« Is the American economy taking too much risk? | Main | A superb post on exchange rates »
Why arranged marriage is costlier than you might think
...when parents are involved in mate choice, sons are significantly less likely to marry college-educated women and women engaged in the labor force, after controlling for individual and family characteristics. I show that these effects are driven, at least in part, by parental preferences and cannot entirely be attributed to correlation between arranged marriages and unobserved characteristics. These results suggest that lowering the incentive for parental control in mate choice may improve investments in women's human capital in India.
That's from Divya Mathur; here is the paper, she is on the job market this year from Chicago. I take the implicit model to be that parents want a wife who will obey her in-laws. Sons want wives who will earn some money and be more interesting to talk to. Put the son in charge and the supply of potential mates responds accordingly.
Posted by Tyler Cowen on November 9, 2007 at 06:55 AM in Education | Permalink
Comments
A classic chicken and egg problem. Woman human capital investment ->lowered incentives for
parental control-> Woman human capital investment. I think the new and better equilibrium will
come via rapid urbanization and successful demonstration effects in otherwise archaic minded
communities (something like "my neighbour's daugther married a smart, handsone man without
any dowry, all because she was herself smart and earning good money"). Already happening.
Prof Cowe, have you considered installing a system whereby the comments on this blog can be
ranked/rated.Most of the posts and comments here are quite interesting and I will like
such a feature.
Posted by: sa at Nov 9, 2007 7:25:24 AM
I disagree with your statement. I'm Indian, and every arranged marriage that I've known usually involves a party who has married "up". Obviously this is based on my own experience, but everyone I've know who has gone the arranged route, usually ended up with someone who was either a doctor, nurse, or accountant. Definitely someone who has a college degree. The article is right in that it is based on parental preferences, but parents prefer their kids marry someone who will be able to provide for them and their families. Its common sense, who would your parents have liked more - the college graduate or the obedient college dropout?
Posted by: Ted at Nov 9, 2007 7:29:00 AM
Attempts to convince these people to embrace the idea of freedom for mate choice are always noble. But I am not sure that the economic route is the most persuasive one. The argument should be a moral one: people should be entitled to decide whom they want to marry to/live with. Besides, she seems to be hesitating in her conclusion ("may improve..."), which unavoidably makes people to wonder whether this is just a leap of faith.
Posted by: Sam at Nov 9, 2007 7:37:36 AM
Sam,
Problem is, people consider "You can't just do what you feel like doing", "It's selfish not to take the wishes of your family into account", and "this is an important tradition we can't just discard because you feel like it" to be moral arguments too. Even in American society, people moan nonstop about how bad excessive individualism is, and putting what other people want ahead of what you want, regardless of whether the actual consequences are even good at all, is seen as virtuous.
Posted by: LisaMarie at Nov 9, 2007 9:02:40 AM
Doesn't Ted's comment actually support Tyler's conclusion? What am I missing?
Posted by: steve at Nov 9, 2007 9:37:28 AM
tyler,
the issue for young men may not be one of /preference/ but of /exposure/. at work, school, and leisure, young people tend to be exposed to people of like human capital so even if young men were indifferent to female HC as a matter of preference they still might end up marrying more educated women than their parents might have chosen because these are the sorts of women they tend to meet in everyday life. (i think the parents probably do have a preference for low HC daughter-in-laws as this implies high fertility).
Posted by: Gabriel Rossman at Nov 9, 2007 10:44:09 AM
Ted - your intuition matches mine. Wouldn't you expect positive assortative matching to break down if parents are doing the matching? Unless they are purely altruistic that is. Parents may match for various objectives, and are probably responsive to financial and political incentives.
Posted by: jason voorhees at Nov 9, 2007 11:29:22 AM
Doesn't the caste system negate this argument?
Posted by: 8 at Nov 9, 2007 12:28:11 PM
Obey the inlaws? There are a lot of other preferences that could express this. How about: more likelihood that the wife would physically and emotionally support inlaws in their old age, less likelihood of the wife divorcing their son, increased likelihood of more grandchildren, etc.
Posted by: Allison at Nov 9, 2007 12:40:23 PM
A peek at the paper shows that the author focuses on middle class, urban families, so it's not a caste thing. My question to Tyler is how the decision making power would transfer to the son? The parents are making a decision (via their child) to maximize their future returns (household services, income, grandchildren), with the "cost"=how pissed their son is with the arrangement. Where cost is again tempered by the degree of power the parents have over the son. The power to choose a mate would only be transferred if the parents have little power over the son, and the son has strong preferences that differ from the parents.
Posted by: v at Nov 9, 2007 1:14:12 PM
Ted, two points:
1) Given that you're reading this blog, you're probably above average socio-economic status - there is a bias here. You're more likely to know the doctors, etc.
2) For every marriage where someone "marries up", the person they're marrying is "marrying down" by definition.
-AK
Posted by: AK at Nov 9, 2007 1:17:53 PM
I agree with Allison. I'd need to see some kind of evidence before I accept the "implicit model."
I think Allison's proposed models are more likely and most likely is a combination of motivations. I doubt the desire for a daughter-in-law that's obedient to her in-laws is the dominant factor.
I could be wrong. But where's the evidence?
(And your model for the sons needs some help, too. How many 20-year-old males do you know of that look for earning potential in a wife?)
Posted by: Bob Montgomery at Nov 9, 2007 2:35:05 PM
Ted's comment makes no sense. The marriages in this study do involve someone marying up (and, by logical necessity, their partner marrying down). What the study says is that the women is more likely to be the one marrying up when the marriage is arranged than when it is not arranged. To be clear, this is a statement about relative, not absolute, frequencies. It might be that 90% of the time men marry up, but still they could be more likely to marry down in an arranged marriage than in an unarranged one. Said another way: this is a study about what direction one particular factor pushes the outcome, not where the average outcome lands after all factors have been accounted for.
Posted by: David Wright at Nov 9, 2007 2:58:40 PM
Given the way de facto polygamy is making it so much more difficult for Beta and Omega men to find wives the "normal" way here in America, could arranged marriage be the next new thing? Stranger stuff has happened.
Posted by: Peter at Nov 9, 2007 3:00:34 PM
de facto polygamy
Would you mind elaborating further on this point?
Posted by: mpowell at Nov 9, 2007 3:12:47 PM
How about this: the parents want grandchildren. They know that educated and working women tend to bear fewer offspring. To maximize the number of grandchildren, they prefer their sons to marry less educated (more fecund) wives.
How to test? Well, for one thing, ask whether the inverse education/fertility relationship obtains in the subject population. For another thing, survey parents on the qualities they seek in brides for their sons.
Posted by: Mark Seecof at Nov 9, 2007 3:15:28 PM
Wow, a female, Indian version of Justin Wolfers!
Posted by: howiewu at Nov 9, 2007 3:27:06 PM
A peek at the paper shows that the author focuses on middle class, urban families, so it's not a caste thing.
But if everyone is in the same caste, isn't the variance between women in the caste smaller? To use an American example, maybe the upper class parents would want their son to marry a middle-lower class farm raised woman without education. But is a woman like that even available in the upper castes?
Posted by: 8 at Nov 9, 2007 3:55:10 PM
De facto polygamy comes in active and passive versions. The active version occurs when a man is romantically involved with two (or even more) women at the same time, the classic example is the high-powered Alpha executive who is married and keeps a mistress. Given the relatively even numbers of men and women in the prime marriageable ages, when some men take two or more women "off the market" they skew this balance.
Passive de facto polygamy occurs when a woman hopes (reasonably or otherwise) that an Alpha will come in and sweep her off her feet, and in the meantime rebuffs any attempts by lesser men to express interest in her. Classic nightclub behavior. Eventually many of these women come to their senses and will settle for a Beta or even an Omega, but not all do.
Posted by: Peter at Nov 9, 2007 4:02:28 PM
By the way, for those puzzled by the terms Alpha, Beta and Omega as applied to men, here is a test of male "dating market value" that should make things clear. Pretty amusing, too. You can find a women's test on the same site, though I won't link directly to it as it's a little NSFW.
Posted by: Peter at Nov 9, 2007 4:07:24 PM
Maybe if you need your parents' help to find a mate, you're just not that great a catch.
Posted by: Keith at Nov 9, 2007 4:55:49 PM
Peter: how interesting.
But regarding arranged marriages, the divorce rate speaks to the success of choosing our own mates.
Posted by: Jane2 at Nov 9, 2007 5:28:54 PM
when parents are involved in mate choice, sons are significantly less likely to marry college-educated women and women engaged in the labor force, after controlling for individual and family characteristics.
I guess that might be true on an aggregate basis. And it might have been true of the arranged marriages I am familiar with three generations back. But for the modern arranged marriages I'm aware of, this doesn't seem to be true at all. Education, at least, is one of the things that stands out as being particularly important as a qualification. Almost make-or-break, frankly. I can't imagine any of the elders among my relations signing off on someone marrying a woman who hadn't gone to college. At the very least.
Labour force participation . . . less so, but again, at least prior to the first children, all the wives in the arranged marriages I know of had jobs. I don't think anyone would exercise a "veto" over the issue, though.
Of course, I'm thinking of Korean arranged marriages, rather than Indian or Pakistani arranged marriages, so there may be significant cross-cultural variation at work. The household resources issue doesn't seem likely to be as significant for Koreans, because the married couple are unlikely to move into (and contribute to the budget of) the family home after they marry. And Korean society is fanatical about education, even for woman, so an untutored daughter-in-law would probably be a kind of social disgrace.
Regarding AK's comment:
1) Given that you're reading this blog, you're probably above average socio-economic status - there is a bias here. You're more likely to know the doctors, etc.
Maybe . . . but doesn't this comparison purport to correct for SES? I can imagine a situation in which wealthy urban clans tend to emphasize educated wives, and poorer rural clans tend to choose uneducated wives, but that doesn't seem to be what is reported.
Posted by: Taeyoung at Nov 9, 2007 6:00:34 PM
"These results suggest that lowering the incentive for parental control in mate choice may improve investments in women's human capital in India."
what about the alternative possibility that the sons would prefer good-looking women who are less qualified than they are? Ray Fisman's study might suggest that - http://www.slate.com/id/2177637/fr/flyout . In that case would you expect 'improved' investments in women's grooming products in India, including weight-loss treatments and plastic surgery?
Posted by: joe at Nov 10, 2007 1:00:27 AM
Aren't some of the folks here, including TC, making a value judgement here? It is not obvious that a college educated female, is superior to one who isn't, from a male perspective. Perhaps arranged marriages provide males with an opportunity to choose from a wider range of females, versus the narrower population of college girls that they might otherwise meet.
In other words, arranged marriages expand the female candidate pool and let males make a more optimal choice (for them) which is, on average, less educated.
The preference of males for lower status females is a remarkably consistent theme across geography, time, and culture. The Koran addresses this issue. However, the idea is hardly unique to Islamic society.
Posted by: Peter Schaeffer at Nov 10, 2007 9:09:40 PM
