once again: college effects do not matter, college major effects are huge

I just discovered an Economist article from last year showing that, once again, which college you go to is a lot less important than what you do at college. Using NCES data, PayScale estimated return on investment for students from selective and non-selective colleges. Then, they separated STEM majors from arts/humanities. Each dot represents a college major and its estimated rate of return:


Some obvious points:

  • In the world of STEM, it really doesn’t matter where you go to school.
  • High prestige arts majors do worse, probably because they go into low paying careers, like being a professional painter (e.g., a Yale drama grad will actually try Broadway, while others may not get that far). [Ed. Fabio read the graph backwards when he wrote this.]
  • A fair number of arts/humanities majors have *negative* rates of return.
  • None of the STEM majors have a negative rate of return.
  • The big message – college matters less than major.

There is also a big message for people who care about schools and inequality. If you want minorities and women to have equal pay, one of the very first things to do is to get more into STEM fields. All other policies are small in comparison.

50+ chapters of grad skool advice goodness: Grad Skool Rulz ($2!!!!)/From Black Power/Party in the Street

Written by fabiorojas

May 11, 2016 at 12:01 am

8 Responses

Subscribe to comments with RSS.

  1. Fabio, your intervention here is a bit sloppy. What college you go to matters a lot because it matters a lot whether you graduate or not.

    This study says there is not a big difference for bay between types of institutions for DEGREE EARNERS in a similar field. Students that go to a selective institution and a wealthier institution are more likely to graduate than equivalent students at less selective institutions.

    Also, the sample for this is just 240 institutions. And there are few institutions in the scatter with 100% admissions, meaning that for-profit institutions are probably way undersampled. Studying business or computer science at a for-profit college is a one-way ticket to dropping out and defaulting on student loans. See Gelbgiser, Dafna. “College for All, Degrees for Few: For­ Profit Colleges and Socioeconomic Differences in Degree Attainment.” In American Sociological Association 2015 Annual Meeting, 2015.

    Liked by 1 person

    Charlie Eaton

    May 11, 2016 at 12:17 am

  2. Charlie: You raise an excellent point. There is a huge difference between non-profits and for-profits. But still, conditional on your academic skills, I would wager these results hold. These are not results to be ignored.



    May 11, 2016 at 1:43 am

  3. It is important to highlight that this is exclusively about pay. It does not involve things like satisfaction or occupational prestige. I see this in the alumni surveys of both my private, elite alma mater and my public, virtually open access current institution. The difference in terms of pay for the two groups is surprisingly small. But that is more because the former group is more likely to be working high level but low paying jobs in prestigious non-profits while the latter will be middle management. College may not matter much in their salaries, but it certainly matters in determining who is going to be executive director at the red cross and who is going to be store manager at Men’s Wearhouse.

    Liked by 1 person


    May 11, 2016 at 1:52 am

  4. Dlp: True, but as college gets more expensive, the value added becomes more salient. If it were cheap, these question would have little policy relevance.

    Liked by 1 person


    May 11, 2016 at 2:32 am

  5. “High prestige arts majors do worse, probably because they go into low paying careers, like being a professional painter (e.g., a Yale drama grad will actually try Broadway, while others may not get that far).”

    Isn’t this backwards? Doesn’t a low admissions rate mean the institution is more prestigious? In other words, institutions become less, not more prestigious, as they move right along the X-axis.



    May 11, 2016 at 3:27 am

  6. My bad. I was reading 1-X in my head. You are correct.



    May 11, 2016 at 3:28 am

  7. Fabio,

    Is ROI the best way to measure this? The total four year cost of these schools, which appear to be what the ROI is based on, range from $50k to $250k. A 10% ROI should vary sizably depending on school cost. So shouldn’t your first bulleted piece of advice be: in the world of STEM, go to the most expensive school! Just to beat a dead horse, take as an example Southern Polytechnic State University and Harvard University, and bracket for a moment the college major. Based on the data grabbed by the Economist (, both schools have an overall annual ROI of 14.3%. But because Harvard’s more expensive, the 20 year net ROIs differ by about $275k. Seems like a smaller ROI from a more expensive school might end up being more lucrative than a larger ROI from a less expensive school, and so the hypothetical high school student interested in maximizing their lifetime earnings should indeed consider school choice along with major choice.


    Tom VH

    May 11, 2016 at 8:14 pm

  8. “Tom VH:” You raise a good point. The ROI is the standard measure of investment because it does not depend on the amount given or on the opportunity costs. It is standardized and agnostic toward alternatives. It simply says, “if I give you $100, what will you give me back a few years from now?” So to answer the question, per dollar, does it matter which college I go to if I care about my pay increase down the line? The answer seems to be no.

    But your question is different. You ask, “should I go to a more expensive school because my profit (not ROI) is higher?” To answer, you must ask about the opportunity cost.

    Let’s take your example: Harvard vs. Southern Polytechnic for a STEM major. Option 1: a 14.3% return on $250k at Harvard. Option 2: 14.3% on $50 at Southern and $200k on something else.

    Clearly, there is no single answer and it depends a lot on context. For example, if you had to take a high interest loan to make the Harvard tuition, then it may make sense to go for Southern. Or, if you have to take a second job to make the Harvard tuition, then Southern may make sense. If you already had the money on hand, you could spend $200k getting a public school medical degree on top of the BA and be debt free, or buy a home, or simply drop it in the stock market and use it for retirement.

    I would also add that if you needed to take a loan to cover $200k, you probably will have to pay that over about 20 years and you will need a stable job immediately to make the monthly payments. That is a serious cost.

    The bottom line is that,what you should do with an extra $100 depends a lot on what the opportunity costs are and it is not clear that just giving money for more expensive schools makes a whole lot of sense.

    PS. Bonus round – if you buy this study, it shows that some majors are horrid investments even at fancy and highly selective schools. A lot of major/school combinations have low returns. If the stock market gives you somewhere between 5% to 10% over 20 years, then you should simply be a cheap as possible (e.g. if you really want a drama degree, go cheap because you’ll lose money in almost all cases).

    PSS. In my own case, I got a STEM degree from a cheap, but fancy, school (public flagship before the tuition spike of the 1990s) and I never regretted it. I turned down fancy private schools because I suspected that it would be insane to pay off $100k in 1989 dollars for a degree that could be had for a third of the price. I’ve now seen my peers spends decades of their lives paying off student loans. I doubt the mental boost from a private school degree would outweigh the empty bank account in my mind.



    May 11, 2016 at 8:59 pm

Comments are closed.

%d bloggers like this: