In this paper we argue that religion and welfare state spending are substitute mechanisms that insure individuals against adverse life events. As a result, individuals who are religious are predicted to prefer lower levels of social insurance than will individuals who are secular. To the extent policy outcomes reflect individual preferences, then countries with higher levels of religiosity should have lower levels of welfare state spending. In formalizing our argument we also suggest that if benefits from religion are subject to a network externality (I derive greater pleasure from religion when others are also religious), it is possible for countries that are similar in terms of underlying conditions to exhibit multiple equilibria with respect to religion and social insurance. We empirically test our predictions using individual-level data on religiosity, individual-level data on social insurance preferences, and cross-country data on social spending outcomes. The findings are strongly supportive of our hypotheses.