Monday, April 21, 2008

Divorce in the SWF: Part II

Shortly after my previous post on Divorce in the Social Welfare Function, a relevant NBER working paper has appeared. That earlier post was a result of a publicized study on behalf of a number of groups concerned about the tax effects of single parent families on children. Just in time for this study is the following NBER working paper by Kieth Findlay and David Neumark titled "Is Marriage Always Good for Children? Evidence from Families Affected by Incarceration." From the abstract:
Given that changes in the availability of men in the marriage market should affect marriage decisions, we use incarceration rates for men as an instrumental variable for family structure in estimating the effect of never-married motherhood on the likelihood that children drop out of high school, focusing on blacks and Hispanics. Instrumental variables estimates suggest that unobserved factors rather than a causal effect drive the negative relationship between never-married motherhood and child outcomes for blacks and Hispanics, at least for the children of women whose marriage decisions are most affected by variation in incarceration rates for men. For Hispanics, in particular, we find evidence that these children may actually be better off living with a never-married mother. [Emphasis provided by J. Ross]
I expect for this paper to carry no influence on the sponsoring child advocacy groups. As much as I like the result, I'll take it with a grain of salt.

1 comment:

spmosher said...

Yes, a grain of salt indeed.So clearly, we need more married families to cut the current costs of single headed familes and unwed mothers.

How about Gay Marriage, Gay Adoption and advocacy of both! We could take up a lot of slack there.

But they won't see that as a solution. The study was clearly undertaken in order to show how we have become immoral, and should return back family values.

Blech! While I agree with some of the study, they are making some huge assumptions, most of which support their findings... But of course.