Thinking About College

Data to Consider

As my children get older, I have recently had reason to thoroughly evaluate the landscape of options for post-high-school study. Most of my comments will be related to comparing destinations for the “traditional college experience” - that is, a four-year residential education. This, however, deserves a huge caveat.

College education was ridiculously expensive in the 90s when I was in college and is doubly so now as tuition continues to outpace inflation. The benefit of a college education, which was obvious for Baby Boomers when college degrees were both uncommon and cheap to attain, is more questionable today. In the 1960s, students could pay tuition with a part-time summer job, and corporations were eager to hire generalists with any type of degree. Today, computers have largely replaced middle managers. GE isn’t hiring history majors anymore, and the mass expansion of university education has made it less distinctive in the job market. So the value of a degree is much less while its costs have gone way up. Colleges and banks now capture, through tuition and fees funded by student debt, much of the net earnings surplus a college degree used to provide.

Post-Covid, I expect these trends to accelerate. The life insurance mortality data emerging now is pretty brutal: for disputed reasons, there was a massive uptick in deaths among the working-age population in 2021. Some claim this was due to vaccine mandates, but the data is less clear. Mortality accelerated mostly in the third quarter after the mass availability of vaccines, but around the time of many vaccine mandates after full FDA approval, but also around the same time as the emergence of the deadlier Delta variant. The effects are easier to discern than the cause.

These deaths undoubtedly contribute to the shortage of workers today. Economically, fewer workers mean less production and specifically less demand for specialized production. Until when and if demographics recover - and that seems unlikely for a long time given the collapse in fertility - we can expect the economy to regress away from specialization and towards basic production. We see this in the wage data. The workers best keeping up with inflation are those with blue collar jobs. A less specialized, poorer, depopulated economy needs more truck drivers and fewer human resource directors, more people doing necessary physical work, and fewer people doing digital paperwork. We are approaching a point where working in an Amazon warehouse may pay about as much as entry-level, non-technical jobs ostensibly requiring a college degree. Technical jobs that optimize the movement or production of basic goods (including worker productivity) should remain at about the same or increased demand.

Students today face a difficult choice. While there are some hopeful emerging trends of alternative routes to professional employment, those trends are far from robust at this moment. More than ever, there is a potential tradeoff between the costs and benefits of marginal college degrees. For families who cannot afford the luxury of the now-overpriced “traditional college experience,” the rational economic choices seem to be:

1. Go to the best school you can that will offer a scholarship minimizing costs relative to alternatives.

2. If one must pay retail, a degree from an elite private university (truly elite by SAT scores, like Stanford, CalTech, MIT, Duke, Rice, or Ivy League) or a 3.5+ GPA from a major state institution. These likely exceed their cost in increased earnings.

3. A near-4.0 GPA from a regional state university, commuting and testing out of as many classes as possible, or taking classes that transfer at a community college, to save money.

4. Don’t go to college and develop a trade or skill that will be more valued than a mediocre generalist degree.

Marc Andreessen believes that the college system is taking longer than some expected to replace despite its increasing economic costs because it offers a hard-to-replicate “bundle” of services: some skills acquisition, a “holding tank” for 18-22-year-olds whose prefrontal cortexes are not fully developed, a social/dating scene, and networking among future elites at a time when brains are primed to form lifelong bonds of friendship. More cynically, he believes the college system could be an unbelievably complex and expensive workaround to direct IQ testing of job applicants that might run afoul of disparate impact legal rulings. Nevertheless, since we live in the world of what is instead of what ought to be, he comes to the same reluctant conclusion I do: if you can go to college, you probably should, even though it’s inefficient and wasteful now more than ever.

So if college is still the right choice for many, how does one choose among them? Here are things I considered in making these choices for my own family.

SAT Scores

SAT scores represent the relative ability and achievement of the student body. Employers, however, seem to irrationally weigh college GPA more heavily in hiring than considering the “competition” a student faced for top grades at a given institution. At the same time, friendships are more easily formed among students with similar academic ambitions, and networking with other high achievers is helpful for future opportunities, so the solution isn’t to attend a non-challenging school where a 4.0 can be achieved easily. Balancing these two factors, I would tend to look for the best school where a student is at least at the 75th percentile of the SAT range for the student’s chosen major. Any lower and the difficulty of maintaining a 3.5 becomes more difficult.

The exception here are wealthy families who can afford private schools. One way of looking at private universities is as a form of social insurance for the well-off. Many offer generous scholarships and financial aid for academically elite students, whose scores, grades, and accomplishments help establish the institution’s reputation. On the other side of the admissions office, they will lower their standards for “legacy” admits, the children of alumni, especially generous alumni. The “Harvard number” refers the amount of donated money it takes for one’s academically mediocre child (relative to peers) to be admitted to the university. Since private colleges also have the most extreme grade inflation, especially compared to flagship state universities, wealthy families who grease the skids can be assured of a decent GPA upon graduation even for legacy or donation-enabled admits. This is doubly true for second and third tier private universities, and this system enables students who would never be admitted to many state universities, or who would struggle for grades if admitted, to avoid the class stigma of attending a regional public university.

Sex Ratios

For many reasons, young men are trending towards only 40% of the undergraduate population nationwide. This study is suggestive enough that if I had sons, I don’t think I would allow video games in the house. A full discussion of this is outside the scope of this post, but suffice to say that most colleges today necessarily have unbalanced sex ratios. However, if humans are designed for social groups where the sexes are balanced, it follows that the more unbalanced the ratio, the more pathological the social dynamics become. This isn’t good for men or women.

Unfortunately, there are only a few schools that offer a balanced social environment for the sexes. In Texas, among major universities, there’s Texas A&M (50% male), Texas Tech (52% male), or some more selective private universities like Rice (52% male) and Southern Methodist University (51% male). By contrast, the University of Texas freshman class is 41% male, Baylor is 40%, and East Texas Baptist is 45%. For those seeking a Christian environment, to my knowledge there are no Christian schools in the state, outside of seminaries and the engineering-focused LeTourneau University (52%), that offer anything close to a balance.

Ideological Diversity

College campuses are increasingly intolerant of conservative viewpoints. For those with such convictions, or liberal folks who enjoy open debate, it is worth considering schools where the student body is at least not overwhelmingly polarized. The New York Times compiled precinct-level voting maps for the 2020 election. One way to assess ideological diversity is by zooming in on the campus of a given university and seeing how the on-campus precincts voted in the election. While I try to avoid stereotyped thinking, and hoped seeking data would refute them, I was genuinely surprised by the extremity of some of the voting results. Rice University voted 8% Republican, and the University of Texas voted only 14% Republican. The only schools I researched that came close to parity were Texas A&M (40% Republican) and Southern Methodist University (55% Republican). This also demonstrates the biased perception of what diversity actually means in academia, as it’s certainly not diversity of thought or ideas. A&M and SMU are considered conservative schools, yet both are closer to balance in political affiliations than either Rice or UT.

Student Outcomes

Most people don’t realize that college rankings are highly manipulated. What we need is objective, population-level data to show which colleges are producing good outcomes for their students. Thankfully, that’s now available, thanks to the researcher Raj Chetty, who convinced the IRS to share complete collections of anonymized tax data with him, which he then matched with college attendance records. The New York Times put together a great visual tool showing the results of the data (this links to the data for Rice University, but you can click around to see other schools). These numbers are complete and comprehensive data, not a sample or survey, and is the most robust statistical comparison between different universities ever conducted, since only his team was given direct access to complete sets of IRS and college enrollment data. Below are some of the data points I thought were interesting.

Earnings

Expensive private schools often tout the value of their “network” of presumably prosperous alumni as justification for outrageous tuition. In Texas, outside of Rice University, Chetty’s IRS data would suggest otherwise.

Texas A&M University (tuition: $12,000) comes in 2nd in alumni earnings, at $59,400 per year, ahead of all other private schools except Rice in the state charging four to five times more. This is especially impressive in terms of “value added” because A&M’s SAT scores, an indication of their incoming students’ raw talents, are somewhat lower than the lower-earning graduates of the University of Texas, Trinity University, and Southern Methodist.

Marriage

The peer groups formed in college influence values throughout life. Perhaps the strongest evidence of this is Chetty’s findings on the percentage of graduates who are married at age 34. Of course, not everyone will marry for all kinds of reasons, but when applied to large numbers of people it is a useful key indicator of overall life outcomes. I was frankly surprised by how much the marriage rate varied among various schools. For example, only 60% of Rice graduates were married by age 34:

And here are the top schools in the state for marriages:

Again, we see Texas A&M as an outlier among large universities in the state, and also nationwide:

In fact, A&M seems to have a higher marriage rate than almost any schools outside of seminaries, small religious colleges, and those serving either the Mormon or Orthodox Jewish communities. I think the Mormon comparison is apt; the essence of the Aggie is earnestness, enthusiasm, and a lack of cynicism that outsiders can find a bit unnerving. That such a culture, like Mormons, produces marriages at a higher rate is not a surprise. And remember this is not tracking marriages right out of college, but rather the percentage of alumni married by age 34, which is probably fairly representative of lifetime marriage rates. I don’t think there’s necessarily a causal relationship here, but rather the statistics provide an objective window into the student culture at various schools. Looking at these types of numbers can help determine if the student body is a good fit for one’s own values.

A Failed Prediction

In looking at the marriage data, I developed an idea that perhaps one of the drivers of low marriage rates was downward social mobility at private schools. Necessarily, the parents of private school students have higher incomes, while the graduates of most private schools end up having substantially less income than their parents. This isn’t a huge surprise necessarily indicating a problem with those schools (other than the value proposition of their price tag versus alternatives). Families need an extremely high income to afford private college, and when parents have an extremely high income, it is more likely than not that their children will have lower incomes, due to regression to the mean. This happens across statistics when dealing with extremes. Extremely tall parents, for example, will likely have taller children than the average, but those children will probably be shorter than them.

Nevertheless, I had the idea that people perceiving a downward pressure in their prosperity across generations would cause them to delay marriage, because they would feel poorer than the same student who was more prosperous than his or her parents. I was curious enough to download the raw data and performed my own analysis. I was wrong on this. The opposite was actually true, because it turns out that parental income is itself an extremely strong predictor of children’s marriage rates. See the table below showing the correlation between marriage rates at a given school (four year colleges) and median income:

This makes sense when we think of income as a proxy for social class, and it’s consistent with Charles Murray’s arguments in his book Coming Apart. Whatever effect downward social mobility might have, it is overwhelmed by the class-based differences in social outcomes that have broadened significantly over the last few decades.

Coda

I’ll admit the Tom File has been sporadic this year, but more commentaries are in various stages of completion. Thank you for reading and I hope you find my perspectives on things useful.