Americans: Over- or Under-Educated?

I have to hand it to Anthony Carnevale. He is persistent. He and his Georgetown colleagues have issued a series of reports with one overriding message: There is a big, and probably growing, payoff to earning a college degree.

The last of a series of reports does offer some interesting data and observations. It is true, for example, that the choice of college major matters in terms of a personal return on an investment in college, and that there are important differences based on other demographics, such as gender and ethnicity.

But the overall message is that typically there is a big payoff to going to college. The Lumina Foundation is using the Georgetown findings as vindication of the urgency of their call (and that of others, notably President Obama) to sharply increase the proportion of Americans with college degrees.

A quick look at the latest study shows at least six shortcomings. First, and probably the most important, is that Carnevale and his colleagues are comparing apples with oranges. Those who go to college are, on average, brighter, more knowledgeable, more disciplined, and more conscientious than those who go to work after higher school. For decades, economists have written about the screening dimension of higher education. Even if those in the college going population had NOT gone to college, they would have earned more than those others who did not go on—simply because they are better workers.

Second, the Carnevale report admits, not everyone is average, as Neal McCluskey pointed out the other day. For example, it acknowledges the 28 percent of those with just an associate degree earn more than the median of those with bachelor’s degrees. Social-worker majors, for example, earn typically dramatically less than the median or average of the entire bachelor-degree population.

Third, remember that 40 or 50 percent of those going on to college do not graduate—even within six years. There is a substantial risk element to making a college investment, particularly those with attributes (low high-school grades, poor test scores, etc.) that suggest the probability of dropping out is high. Adjusting for this risk factor lowers the expected income gains from college dramatically. Carnevale looks just at college success stories (graduates), ignoring the important part of the college-going population who are far less successful (those not graduating).

Fourth, a small technical point. The Carnevale report estimate that the present value of the lifetime earnings advantage associated with a four-year degree is the better part of a million dollars is based on an assumed 40-year work life—working full-time, year round. In reality, many Americans do NOT do that—they take breaks (in some cases, involuntarily) from employment. In particular, women often take several years off to have and raise children.

Fifth, and very important, is that what is important in life usually happens at the margin, not in the middle (the average). What is important to the future is what job prospects will be for the increased number of graduates coming into the labor market. The past does not always predict the future. Increasing the proportion of college graduates from 10 to 20 percent of the population, or even 20 to 30 percent, probably has a different economic impact than increasing it from 30 to 40 percent (the future). Carnevale acknowledges that possibility, but opines that it does not lower the expected earnings differential enough to negate the positive net payoff to higher education. But who really knows? I think the diminishing returns factor to the higher-education investment is pretty high based on my reading of some labor-market data.

Related to that point, a huge percent of college graduates are taking jobs that are traditionally taken by those with high-school diplomas. Carnevale partially comes back with a counterargument: even for those persons, college grads earn more than those with just a high-school diploma. That is often true, but the differential associated with the college degree is almost certainly dramatically reduced. Do the 80,000 college grads who are bartenders gain the same $900,000 plus advantage from a diploma that others supposedly do? I doubt it. And remember the costs of college are rising, and likely will continue to do so in the future. Those costs, by the way, appropriately include the opportunity cost of income foregone during college years from lost labor income.

Lastly, not all colleges are created equal. If there is a $900,000 advantage associated with a college degree, there is plausibly three times that advantage associated with a degree from Harvard College. But what about the graduates of the University of the District Columbia, Youngstown State University, Chicago State University, or the University of Texas at El Paso? First, most of those enrolled at those schools don’t graduate (at least as far as the U.S. Department of Education can tell us). Second, I doubt the earnings differentials associated with a degree from these schools are equal to the average of all college graduates. That is why we need the IRS to provide us average earnings data by college to help evaluate the differential rate of return on investments in the various colleges in the U.S.

The Georgetown studies are opening up a much needed dialogue on the return to a college education. Several colleagues and I will be joining this debate more substantively in the weeks and months ahead. Indeed, I am asking Prof. Carnevale to join me, and no doubt others, in having a frank dialogue on this issue sometime in the near future.

Return to Top