Matthew Yglesias trots out the old “college graduates make more money than non-graduates, and this is increasingly the case; therefore, we should send more students to college” argument. Not surprisingly, he fails to mention that even today, only 60 percent of college freshmen graduate within six years, and a quarter of employed college graduates in their 20s work in non-college jobs.

As I wrote regarding the college wage premium in general:

Those who graduate from college do so because they have higher cognitive abilities, more perseverance, and a greater ability to defer gratification than those who do not. These qualities would help them succeed regardless of whether they went to college. In other words, to some extent, their graduating from college is not the reason for their success, but a consequence of the reasons for their success.
The usefulness of the college wage premium is limited in other ways. It doesn’t include those who drop out. It doesn’t subtract the various investments (tuition, debt, taxpayer subsidies, time spent out of the work force) that go into four-year degrees. It makes no distinctions between different schools and majors. And it doesn’t factor in cost of living: One recent study found that college graduates tend to live in metropolitan areas where everything costs more — so much more that half the increase in the college wage premium between 1980 and 2000 disappeared when the author corrected for the difference.

Nor does the increase in the college wage premium mean that we should send more students to college:

The economic logic seems sound — if the price is going up and the supply is staying about the same, the demand is probably increasing. From this, it follows that if we can use public policy to increase the supply of college-educated workers, we should seriously consider doing so.
But if there’s such a high demand for college-educated workers, then why, even before the economy crashed, were 25 percent of college graduates in their 20s working at jobs that didn’t require degrees? (The proportion of graduates who utilize their degrees rises, by a few percentage points, until about age 32, but levels off thereafter.) As I pointed out in NR, people who graduate but don’t utilize their degrees get essentially no “college wage premium,” especially once you factor in the debt they’ve accrued and the years of work they missed while attending college.
A big part of the reason is that “college-educated workers” are not interchangeable. The college wage premium, and fluctuations therein, vary substantially by field of study. In other words, the economy doesn’t need more generic college graduates — and in fact refuses to hire many of them. Rather, it needs highly capable people in certain fields. It would probably be better to encourage students acquiring useless majors to switch to these lucrative fields than to send more kids to college across the board.
After all, when you send more kids to college, you’re scraping closer to the bottom of the college-eligibility barrel. The new kids will be less able and motivated, on average, than the ones who are already in college — and thus even more likely to drop out before finishing and to wind up in jobs that don’t utilize their degrees if they do finish.

A key thing to keep in mind with this supply-demand disconnect (Yglesias calls it evidence of a “substantial overall ineffienciency”) is that until we miraculously and dramatically improve the academic performance of high-school kids, we will have a limited supply of individuals capable of graduating college — that is, a cap on the supply. When supply is capped and demand is growing, the price goes up. We just have to deal with it.