Estimating the college wealth premium: Not so easy

Dale Lehman writes:

Here’s the article referenced on Marginal Revolution today. I thought it might be of interest and worth blogging about. It is quite thorough and fairly complex. The results are quite striking – and important. My big concern relates to a critical variable – financial literacy. On page 14 they claim that it is an exogenous variable to education. It is defined in Appendix A. I don’t quite see how it is clearly exogenous to education. I would think that college educated (and postgraduate educated) people are more likely to feel confident in their financial literacy. So, if including this “exogenous” variable eliminates most or all of the wealth effect of education, isn’t it possible that it is just measuring that effect and not controlling (I know you don’t like that word – perhaps we could say adjusting for) it?

My quick thought was that it’s a political thing. The linked post highlights the following claim:

Among non-Hispanic white family heads born in the 1980s, the college wealth premium is at a historic low; among all other races and ethnicities, it is statistically indistinguishable from zero [boldface added by Cowen].

Cowen is conservative and education is a liberal institution, hence he’s supportive of studies that disparage education. This is not to say that the claim is wrong, just that I’m guessing that the conclusion is motivating the post.

But let’s get to the statistical issue that Lehman raises. I agree that adjusting for “financial acumen” seems a bit of an all-else-equal fallacy.

A related question is whether we would expect the college wealth premium to be positive.

Just thinking off the top of my head—the last economics class I took was in 11th grade—I’d think that, if you adjust for enough things, the college wealth premium should be negative. To put it another way, if you decide not to go to college, you should get some money in return. It won’t work this way in practice, because not everyone has the opportunity to go to college—like so many other institutions on society’s escalator, college is often an institution allowing the rich to get richer—but if you adjust for enough things, presumably this would include adjusting for whatever it is that gives you the opportunity and inclination to go to college.

Ummm . . . let me step back for a moment. What’s the causal effect of college attendance on future wealth? Of course we’d expect it to be positive. You learn stuff, you make connections, you get a credential, this turns into $. On the other hand, supply-and-demand and all that: if college is such a great deal, colleges will start to increase the price until on the margin its expected value in terms of future wealth will be zero. Or even negative, if college is actually fun and thus a “consumption good” like bullfight tickets etc.

We could also think about this from a causal-inference perspective. Let’s compare potential outcomes: future wealth of a person who is assigned to go to college, compared to future wealth of a person who is assigned not go to college. But then it depends on the person. The causal effect is hard for me to think about, even in the abstract, in that the effect of the treatment would depend on how it is applied.

48 thoughts on “Estimating the college wealth premium: Not so easy

  1. It also depends a lot on the era you look at. A college education in 1990 was probably a clear positive whereas a college education today if you’re talking about going deep into debt to the govt could easily be a negative. Also it depends on the major.

    One key factor we should distinguish is the benefit to society vs to the individual. People can be doing better more useful and more productive work so that the world is better off, but to them individually the student loan payments and increased competition for jobs can result in low take home income.

  2. “Cowen is conservative and education is a liberal institution, hence he’s supportive of studies that disparage education.”

    Cowen leans far closer to libertarian than conservative.

    Moreover, it does not follow that a hypothetical conservatives would be supportive of studies that disparage *education*, just because *universities* are a liberal institution.

    Plenty of conservatives and libertarians think that *education* is important. Conservative intellectuals’ complaint about universities seems to be more that liberals are in charge, instead of themselves. They aren’t opposed to education, in general. Libertarian intellectuals, by contrast, tend to have a more mixed opinion. Bryan Caplan has argued that much education spending is signalling without conferring value. Other libertarians are skeptical of government financing of education, but not opposed to education in general. Many libertarians also favor credentials that can certify someone knows a subject without necessarily attending a university (legally, it is difficult for US employers to screen job applicants through an IQ tests, colleges are largely a proxy for this).

    • Agreed. “Cowen is conservative and education is a liberal institution, hence he’s supportive of studies that disparage education” is both inaccurate and a cheap shot. It seems quite clear from Cowen’s writings and activities like making free, online economics courses that he’s very fond of education itself.

      More generally, there’s a built-in fuzziness in using the word “education” when what we really mean is “universities and similar institutions.” Many of us are very fond of education, but are increasingly uncomfortable about universities. My own, for example, just announced yet another tuition increase (+4.5% for next year) — factor that into the wealth premium calculation!

      • Raghu:

        1. I don’t think it’s a cheap shot at all. I agree that Cowen loves education—his blog is education, and he’s been doing it for a long time! He also has political views, and I respect that he cares about them. For related comments, see this post from a few years ago.

        2. I’m not just talking about universities. The same issues arise when considering K-12 education, for example in discussing research on the value of school spending.

        3. As noted in the above post, it would make a lot of sense to me for the college wealth premium to be negative. College is pleasant (or it’s often supposed to be), and if something is pleasant, it makes sense that its net cost would be positive, not negative. I do see the argument for college to be cheap or free, on the grounds that a well-educated populace is good for society more general, and this would be consistent with a positive college wealth premium. My point is that the net college wealth premium is a social choice; I’d expect economic forces to pull it below zero, but the social decision of subsidizing college could make it positive.

        • If the supply of education is a competitive market, then microeconomics suggests that the price should be the marginal cost. One way of reconciling this with the rise in college fees and with your comment on people spending money on pleasant things is that what is sold as college is no longer principally education, but “the college experience”. Perhaps it would be easier to work out how to provide an educated workforce without saddling them with ridiculous levels of debt if the proportion of money going to education was increased and the proportion of money going to “the college experience” was reduced.

        • Ag:

          I don’t understand your comment. What I’d say is that the primary function of college is the education but that people also want pleasant experiences. Also, education can itself be pleasant. I’m a soccer coach and we try to help the kids become better players—and they do improve!—but we also want it to be fun. Same with education. It won’t work if it’s too pleasant—learning requires intense effort and concentration–but if it’s all cold gruel, that won’t be so effective either. The rise in college fees is completely consistent with college being all about learning and with this learning being economically valuable. And if it’s fun, sure, people will pay more, just like people will pay more for a gym with nice showers and fluffy towels than with a gym where you’re uncomfortable all the time. But even at fancy gyms, people are paying primarily for the gym, not the towels.

        • Well, the supply of college education is clearly not a competitive market in your sense. Applicants clearly do not shop around for the cheapest place to have an education that they consider acceptable. In terms of the marginal cost, well the cost is essentially that of attracting the best staff. There is no reason to believe that such costs would diminish over time. Major universities do not make large profits to my knowledge.

        • I don’t know that for typical students college is pleasant. One of the big issues in talking about college or education by people who were successful in it or who attended “traditional” higher education is that that experience is a small portion of the total higher education experience. Most people go to community colleges. A lot of others go to commuter schools and are working full time while taking care of families. They are not lounging in the quad or bonding with roommates.

        • Elin:

          I actually had a bad time at college. I enjoyed high school a lot more. Most of the students at Columbia seem to be having a good time, but you have a good point that they have it a lot easier than the typical college student in this country.

      • The question is not whether conservatives tend to be anti-education. That’s like asking if conservatives tend to be anti-justice. Obviously, conservatives are supportive of a conservative notion of education, just as they are supportive of a conservative notion of justice. The question is, Are conservative notions of education consistent with the true purpose of education? Or, to be less subjective, what do conservatives believe is the true purpose of education?

        Having followed and participated in the long-running debate over the purpose of education, I find that conservatives tend to believe one or more of the following: education should be designed to prepare one for a career, not to foster secular enlightenment; to the extent that education goes beyond what is needed for career preparation, this is a private luxury, not a public good; and public investment in providing this kind of enrichment universally is misguided because not everyone is able or willing to benefit from it, either because they aren’t intellectually inclined, or because their circumstances (e.g., generational poverty) are burdensome and are better alleviated by career advancement than by intellectual stimulation.

        From what you say, I infer that your disenchantment with universities is a liberal concern–it’s too expensive and therefore not widely accessible without taking on massive debt. Liberal policies, like greater public investment in higher ed, combined with greater regulation of costs passed on to students (e.g., the textbook monopoly, the luxuries built into campuses, expensive sports programs), might resolve this problem.

        • Your point about conservative views of education reminds me that during the Free Speech Movement in Berkeley, one student reported that when she went home for the Christmas break, her father told her “I sent you to college to get an education, not a bunch of ideas.”

    • John:

      Thanks. To clarify: I’m not saying that Cowen’s views on the general subject of education are so strong that they would determine his position on this particular issue. And of course he teaches at a college himself, as do I. What I’m saying is that education in general (K-12 and beyond, not just universities) is a liberal institution (with some exceptions), at least in this country, and so I think that will make a conservative (or a conservative-leaning libertarian) sympathetic to studies that make existing U.S. liberal educational institutions look bad. If a study came out showing that private K-12 schools or religious colleges outperformed their public and secular counterparts, I’d expect a conservative-leaning libertarian would be sympathetic to those studies too, because private K-12 schools and religious colleges are more conservative than their public and secular counterparts. But, some private K-12 schools are liberal, and I’d expect a conservative-leaning libertarian would be happy to hear that liberal private K-12 schools perform worse than their public counterparts. Even though that last example is counter to libertarianism (it’s public outperforming private), I think it’s still the kind of story that fits the conservative-leaning libertarian worldview.

      Anyway, that’s just part of the story. Ultimately the research claim needs to be evaluated on its merits, not on its political implications. I have no problem with Cowen writing about a claim that aligns with his pre-existing views. We all do this sort of thing.

      • I appreciate the reply, and I agree that “ultimately the research claim needs to be evaluated on its merits, not on its political implications”.

        I like that you address that there are several dimensions at work here: liberal vs. conservative, private vs. public, secular vs. religious. Libertarians are generally more interested in the private vs. public issue. With respect to liberal vs. conservative, I would also distinguish this as liberal education philosophy vs. conservative education philosophy and liberal content vs. conservative content. I’m not really sure I have a good handle on the first one.

        • John –

          > . With respect to liberal vs. conservative, I would also distinguish this as liberal education philosophy vs. conservative education philosophy and liberal content vs. conservative content. I’m not really sure I have a good handle on the first one.

          I’m wondering if you could at least try to outline what you consider to be conservative education philosophy and conservative content?

      •  Cowen’s […] a conservative (or a conservative-leaning libertarian)

        I tend to think that neither of these descriptions is apt. Cowen simply takes too many positions that don’t fit any conventional contemporary understanding of “conservative.”

        When a man

        is an outspoken supporter of Black Lives Matter, and was outspoken years before the cause was popular;

        argues that America’s power is founded in “land theft and massacres”;

        doubts that genes are important determinants of life outcomes;

        is saddened by Trump’s election;

        is enthusiastic about contemporary art, including contemporary painting;

        supports sharp increases in immigration;

        repeatedly warns about the dangers of air pollution;

        sounded alarms about COVID a month before most others did, criticizes the libertarian response to COVID, and repeatedly argues that people should sharply curtail their social activity during the pandemic

        — when someone does all of these things, it doesn’t seem right to pigeonhole him as conservative or even as “conservative-leaning.”

  3. I haven’t looked at the study itself, although I agree that treating confidence in one’s financial literacy as exogenous is weird.

    The more general problem, and I’m speaking now as someone who hasn’t directly done Mincerian or similar education-wage-wealth studies but has used or critiqued them in related applied work, is that education is not a free-standing intervention that can be given a quantitative impact measure. In theory, what you ought to be doing is comparing the effect of attending an institution and getting a degree to not (or not fully) for the same individual. You could do this at a sample level if you were a totalitarian dictator, setting up treatment and control groups. Absent this we have observational studies.

    But education is tightly enmeshed within a host of other factors. Some of this is about access, as Andrew says. Another important set of influences operates on or through student “choice”. Some people get less out of classroom learning for a variety of reasons. Some people like or dislike them more for other reasons even though they might benefit about the same. And then there is the whole issue of opportunity costs, which are not only economic (and ought to be factored into access issues) but also psychological, familial, etc. I’ve had innumerable conversations with students informing me they were dropping out, and the list of reasons is very, very long.

    So the moral is: the role of education is simply not separable from a complex of other factors we would have to account for in observational studies. Instead of trying to estimate the bottom line effect of “education” on wealth or other outcomes, it is far more productive to study the effect of specific mechanisms like degree completion when it can be given a random or near-random interpretation and choice of major. Even these can be tricky, but they are more tractable.

    Politics: I suspect the motive is not so much pro-libertarian as anti-liberal (US liberalism, not classical). Liberals have been arguing for decades that the primary cause of increasing inequality is insufficient uptake of higher ed and that government programs to increase affordability, standardize “quality” and otherwise upgrade “human capital” are the solution. Cowen enjoys poking holes in liberal shibboleths, and this looks like such a beast. Historically, conservatives (here including libertarians) have argued that inequality reflects underlying differences in talent and character, so attacking the liberal belief in the education fix fits.

    For the record, I’m also skeptical of the liberal story but for other reasons.

    • There is something inherently incoherent about these education/wealth studies. I can understand education/income although I think all the points you make are valid critiques, but the situation is even worse when someone tries to study the effect of education on wealth. Education itself is a form of wealth. If you teach me piano, then what I have gotten from the piano lessons is knowing how to play piano. If you are going to “measure” the effects of those piano lessons on my wealth, but you aren’t going to assign a value to my knowledge of piano playing, then you have biased any measure against education. It is like measure the effects of home ownership on people’s wealth, but not counting the value of the homes.

    • Peter:

      I’m not sure how you can distinguish “conservative” from “anti-liberal” (in the U.S. sense), just as I think it would be hard to distinguish “liberal” from “anti-conservative.” I’m not saying that the political system is entirely unidimensional, but it’s zero-sum enough that I’ll associate liberalism with general support for liberal-leaning institutions such as education and liberal-voting regions such as the west coast and northeast, and conservatism with general support for conservative-leaning institutions such as religions and conservative-voting regions such as the south and rural areas. Not 100%—there are lots of educational institutions that conservatives like and lots of religious institutions that liberals like, etc.—but generally so.

      I agree with you that criticizing the educational wealth premium is a way of conservatives to criticize liberals on their own turf. A comparable salvo from the liberal side would be a studies that finds churchgoers to be more likely to engage in some immoral-seeming behavior.

    • There exist studies using observational, as opposed to experimental, data that try to tease apart the same-individual’s-counterfactual point in estimating the causal effects of education on earnings — the difficulty you point out in your second paragraph, using instrumental variables. Thus, the depending upon birth month, some students could legally drop out one grade earlier than other students. As to college, the Vietnam draft lottery was like random assignment of a damn good exogenous reason to go to college to some men, as opposed to others.
      Everything I have ever seen says that the future wealth gained from more schooling is, for almost everyone, greater than the future wealth gained from not going to school and working instead (what Mincer said the choice would hinge on). A big average treatment effect. Obviously, if you are Bill Gates and will be able to go get rich all the faster if you drop out of college, you will tend to do so. That such cases exist is unlikely to affect average treatment effects, I would think. (Although, I suppose, if estimation weighted Gates by his wealth, …. Krugman likes to point out that if Gates walks into a bar, the average person their becomes a billionaire.)

  4. Education was immensely worthwhile for me personally, and moved me from working class to the upper strata of the middle class. This was true in my era starting college in the mid 1960s and entering medical school in the later 1960s. Things have changed, and they will continue to change. A 17 year old junior in high school now has to consider what the economic climate will be from 2025 to 2055. What will be the good careers in that timeframe? What parts of the country will experience good growth, and what parts will lag? My decisions were easy, and the Vietnam War made it very easy to see that medical school was a good idea. There are reports that intergenerational upward mobility has been declining and that there is polarization of our society into a small number of extremely rich with a large number of the not so rich. In the the long ago past, the way to get wealthy was by birth; more recently education could propel some like me into a higher level. I understand the argument that the value of education has been largely neutralized. Probably, exceptionally talented young people will find a way up, but what is the path for someone within two standard deviations of the mean for talent during the next forty years?
    Professor Gelman, what has been the outcome for one of your students of average talent in the last ten years?

    • > and the Vietnam War made it very easy to see that medical school was a good idea

      This morning I got a military advertisement, something like “This student is smiling because they have no student debt. Sign up part time…”

      Anyway, so the education-military relationship is now join the military to go to college, which seems quite different.

  5. You’re conflating income with wealth. It’s an important distinction here because the result can be explained by simple tempo effects of wealth accumulation. The cross-sectional wealth premium is ambiguous even if college grads accumulate substantially more wealth in their lifetimes.

    This paper is garbage.

  6. A couple of flaws in your reasoning. First, non-profit universities aren’t in a true capitalist market: public ones are heavily subsidized by government (though less so than they once were) and many private ones have endowments, just to give two examples. Universities can also gain reputation, or alleviate public criticism, by lowering or eliminating costs for those who cannot afford to go and/or are in underrepresented groups.

    Second, for-profit universities can’t raise costs to meet the value of education and still compete with their subsidized competitors. But they have another option: they can lower the value of the education provided to meet the cost. Of course, only fools would knowingly pay full price for half the education they can get elsewhere, which is why the Obama admin tried to provide consumers with better information about return on investment by institution, and why DeVos tried to undo that.

    • IMO Andrew’s point about supply and demand is pretty much irrefutable. Students have obviously picked up on the fact that some fields are oversupplied (English literature) and others are undersupplied (math/science/coding etc) and are behaving accordingly.

      I’m not sure I agree with your point about private education. It may be that in practice for other reasons private schools like DeVos suck, but OTOH public universities are hardly known for their lean administration. America’s public universities haven’t been reaching new heights of greatness recently, so I wouldn’t think of them as an intimidating target for private schools with sound leadership.

    • Nothing is in a ‘true capitalist market’. From my vantage point- inside the Uni – the collapse of higher Ed has been mostly driven by the Uni system adopting the worst of American corporate managerial ideas, combined with the state-backed debt racket. Rent seeking behavior is nourished and cultivated, and the primary function of administration is to create more administrative overhead while competing on entertainment value to attract more students and glitzy ‘initiatives’. It pains me to have to be so cynical, but there it is!

  7. The financial acumen variable isn’t serving the purpose you seem to attribute to it. the authors’ point is that while income is higher for college graduates (though the premium is declining somewhat) that wealth premiums are declining much faster. Since income is the primary input into wealth, this seems puzzling unless a lot of people are pissing away their income. The authors suggest that (a) higher college costs may leave with them with less net income; and (b) financial liberalization may indeed cause them to take on debt which actually makes their wealth situation worse. But those with better financial acumen don’t have this effect (actually, it’s a little hard to see the net).

    One issue not discussed (in a cursory reading… maybe I missed it) is that the growth of college-bound cohorts ought to shrink the income premiums for two reasons: increased competition between college graduates and, more importantly, decreased advantages for those not ready to learn anything from college.

    The bigger issue here (to me) is the regression strategy: they note, quite rightly, that just lumping all racial groups together, and sensing that mere intercept differences won’t be enough, they estimate four separate regressions, even though the sample sizes for Hispanic and Black racial groups are really small. If only there were some sort of method to flexibly allow partial pooling of the factors; that might allow more robust conclusions, or at least conclusions that you could read sensibly in aggregate. But I really can’t imagine that in the hierarchy of methods there is one they could have used, but maybe I shouldn’t get all hyper about it.

    • Is income the biggest determinate of wealth? My parents paid for my college, my children also did not take on debt for college. So they start at a big advantage in terms of wealth by starting at 0, even if they hadn’t been able to save birthday gifts. Their parents and grandparents all own or owned homes and had fully funded pensions. I suspect that parental/grandparental family wealth is the biggest determinate of wealth.

  8. I just checked and thus far, the term, “supportive of” has appeared six times–soon to be seven if my comment gets through the filter. What decade did the vigorous verb,”support” get replaced by the wordy,”supportive of”? Who knows what abomination is on the horizon and who will be supportive of it?

  9. First of all, I think Dale Lehman’s question is perhaps misunderstanding the study. It’s not the *student’s* financial acumen that’s incorporated into the analysis, it’s the *parents’*. It is irrelevant, therefore, if going to college is making the students more financially savvy.

    Beyond this:

    “My point is that the net college wealth premium is a social choice; I’d expect economic forces to pull it below zero, but the social decision of subsidizing college could make it positive.”

    As always, there are excellent comments here. I think the post and some of the comments, however, muddy the implication of the study, and also this really interesting comment from Andrew.

    Regardless of what you and I think about the wonderful benefits of education in terms of personal growth and benefits to society, many students look at university-level education as a way to increase their future income and material well-being. This is especially true of students from poorer backgrounds. It is not an unreasonable view; it’s easier than ever to learn on one’s own without a university, and states’ diminishing funding for higher education shows how much we care about its social benefits. (My “public” university, the University of Oregon, gets $3k/student/year from the state.)

    Earlier today I was working on plans for a summer outreach activity I’ve been co-organizing for many years, that targets low-income high-school students. The question of what, ethically, one should tell these kids is a very real one. A numerical assessment of what the gains are of a university education is valuable. The messages from the university itself, or society more broadly, are not at all clear.

    • > A numerical assessment of what the gains are of a university education is valuable.

      Well, I don’t think that is being denied here.

      The difficulties Andrew points out in framing the causal question seem real. I like what Peter suggests here about trying to simplify the question (https://statmodeling.stat.columbia.edu/2021/03/29/estimating-the-college-wealth-premium-not-so-easy/#comment-1772905).

      But maybe a critique of this is that we already do try to simplify the question here with proxies and they end up getting gamed or being misleading. I think I’ve read stories about colleges gaming numbers like % of graduates who are employed after graduation, but maybe this is still easier than trying to sort out the casual question.

    • Raghu
      Thanks for the clarification. It is the parents’ financial acumen. But I still don’t see how that is exogenous. Isn’t it likely that the parents’ financial acumen is related to the child’s? If parents discuss financial planning, their children are likely to have spent years hearing such discussions, and developed such abilities themselves.

      • Yes, certainly; as the authors write: “Clearly, parents’ education and financial acumen were important variables…”. But isn’t that the point of separating this variable to infer the impact of the university education in itself? In other words, we want to know, for students whose background financial literacy is similar, how much of a financial gain going / not going to college will bring. Relatedly, one might want to know the impact of improving the students’ “financial acumen,” separate from higher education, but that’s another question!

    • “The question of what, ethically, one should tell these kids is a very real one. ”

      My suggestions:

      1) no degree is a ticket to anything
      2) the *relative value* of any degree depends on you
      3) regardless of what degree you get, who you know and what they think of you matters the most to your success
      4) despite the best efforts of everyone around you to make it otherwise, knowledge, intelligence and ability are still useful.
      5) there are more jobs in machine learning than in art history

      Oh, yeah, the stuff that has nothing to do with college:

      0) cash is king. Get it. Suffer to save it. Invest it. Make it grow.

        • FYI quiz link dead

          “no one is deciding between machine learning and art history…Univ. of Oregon awards about 10 degrees per year”

          Great news! Conceivably that’s an employable number of art history majors on a nationwide basis.

          For your low-income high-school students, it’s great that you’re looking for realistic assessments of the value of a degree.
          There are potential financial benefits – but those are degree- and person-dependent. There can be intangible benefits, but those are also degree- and person-dependent. A degree is an investment with risk of failure. The cost can outweigh the benefits if it’s not managed well.

        • Ha! Great quiz. I had absolutely no idea on any of them. I managed to get a few right. Well not true, I had medium conviction for the rising economics degrees.

      • People don’t get “a job in art history” they may get jobs in the culture industry, may go to law school or may just get regular jobs that don’t require a specific major. Very few under grads would get a job “in machine learning.” CS or software engineering grads would often start at entry level developer positions, but of course many work at small businesses.

  10. I’d think that, if you adjust for enough things, the college wealth premium should be negative.

    This goes for anything. Keep trying different adjustments, interactions, etc and you can choose between positive, negative, and negligible coefficients depending on your biases.

    The coefficient is an arbitrary number. Then people use that number to get an arbitrary p-value that they compare to another arbitrary number to see if it is “significant”. This is 99% of what gets called science nowadays.

    • Any college degree makes you eligible for tons of jobs that you won’t be eligible with a high school degree only. Any major where you do a lot of writing and analysis will serve you well. I also think you are probably mixing up wealth with income.

  11. I once had a dataset, I think this was in the late 1990s, that had annual earnings for a cross section of workers over a period of time -I can’t remember if the observations were one year apart or several years apart – when there was a noticeable jump in earnings inequality. Tabulated by education, the earnings rose much more for the college educated. But when we ranked the workers into percentiles, and looked at the effects in narrow percentiles, it was the lesser educated who tended to have the higher increase in earnings. There is a Simpson’s “paradox” or one of those paradoxes going on here: on average the college-educated appeared to benefit more from the rise in earnings inequality because they were more concentrated among higher earners. But conditional on relative earnings, the lesser-educated appeared to be receiving more of the gains. I’ve always wished I’d pinned down that relationship in more detail or been able to check it in other periods, but it was too tangential to what we were studying.

  12. I agree that the causal effect is hard to think about. Here are some reasons that it is complicated.

    If you buy a house with a mortgage, that decreases your net worth because you now have that huge amount of debt, and you don’t know the true value of the asset. So in wealth calculations it looks quite bad, yet if the housing market doesn’t collapse when you are trying to sell, a house can be a good investment.

    If you borrow for school you have debt, but many low income students do not borrow even though for some high salary majors (such as nursing) it may make sense to borrow compared to postponing until you have the money.
    At CUNY (second biggest system in the country, CSU being the biggest) 79% graduate debt-free. (https://rockinst.org/blog/a-deeper-look-at-student-loan-debt-in-new-york-state/)

    There is also a complicated gender issue. Of course, overall, there are a lot more women in college than men (this is a huge change from when I was growing up). Without going down a rabbit hole of why, let’s say it’s well established that women have, on average, lower earnings than men. The workforce is still pretty gender segregated, in particular I’m thinking about social work and education as two areas in which the students are overwhelmingly female and the salaries are low. (Again, let’s not get distracted about why.) Yes, it is hard for people in those roles to accumulate savings but they are also good, steady, white collar jobs. In some places they are unionized and have good health benefits and retirement programs. So some would say “that’s terrible, they should have learned machine learning” but lots of people think these are good jobs.

Leave a Reply to Elin Cancel reply

Your email address will not be published. Required fields are marked *