For males, the college premium and the marriage premium are roughly equal. In the NLSY, for example, you earn 34% more if you’re a college grad, and 44% more if you’re a married male*:
Funny thing, though. When people – economists and non-economists alike – look at the size of the male marriage premium, they barely respond. On a personal level, that 44% premium doesn’t lead them to urge men to marry. On a policy level, the 44% premium probably wouldn’t even increase opposition to the marriage tax – much less inspire support for a massive government effort to encourage men to wed.
Why the discrepancy?
1. You could point out that (a) married women earn 10% less, and (b) more men can’t marry unless more women marry. But the male marriage bonus vastly exceeds the female marriage penalty. Indeed, the net premium for a couple almost exactly equals the college premium.
2. You could object that the marriage premium is largely selection rather than treatment. But like the college premium, the real story is probably that it’s a mix of both.
3. You could object that men fail to marry despite the high premium because they would hate being married. But you can say the same about school: Students give up because they find it super boring.
4. You could object that encouraging marriage restricts people’s freedom, but encouraging college doesn’t. But this makes no sense. If using taxes, subsidies, and regulations to make college more attractive doesn’t “restrict freedom,” why would using taxes, subsidies, and regulations to make marriage more attractive “restrict freedom”?
5. You could say that education has positive externalities, but marriage doesn’t. But this is irrelevant, because the people pushing college are focusing on the private return. In any case, the externalities of marriage are far less debatable than the externalities of education.
I could be missing something; if you think so, let me know. My considered judgment, though, is that the double standard is all too real. People should push both education and marriage – or neither.
* I interact married and number of children (CHILDNUM) with gender dummies to allow the effects of family status to vary by gender. AFQT is the NLSY’s IQ measure.
READER COMMENTS
Hume
Jan 23 2012 at 11:14am
It has nothing to do with income levels and income premiums. It has to do with autonomy and the value of freedom for individuals. Even libertarian/classical liberal Loren Lomasky recognizes the plausibility of a political community determining that education is among the ‘welfare goods’ children require as independent project pursuers. Although arguing against a state monopoly in education, Lomasky notes that a political community can make a plausible determination that “education is a crucial requisite of successful living as a project pursuer.” [Lomasky, Persons, Rights, and the Moral Community at 174].
It is entirely reasonable for a political community to determine that education, even a college education, is an important choice necessary for individuals-as-project-pursuers to lead valuable lives, while at the same time denying the same importance for marriage.
You should stop focusing on the ability to earn a high income as the end of all our aims. I do not understand how an academic nerd like you (and I mean that as a compliment) can not recognize the potential importance of education to all. Did you choose a life path based on the potential for a high income? Or were you more concerned about other values?
South Street
Jan 23 2012 at 11:21am
Correlation isn’t causation! In econometric terms, you have a problem with unobservable characteristics (which you seem to adress in 2., but according to the study you link to, only accounts for less than 10% of the marriage premium), but there still remains the issue of reverse causality: a lot of marriages never happen or break up because the man is economically unsuccessfull! Hence I suspect that a large part of the causality goes from earning to marriage, and not only vice versa.
Importantly, you do not have so much of this effect with education. In the vast majority of the cases, education comes first (between the age of 6 and 25, let’s say), and profession later, while the earning age/marital age coincide (let’s say from 20 – 70). The causal link from education to earnings is therefore much stronger than the causal link from marriage to earnings, and I suppose this is the (good) reason why people care more about the education premium.
South Street
Jan 23 2012 at 11:31am
little add on:
So the first question should instead be, “Is there really a marriage premium, and how large is it?” (which your regression does not answer).
Phil
Jan 23 2012 at 11:50am
Perhaps the difference, as “South Street” touched on, is that people think education *causes* increased income, but they don’t think that marriage does. They probably believe that the kind of personality that earns higher incomes is the same kind of personality that marries.
If you suggested to them that if a man gets married, he is likely to earn more, they would probably laugh.
PrometheeFeu
Jan 23 2012 at 12:04pm
My guess is that subsidies to college are not actually about encouraging people to go to college but rather about making college affordable for poor people. (Many subsidies to college are means-tested) But marriage costs almost nothing. A couple bucks for the marriage license. A trip to the courthouse during the lunch hour and bam! You’re married.
If you think about it this way, there is no paradox. It’s all equity-driven policy.
As for the marriage penalty I’ve recently conversed with tax accountants who assured me that in their experience, the marriage penalty has disappeared except for a few people here and there.
PrometheeFeu
Jan 23 2012 at 12:16pm
@Bryan Caplan:
2 possible explanations:
1) Causation runs the other way: a higher income makes you a more attractive mate.
2) Marriage is chaotic. You get married once and then you are married for a long time. As time passes your chances of being married increases. At the same time your income increases as you age.
Walter Wessels
Jan 23 2012 at 12:28pm
A college education creates an externality (of sorts): college educated people earn more income and pay relatively more tax. Subsidies are needed, up to a point, to internalize this externality. Marriage has the same effect (increase income) but depending on the education level, may not generate relatively more taxes (compared to the taxes that lower the opportunity cost of getting the education, that is, lower forgone wages due to taxes). But your main point, I suspect, is that much of the increase in earnings reflects selectivity (more able get more education and are attractive marital partners). I would argue that a spouse and an employer are both equally able to identify those who are more able, reducing the role of selectivity in education and increasing it in marriage.
GD
Jan 23 2012 at 12:30pm
“But marriage costs almost nothing.”
You must be single…
Vipul Naik
Jan 23 2012 at 12:30pm
You write:
The average of 44% and -10% is 17%, not 34%. So the college premium is twice the marriage premium, assuming that the pre-marriage earnings of men and women are equal.
Tom
Jan 23 2012 at 12:42pm
“If you suggested to them that if a man gets married, he is likely to earn more, they would probably laugh.”
Doubtful. It was (years ago) a common practice to hire married guys as salesmen. They, having more bills to pay, had greater incentive to make more money.
John Thacker
Jan 23 2012 at 12:49pm
@Hume:
Err, why? You seem to be arguing by assertion. Your next paragraph makes it clear that it’s only about imposing your “values” on everyone else. Did you write this as a parody of the pro-college side to discredit it?
@PrometheeFeu:
It is the case that “Many subsidies to college are means-tested.” But many are not. And the means-tested subsidies that go to the poor are far more controversial in politics (even among Democrats) than the subsidies to the middle and upper middle class, like government student loans.
@Vipul Naik:
Your math is only correct if the average earnings of unmarried men and women are the same. Are they?
John Thacker
Jan 23 2012 at 12:52pm
@Hume:
Education may be necessary for people to maximize themselves. But is schooling? There are forms of education and life experiences not found in a classroom. Don’t you value autonomy and the freedom for individuals? Or do you insist that everyone share exactly your same values and want to do things in the same way?
Your conception of education is as stunted as those who people that people cannot possibly be complete outside a marriage, in my view.
Hand
Jan 23 2012 at 1:10pm
If you have fewer bills so you work less and make 44% less are you necessarily less happy? From a welfare standpoint I’d think not.
Hume
Jan 23 2012 at 2:03pm
@John Thacker
“You seem to be arguing by assertion.”
Perhaps, perhaps not. This is a comments section to an economics blog and I was noting the philosophical work of a splendid libertarian/classical liberal. I have spent some time around here before making long arguments, but this one cannot be done in the space available. As a result, I will simply point to Lomasky’s Persons, Rights, and the Moral Community. I am not necessarily advocating those views, only bringing them up to show the lack of depth in Caplan’s post.
“Your next paragraph makes it clear that it’s only about imposing your “values” on everyone else. Did you write this as a parody of the pro-college side to discredit it?”
I do not. Again, I was noting a work done by a libertarian philosopher. Unfortunately, most pop libertarians prefer to scream “rights!” or “liberty” or “freedom!” without providing the moral foundations for such a view. As noted by Lomasky, “[r]ights without foundations are treacherous entities. … Rights are so easy to claim, but so terribly difficult to justify. Naked appeals to intuition or moral insight too often supplant analysis, and, not surprisingly, one person’s right is another’s fantasy. The result can be pleasing only to the moral skeptic.” Lomasky, Persons, Rights, and the Moral Community vii (1987).
This is Lomasky’s project, to answer Nagel’s (sometimes unfair) objections to Nozick. But in the process of explicating the ethical basis of rights and freedom, you will realize things are not so simple as pop libertarians make it seem. This is because you will realize that one is necessarily invoking values when you make rights claims or claims in the name of freedom. Thus, your assertion that I was simply imposing my values on everyone else colors all the waters we are swimming in. Most importantly, the libertarian theories of property rights, contract rights, torts, etc. are derived from certain fundamental moral principles and values, but they do not provide for an all-encompassing, self-contained, neatly packaged institutional system with no room for reasonable disagreement. And the moral principles that entail the necessity of moral rights may also entail certain conclusions regarding educational provisions and the educational system.
This is all very off-the-cuff. This is a comments thread.
Ray
Jan 23 2012 at 4:49pm
Bryan, there is another “explanation” for the discrepancy. Education is INTENDED to increase productivity at work, and therefore income. Marriage isn’t. And most people cannot adequately separate intention from effect.
This is related to why some people don’t like the argument that free markets help the poor. They aren’t INTENDED to help the poor.
Ray
Jan 23 2012 at 4:55pm
To follow up my previous comment:
Intention effect also explains why people have a hard time with the idea that Obamacare won’t improve medical care. It’s INTENDED to solve the health care problems! Same with minimum wage not increasing workers’ wages.
Mr. Econotarian
Jan 23 2012 at 5:18pm
1) My wife has fixed me up and made me far more presentable and efficient at working than college or grad school ever did.
2) Unemployment is different for married guys, because you can’t really sit around all day and watch TV because you’ll be put on chores or encouraged to go find a job by your spouse.
MS
Jan 23 2012 at 5:19pm
This post is, well, strange. There are plenty of variables that, in a naive and simple regression, will show up as having a ”wage premium”. Some we believe are more due to causal effects than selection, and for some we think the opposite. For most people, education goes more into the former category and marriage more into the latter. But there are 25+ other variables that we make the same distinction for as well.
Hence, there is nothing special or discriminatory in the view that (many of us hold) the “marriage wage premium” is mostly due to selection. There are many other variables that are linked to higher wages for which we also view them as mostly due to selection.
It of course linked to that there is a potential theoretical mechanism for why education would have a larger causal mechanism on wage that is rather intuitive and easy to believe (true or not). But, there are also several empirical papers that can be used to indicate support for such a view (e.g. there are also papers indicating a causal positive effect on IQ from more education, as well as well-conducted papers indicating non-cognitive (causal) returns from education).
It is more difficult to believe in the (potential) theoretical mechanisms for why marriage causally increases wages (to a significant degree – esp. in terms of the magnitudes from your naïve regression). There are definitely fewer empirical papers that can be used in support for such a view as well (even though they are not non-existent).
OneEyedMan
Jan 23 2012 at 5:34pm
Vipul Naik is right that percentages do not work the way Caplan roughly suggests they do.
However, Men earn more than women do, to the positive marriage premium for men has a larger effect on family income than the female decline.
Median male income is about $47k and female about $36k. Unmarried, the two of them make $83k. With the premiums above, the two of them married make 100k or about 21% more. Still smaller than the college premium but closer.
Glen S. McGhee
Jan 23 2012 at 6:24pm
PrometheeFeu could be right:
1) Causation runs the other way: a higher income makes you a more attractive mate.
Earning potential, like wealth, is an asset in the marriage market.
These are social institutions you are talking about, right?
John Thacker
Jan 23 2012 at 8:09pm
@Hume:
My objection stands. You provide a quote from a philosopher about education. What Bryan is complaining about is schooling. It’s a poor and stunted view of the world that thinks that schooling is the only kind of education in the world. Insisting on it shortchanges people, and their right to develop as a person.
Hume
Jan 23 2012 at 11:14pm
@John Thacker
“Insisting on it shortchanges people, and their right to develop as a person.”
I am not insisting on anything. In my original post I wrote “It is entirely reasonable for a political community to determine that education, even a college education, is an important choice necessary for individuals-as-project-pursuers to lead valuable lives”. Notice that I referenced a choice that is necessary. This does not entail requiring individuals to attend school; rather, it suggests that a choice must be provided.
“It’s a poor and stunted view of the world that thinks that schooling is the only kind of education in the world.”
That’s great, but you are now insisting on a competing interpretation of what is morally valuable and required for freedom to have any moral standing (thus reinforcing my original objection to Caplan’s shallow analysis of the importance of education as relating only to income generation). However, it is possible that other people living within the political community disagree, arguing that “schooling” is necessary for freedom to have any value, and thus a choice must be given to all if we are to insist that freedom is a value for all. Unless you profess to transcend the realities of the human condition and our limited cognitive capacities and knowledge, you will recognize the potential fallibility of your own position. This does not mean you should give up in setting forth a conception of justice and freedom; rather, it means you ought to recognize that reasonable people can and do disagree.
To be clear, I am not insisting on this position, only that it is reasonable and consistent even among classical liberals/libertarians (hence my citation to Lomasky). There is nothing inherently unreasonable or incompatible with classical liberalism to suggest that schooling is fundamentally important to obtaining an education, and as such, is fundamentally important for individuals-as-project pursuers. In consequence, it is not unreasonable or incompatible with classical liberalism to suggest that a choice must be provided in order to respect the freedom of all. Again, I am not arguing for this position, only noting that Caplan’s analysis does not cover this argument, an argument that many people who seek the further encouragement Caplan dispises.
Jack
Jan 24 2012 at 8:17am
Prof. Caplan: Intuition tells us that education causes higher incomes, but that higher incomes cause marriage. That’s all. It very well may be wrong (use IV, instruments, etc.) but I think the (incorrect?) story is that simple. Priors are hard to shake off.
Tom
Jan 24 2012 at 8:21am
My gut instinct is that you’ve got the causality the wrong way round. Surely men who are doing well and earning more are not only more likely to “attract” a partner but are also more able to afford the very expensive ceremony that is a wedding.
PrometheeFeu
Jan 24 2012 at 11:00am
@Everyone who says weddings are expensive:
While you indeed can blow a ton of money on a wedding, you don’t actually have to do so. I’m not at all convinced that poor people choose not to marry because they can’t afford some sort of gigantic ceremony.
Roger Sweeny
Jan 24 2012 at 12:36pm
Nice coincidence. I mentioned this just last week in my high school physics class, in the context of hypothesis testing. If you run a controlled experiment, you have two groups that only differ by the variable you are testing. If you’re using data from something that happened naturally, you have to break up your population into two groups that differ only on that one variable. If that can’t be done–and it usually can’t–you try to “control” using statistical techniques. But then you have to be sure you know all the relevant differences, and have data on those difference, and you almost certainly don’t.
I mentioned the marriage studies first, and everyone agreed that people who get married are different (on average) than people who don’t. They are probably easier to get along with, more reliable, have better economic prospects, etc. So the fact that married people make more money may have little to do with marriage itself and much to do with the kind of people who get married. Students seemed to think that government actions encouraging marriage would not have the effect of making the country richer.
I then mentioned the college studies. I suggested that they had the same problem as the marriage studies. People who get a college diploma are very different from people who don’t. They are harder working, more able to resist temptation, smarter, a whole boat load of things. If college didn’t exist, they almost certainly would be more successful than the people who presently don’t get a degree.
Roger Sweeny
Jan 24 2012 at 12:40pm
Hume,
As a teacher of high school students, I can assure that many, many high school students do not feel that they are free not to go to college or that by going to college, they are exercising autonomy. They have been told for years, “If you don’t go to college, you won’t get a good job. So you better go to college.” Legally, an employer can reject you if you don’t have a diploma, or offer you a job because you do, even if the diploma has nothing directly to do with the job, or even if the job doesn’t require any skills beyond high school.
Collin
Jan 24 2012 at 10:19pm
Am I correct in understanding that the higher number of children DECREASES the income?
What are the implications of that in the modern economy? Has the job market gotten so competitive that it economically hurts families to have children?
Hasdrubal
Jan 25 2012 at 2:17pm
I think the number of children is inversely related to mothers’ education, and since education is highly correlated with income, children are naturally negatively correlated with income.
There’s also the fact that immigrants tend to have more children than natural citizens, and immigrants tend to be poorer than average. Another negative correlation between children and income.
Finally, this is a worldwide trend: As GDP per capita increases, the birthrate tends to fall everywhere from Italy to Japan to sub-Saharan Africa.
Glen S. McGhee
Jan 30 2012 at 9:11am
What if college cuts into marriageability, as it apparently does for the disadvantaged? The question is, why it does not for the rest.
http://www.news.cornell.edu/stories/Jan12/CollegeMarriage.html
Comments are closed.