary considerably across the models and the samples, but their association with happiness is generally weak. Fourth, two measures of happiness (life satisfaction) yield similar sets of estimates. Fifth, there is some indication that transition economies are marked by lower happiness while Latin American countries are generally happier. Last, a reasonable test indicates absence of any significant specification error and mitigates worries about possible endogeneities.