Expensive colleges deliver zero value for computer science majors?

“Do Elite Colleges Lead to Higher Salaries? Only for Some Professions” is a WSJ article by Eric Eide and Michael Hilmer, academic authors who studied the link between prestige of undergrad institution and mid-career earnings.

CS grads from four wildly disparate schools in a chart ended up earning roughly the same by mid-career. Here’s what the authors had to say about science/tech fields in general:

What we found startled us. For STEM-related majors, average earnings don’t vary much among the college categories. For example, we find no statistically significant differences in average earnings for science majors between selective schools and either midtier or less-selective schools. Likewise, there’s no significant earnings difference between engineering graduates from selective and less-selective colleges, and only a marginally significant difference between selective and midtier colleges.

Our findings are crucial for families to understand because chasing a prestigious STEM degree can leave students burdened with huge amounts of unnecessary debt. Financial aid can certainly help, but for many families, the cost of education can still differ dramatically across schools. For example, if an engineering student chose to attend the University of Pennsylvania instead of Texas A&M, the average starting salary would differ by less than $1,000, but the tuition difference would be over $167,000. At that slightly higher salary, you’d have to work for more than 150 years before you make up for that vast tuition difference.

The authors didn’t look at programmers without degrees. I wonder what they would have found if they’d studied people who started coding at age 18 rather than spend four years in college. Perhaps they would have significantly out-earned the college grads due to always having four additional years of work experience.

[Separately, the researchers may have left out an important factor: the wealth level of potential mates at different universities. From the introduction to Real World Divorce:

“When young people ask me about the law as a career,” said one litigator, “I tell them that in this country whom they choose to have sex with and where they have sex will have a bigger effect on their income than whether they attend college and what they choose as a career.”

From the history of divorce chapter:

What Weitzman did not predict is the tremendous variation in child support profits from state to state, despite the fact that federal law requires each state to follow a similar procedure in developing child support guidelines. How profitable can it be? A professor of economics in Massachusetts, a typical “winner take all” state, said “The best career advice that I could give to a female freshman would be to drop out and stop paying tuition. Get pregnant with a medical doctor this year. Get pregnant with a business executive two years from now. Get pregnant with a law firm partner two years after that. She’ll have three healthy kids and a much higher after-tax income than nearly all of our graduates in economics.”

Choosing a college where fellow undergraduates disproportionately come from rich families and where the campus is located in a state featuring the highest profits from short-term marriages and/or out-of-wedlock pregnancies could result in a high return on the tuition investment. It is easier to meet and have sex with a rich college kid if one is a student at the same college. (See also “Assembling a longer-term annuity from child support via frozen embryos”) Of course, majoring in CS is probably not the best choice if one is going to pursue wealth through sexual encounters…]



  1. Anon

    February 1, 2016 @ 5:53 pm


    Dr. Greenspun, thank you for sharing this article. I wonder what these finding suggest about the nature of these professions. Could one argue that CS and engineering jobs are more merit oriented since it’s easier to quantify contributions from an individual than someone who is a management consultant or a researcher at a social science think tank (Rand, etc)?

  2. alanc

    February 1, 2016 @ 8:42 pm


    I would have to go with essentially dishonest for them to compare costs between a private Ivy league (Penn) and a mostly free resident state school (TexasA&M). At a minimum they should compare the non-resident cost of Texas A&M.

  3. G C

    February 1, 2016 @ 9:50 pm


    Great article, thanks for sharing. Overspending on college is very easy to do, IMHO.

  4. philg

    February 1, 2016 @ 10:10 pm


    alanc: It is dishonest to include a low-cost state school because non-Texans would have to pay more? Isn’t there a low-cost state school option available to most Americans? E.g., if they live in Kansas they could attend University of Kansas ($11,000/year in tuition) or Kansas State at less than $8,000 per year (see https://www.k-state.edu/admissions/finaid/ where they also disclose a “privilege fee” of $832/year; this raises the question of whether there is a “white privilege” fee that is different…)

  5. Jackie

    February 2, 2016 @ 5:06 am


    It looks to me like they purposely chose strange pairings to rig their results. Notice that they compare Penn (an expensive Ivy League university but one with a relatively weak engineering school) with a strong state school. Would the results have differed if they chose MIT for engineering and Penn for business?

  6. alanc

    February 2, 2016 @ 10:06 am


    philg – they could easily have said “for a texan Texas A&M versus Southern Methodist” (SMU tuition higher than Penn) or “for a pennsylvanian Penn State versus U Penn” – but much like your favorite subject, there is a great varience in in-state costs – many Californians go to Idaho for college where the out of state tuition is less than the California in state tuition.

Log in