Our basic result is a conditional U‐shaped relation between happiness and religiosity. Countries with high levels of religiosity and countries with low levels of religiosity report higher levels of well‐being than countries with medium levels of religiosity, conditional on the level of income. We argue that this outcome reflects network and long‐run transition effects. P. 18.
That is from the conclusions of this new published paper by Gundlach and Opfinger (Review of Development Economics, August 2013). A draft (2011) is here.