The Effect of Income on Religiousness

A majority of the world’s population claim adherence to a religion. But while there is a growing literature on the effects of religiousness on economic behaviour, we know very little about how economic factors affect religiousness. How income affects religious behaviour and affiliation is an especially important question for developing countries experiencing a rise in the average incomes of poor and lower middle class households. Whether these families will become more or less religious (or even change their religion) as they grow richer will have an impact on the way their societies develop.

That’s the opening paragraph of a new paper recently published in the July edition of the American Economic Journal: Applied Economics. Interestingly they found “significant effects of income on religiousness. Households that earn more go to church more often. Households that earn more are also more likely to be members of an evangelical community rather than of the mainstream catholic church.”

This outcome is contrary to my priors. My experience has made me think that the less money people have the more religious they tend to be. The explanation (in my mind) is that the less money you have the less capable you are to do things on your own. You need constant help from others and you are giving constant help to others. This sort of behavior breeds humility and orients an individual well towards a healthy spiritual relationship with God. On the flip side, more money breads a certain kind of hubris a “I can do it on my own” kind of attitude. 

The study uses a regression discontinuity design to establish a causal relationship between income and several measures of religiosity. For those unfamiliar with regression discontinuity, let me briefly explain (as it is quite a neat tool). In Ecuador there is a cash transfer program that gives poor households a monthly transfer of $35. What a regression discontinuity empirical strategy does is exploits the cut-off point where eligibility for the cash transfer program is determined. The purpose is quite intuitive: households on either side of the cut-off point are quite similar save for eligibility in the cash transfer program. Therefore any difference between households, on average, can be interpreted as an average treatment effect. See the figure below for the results of this study.

Of course, the usual caveats exist: This experiment was carried out in Ecuador, and the results may not hold everywhere. The results are likely very sensitive to definitions of “religiousness”. It’s not clear these effects will persist (in magnitude or direction) at higher levels of income. Nevertheless, here’s the last paragraph of this interesting paper:

Sociologists have long adhered to the so-called secularisation theory which posits that societies become less religious as they develop. This theory has come up for criticism, not least because of the continuing importance of religion in the US. Our results show that it is far from clear that higher income leads to lower religious participation. We should therefore not automatically expect other societies to follow the European example and become more secular as they grow richer. Rather, church membership and attendance seem to be similar to membership and participation in social clubs. They are costly in terms of time and money and, for the households in our sample, have a positive income elasticity.

HT: Bruce Wydick