Friday, May 05, 2006

The Price of Religion in the Family


I want to look at the effect of spouse’s income on the decision to convert to the same religion as the spouse. That is I would expect people to be more likely to convert in order to obtain a more desirable (higher earning) spouse. In this way the higher earner has greater bargaining power over choosing the religion of the family unit.

Similarly I expect people to be more likely to forgo preferences for a spouse of the same religion if the potential spouse is wealthy. Thus we should see that having a rich spouse would mean they are less likely to have the same religion.

Empirical Strategy:

For my primary hypothesis I will be regressing conversion to the same religion as spouse on spouse income and assets, differences between respondent and spouses incomes and assets, and controlling for age, gender, and religiousness of respondent and spouse as measured by services attendance. I would test the second hypothesis by regressing whether they have the same religion as spouse on the same dependent variables and controls.


My main concern is how I assume that someone converted for marriage if their current religion is different from the religion they were raised in, if their current religion is that of their spouse. However, there's clearly the possibility that they decided to convert and then found a spouse in that religion. I don’t feel this would bias my results in a particular direction, so I’m not worried about taking the alternate situation into account. It is also important to note that income and spouse’s income and assets are self reported by the respondent and susceptible to bias.


The hypothesis is nice an clean. There is a clear incentive and we are interested in discovering if, and by how much, people respond.

Your identification strategy is reasonable. The lack of precise conversion data is going to make it harder to be very convincing, but we'll just have to deal with that.

I think that there are some additional identification problems that you will need to think about. The ideal test of this would be something where we randomly match partners of different religions to each other and we see if those randomly assigned to someone wealthier are more likely to convert and stick it out.

Given that we self-select into relationships your estimate will not precisely capture the willingness to convert when your partner is rich. You'll want to think about how the self-section of partners will affect your estimates.
I think of the identification problems that you might want to think about is whether different relgions elicit more conversions. If a religion that requires conversion for marriage is highly correlated with wealth in your sample, then that might affect whether you can drawn a conclusion that income affecs conversion. In general, I would be worried about confounding factors such as characteristics of different religions, etc. Overall, I think this is a clear hypothesis, and the regressions are logical.
More on what Jimmy said...There are higher effort costs associated with converting to some religions versus others. Perhaps, you could do individual regressions where each one has the relationship between conversion to X religion and spouse's wealth. I'm also wondering if you can just include fixed spouse religion effects. ...Bryce?
Post a Comment

Subscribe to Post Comments [Atom]

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to Posts [Atom]