As higher education costs escalate faster than inflation, while the global job market remains largely depressed, increasing numbers of commentators are questioning the value of higher education as a whole. The question becomes more urgent, as state budgets become increasingly burdened by higher education’s further expansion. However the answer to the question of education’s value seems pretty clear to me: it depends who is getting educated in what.

The recession has exacerbated a trend that was already in place, reversing a development of the 1970s through the 1990s. In 1973, college was cheap compared to today, but the differential in average earnings between college graduates and high school graduates was also smaller than today. The Census Bureau’s figures, comparing incomes of those with a full high school diploma with those with a bachelor’s degree and no more (and including the unemployed, an important factor in 2009) showed that men with high school degrees had average incomes of $46,741 in 2009 dollars, while those with college degrees had average incomes of $60,148, a differential of $13,407. For women, the figures were $17,131 and $26,814 respectively, a differential of $9,683 — pay levels for women at work were not quite as grim as this indicates; in those days (and to a lesser extent even today) many women were full-time housewives. This skews the figures, but in this analysis it’s impossible not to skew them somehow. In any case, the conclusion that a woman who graduates and immediately gives up work for motherhood earns nothing from her college degree is logically correct, however misleading we may think it in reality.

With the average cost including room and board of a four-year college being $9,295 in 1975 in today’s dollars (and 1973 won’t have been much different – the Great Inflation of college costs had not really begun) the pre-tax payback period from earnings for a four-year college was 2.77 years for men or 3.84 years for women. That’s rather too favorable, because college graduates had to pay more tax on their increased earnings. Assuming a 30% average rate of tax on earnings the payback period became 3.96 years for men and 5.48 years for women. For the state, subsidizing the college costs of poor students or providing state colleges with subsidized fees, and receiving tax payments after graduation at 30% on additional earnings, the payback period can be calculated as 9.24 years for men and 12.8 years for women.

Looked at in a different way, if we assume the average career in full-time employment lasts for 30 years, so that the differential between college and high school earnings becomes a 30-year annuity (with both college costs and earnings being at 1973 levels, corrected for inflation), then 1973’s pretax return on investment for a man from college education becomes 36%, for a woman 26%. Post-tax the returns become 25% and 18%. For the state, the returns on subsidization are 10% for men and 7% for women. All healthy figures, confirming the belief that in 1973, even with the earnings differential between high school and college less than it later became, a college education was on average a pretty good deal. Even allowing for the fact that the average includes many students whose benefits from college education were lower than average, college must have been an adequate or better deal for pretty well all those undertaking it.

Move forward to 1991. By this time, the annual cost of college in today’s dollars had increased by 29% to $12,017. However the earnings premium from a college degree had also increased, by 66% for men and 61% for women. Thus even though the financial burden of college had increased, the benefit from going there had increased even more, raising the return on college education even higher and making it easy for colleges to increase their tuition costs, professorial pay and numbers and administrative burden.

By 2001, the premium for the college educated over high school graduates had increased by another 18% for men and 19% for women, making college even more attractive. However the cost of college had continued to escalate, by a further 30%. Returns on that college education investment were no longer so high as in 1991, although still considerably higher than in 1973.

Now look at 2009. The cost of college in real terms has escalated by a further 29%. However with the lackluster economy of the 2000s followed by a further grinding recession, the earnings premium for college-educated workers has declined, by 9% for men and 5% for women. It still remains around 80% higher than the 1973 earnings premium, for both sexes, but the effect of continued ratcheting up of college costs and the modest decline in the college earnings premium has made college dramatically less attractive on a return-on-investment basis. Instead of the healthy 25% and 18% post-tax returns that had been achieved in 1973, and the higher returns in the intervening years, the post-tax return from college investment has fallen to 21% for men and 18% for women. Because of the huge increase in costs, the college investment is also much riskier in relation to people’s earnings and assets than that of 1973. As for the government, the tax return from providing full scholarships has declined into single digits, 8% for men and 5% for women.

On average, college remains financially attractive. However the average conceals large differences, and two countervailing factors. As to the differences, college remains highly financially attractive to those enrolled in the best schools. If you can get into Princeton, you should go there; it costs little more than other private schools and your potential earnings uplift is stunning. Furthermore, your return is better from a scientific or technical degree than from a liberal arts degree. That should not be taken too far; petroleum engineers currently enjoy the best average education premium, but that reflects current buoyant hiring trends in the oil business, which are unlikely to last over an entire career – as a petroleum engineer in the oil price downtrend of 1986-2002 your earnings stagnated and you may well have suffered periods of unemployment. Nevertheless, a degree in biotech today is unquestionably worth more than one in women’s studies.

As to the countervailing factors, the positive one is that college is more fun for almost everybody than the first four years in the workforce. For some few aesthetes (though I suspect only a small proportion of the ex-student population) the intellectual or aesthetic benefits of a really good college education also provide additional pleasure throughout their life, beyond their increased income.

The negative countervailing factor, and it’s a big one, is that the income benefits outlined above may not in fact be caused by college degrees but simply coincidental with them, because those who go to college are on average smarter than those who don’t. After all, men with doctorates earn 66% more and women 82% more than those with bachelor’s degrees, but that doesn’t mean we should educate everybody to PhD standards, it simply means those capable of mastering PhDs are a generally pretty smart lot, at the top end of the curve. (However intelligence isn’t everything; those with professional degrees (lawyers, doctors and MBAs) — presumably not quite as smart as PhDs — earn more still. If intelligence is good, a ticket of entry to an artificially restricted profession is even better!) So at least some of the earnings benefit from a college degree is in reality an earnings benefit from being smart.

For society as a whole therefore, the benefits of college education may be diminishing, though still considerable. Pushing to educate students ever further down the ability spectrum seems likely to have costs exceeding its benefits, unless the students are very highly motivated to undergo the college experience. Rather than expanding the U.S. college system it would make more sense to streamline it, closing marginal colleges with a population of indifferent students and cutting back subsidies to programs of low economic value. Likewise, subsidizing low-ability students in for-profit colleges, generally without campuses, generally adds little value.

Conversely, college is still relatively underused for mid-career training or re-education, especially in areas directly related to burgeoning sectors of the economy. Here, students with technical training that has become outdated (that computer science BA completed in 1988!) can re-equip themselves for the second half of their working life. Many technological subjects can become hopelessly outdated within a decade, and the instability of working life is such that it is unrealistic to expect employers to pick up all the slack through in-house training. Mid-career education, while it needs to be of the maximum possible quality, is however different in nature to initial college education; it needs to be closely oriented towards practical skills and completed as quickly as possible. If the state subsidizes education, these courses and these people should receive the majority of its subsidies.

For individual students, college should only be undertaken for those who expect to gain especial benefit from the college experience, either academically or through its other offerings. The initial four year college should thus be the province of no more than 15-20% of students. Most of the next 60% of the ability group should get higher education in smaller doses, a year or two initially, the rest spread throughout their careers. With this approach, there will be far fewer students wasting four years studying political correctness and binge drinking – but, far more important, many fewer steel workers and auto workers finding themselves made redundant at 50, and unable to transition to a new life phase of renewed productivity.

Assuming the U.S. economy of 2011-2015 is more like that of 2009 than that of 1999, the market is telling us something about the changing economics of college. We should heed its message.

-0-

(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)