In the course of commenting on our article on religion, income, and voting, David Beckworth links to this interesting paper on religiosity and the business cycle:

Mainline Protestant denominations–which tend to have higher income earners–do well in terms of growth during economic booms while evangelical Protestants denominations–which tend to have lower income earners–actually struggle. (During economic downturns the outcomes are reversed–evangelicals Protestant denominations thrive.) In general, I [Beckworth] find mainline Protestants to have a strong procyclical component to their religiosity while evangelicals have a strong countercyclical component. These findings can be explained by again appealing to the labor-leisure choice explained by economic theory.

I haven’t had a chance to really look this over, but it seems important, and it reminds me of Robert Putnam’s comment that, although we think of religious attendance and denomination as fixed demographic descriptors of people, it’s pretty common for people to change denominations–even to switch between Protestant and Catholic. Putnam also said there was evidence that people switch religions to match their political beliefs.

Regarding Beckworth’s paper itself, what it really needs (from my perspective) are some graphs that directly map the data to the findings. Regressions are great, but I need some scatterplots to really be convinced. And then the challenge is to map the graphs to the regression estimates. This takes work–sometimes a lot of work–but the payoff is a new level of confidence building, a step beyond mere statistical significance.

P.S. Minor comments: the axes are overlabeled–this is standard. On figure 1, the y-axis should be labeled as 4%, 5%, 6%; in Figure 2, the y-axes should go down to 0 and be labeled 0, 5000, 10000, 15000; and the x-axes should be labeled every 10 or 20 years without those tick marks. Figures 3 and 4 are unreadable, and Figure 5 is unclear. I can’t figure out what’s going on there. A long caption would help. Umm…I don’t have lots of comments on the tables since I prefer graphs, but I could make the usual comments, e.g., “42.01%” should be “42%”, etc. And in Table 3, the denominations should be listed in a natural order, not alphabetically. Some possible orders are: decreasing size order, or ordering from most liberal to most conservative.

The bottom line, though, is that this is interesting stuff.

P.P.S. To return to Beckworth’s comments on our article, he quotes us as saying, “economic concerns are more important in poorer areas, with social and religious issues mattering more among the rich” and then writes,

If one assumes that voting for moral and cultural reasons also implies other time commitments to these issues (staying informed by reading and watching TV, discussing issues with friends, supporting rallies, etc.), then these results could also be interpreted by using what economic theory tell us about the labor-leisure choice via the substitution and income effects. For those individuals in the lower income group, it is likely that the substitution effect–which says the opportunity costs of social activism, forgone earnings, is too high–dominates. Their time is better spent working than worrying about social issues. On the other hand, those individuals in higher income groups most likely have the income effect–which says they can afford leisure activities like social activism–dominate. Simply, as individuals become richer they can afford to become more engaged in these social issues, if that is what they want.

This makes sense to me. I don’t see this as an alternative explanation of the post-materialism idea of Inglehart and others; rather, I assume this is a standard part of the story: less concern over the necessities of life, more time to think about other things.