Some studies have found that church attendance is negatively associated with economic growth and that belief in hell or heaven is positively associated with economic growth. The implication of this assertion is that 'believing rather than belonging' is what matters for economic performance as far as religion is concerned. Using logistic regression model on the data set of a metropolitan area, we have provided another evidence to further the empirical scholarship on the effect of religion on economic performance .We showed that church attendance perversely affects savings but belief may not matter significantly for savings decision of household after we have controlled for liquidity constraint, family ties and income uncertainty. We found that savings is negatively related to church attendance but belief in hell does not matter for household savings.