Not sure I like the easy assumption here that "panel study" = causality. Suppose I get happier. Then I go to church more. The use of some complicated-ass Grange

Not sure I like the easy assumption here that "panel study" = causality. Suppose I get happier. Then I go to church more. The use of some complicated-ass Granger-causality-style technique also makes me trust it less. This natural experiment looks more interesting, although it exploits just a temporary shock (variation in the length of Ramdan): https://academic.oup.com/qje/article/130/2/615/2330341?login=true https://academic.oup.com/qje/article/130/2/615/2330341?login... "We study the economic effects of religious practices in the context of the observance of Ramadan fasting, one of the central tenets of Islam. To establish causality, we exploit variation in the length of daily fasting due to the interaction between the rotating Islamic calendar and a country’s latitude. We report two key, quantitatively meaningful results: (i) longer Ramadan fasting has a negative effect on output growth in Muslim countries, and (ii) it increases subjective well-being among Muslims. We find evidence that these patterns are consistent with a standard club good explanation for the emergence of costly religious practices: increased strictness of fasting screens out the less committed members, while the more committed respond with an increase in their relative levels of participation."