There may be no more highly-placed booster for the importance of college education than editor/journalist David Leonhardt of the NYT (disclosure: an editor of my occasional pieces at Upshot and someone I’ve known and respected for many years). Every few months, he writes something on this topic and he’s right to do so, both from a macro perspective–we want the best human capital inputs we can have in our national production function–or a micro one–we want people to achieve their full intellectual potential. (And FTR, I write pieces extolling full employment numerous times per week, including this one…read on.)
In today’s piece, he’s riffing off of a new study that uses an econometric technique that wasn’t around–or at least wasn’t in my tool box–back when I was learning the trade: “discontinuity regression.” It creates a quasi-experimental design to test the impact of an intervention by comparing outcomes of people on each side of the borderline to entry to the intervention. In this case, David’s reporting on kids who just made it into college with kids who didn’t quite clear the bar. Presumably, both groups have similar cognitive skills (or at least similar SAT scores), so there’s a kind of natural experimental and control group.
…the data show that the students just above the admissions cutoff earned substantially more by their late 20s than students just below it — 22 percent more on average, according to the Florida study, which was done by Seth D. Zimmerman, a Princeton economist who will soon move to the University of Chicago. “If you give these students a shot, they’re ready to succeed,” said Mr. Zimmerman, adding that he was surprised by the strength of the findings.
Well, maybe the strength/magnitude, but the existence of an historically high college-wage premium (the wage advantage of college over non-college educated workers) is one of the best known facts in labor economics and it’s one of the reasons why almost every politician/policy maker touts the importance of higher education.
So why push this idea so hard when it already has so many converts?
As David notes, he’s pushing back on a pushback. That is, after decades of “go to college!” as the central argument against wage stagnation and inequality, two other arguments have evolved. One, not everyone will benefit from college, and two, more people with college degrees won’t fix all that ails us in terms of wage trends.
The former is covered in the piece, as various authors and economists have stressed two-year degrees, associate’s degrees, or apprenticeships as alternatives to four-year degrees. Both sides have a point (as David acknowledges) but I must say, the people I hear making the (correct) argument that not everyone benefits from college all went to college. Most have PhD’s.
The danger here is that the emphasis on four-year college discredits the rationale for seeking a lower attainment level. Research by labor economist David Autor initially suggested that there was a hollowing out in the middle of the job market, with lots of low-skill (non-college) and high-skill (college) jobs. But as Harry Holzer has pointed out–and Autor’s more recent work agrees (I think)–there is a critically important labor market space for certain middle-skilled jobs that don’t require four-year degrees (and, yes, here’s another Ph.D. making that case…actually two).
The way I think about it is a high-school educated home-health aide (a fast growing occupation, mind you), with an associate’s degree in gerontology, becomes a higher paid, higher status home health care provider, perhaps a manager in her field.
But where does that leave us as to who should go to four-year college? As per David’s piece today, pretty much anyone who can make it through, but that, in fact, is where the sleeper challenge is invoked.
If policy makers sincerely want to increase the share of the workforce with at least a four-year college degree, we need to worry as much, if not more, about completion than access. Policy, especially tuition-based policy, has always stressed the access margin—helping kids who face barriers to higher ed get into school. But the share of college-educated workers is a product of three probabilities: that of finishing high school (around 90%), getting into college (around 60%), and completing (also around 60%)—those are all off the top of my head but I believe they’re ballpark.
In other words, the completion margin is no larger than the access margin and logically deserves as much attention. In fact, if we follow David’s prescription, I’m certain the completion margin would become even more important, as more kids from the lower side of the “discontinuity” cutoff go to 4-year schools.
Which policies help boost completion? Income-based supports are important, especially among older students who have to finance their lives while in school. Also remedial supports, including not just academic counseling by “completion counseling” for those inadequately prepared.
The second motivation’s for David’s pushback is the type of argument you hear from Paul Krugman, Larry Mishel, and sometimes myself: college education will not insulate you from the same weak demand trends that have beset many wage earners for years. Here David’s essay does not say enough.
It is true—and importantly so—that the college wage advantage is at or near an all-time high. But it does not appear to be rising much, and EPI has extensive data showing stagnant real wages among college graduates (to be precise, those with 4-year degrees; those with advanced degrees have done better but they’re still only 12% of the labor force). Weekly earnings data from Autor show flat 4-year college earnings since 2000 (more so for men than women).
Here’s how I squared these two sides of the story in a recent report (not yet out):
…the education/wage debate needs clarification. On the one hand, there clearly exists a positive wage gradient by education level. On the other, the [get more skills, earn higher wages] story is incomplete in that more education alone won’t solve the [problems of wage stagnation and wage dispersion]. It is not hard, however, to square these observations. On average, an individual is better off with more education or training, much as marginal product theory would predict. But a) that doesn’t inoculate him or her from stagnant trends within educational classes and b) it doesn’t speak to the wage needs of those who are not likely or able to move up the education ladder. A comprehensive wage policy agenda must be mindful of all of these nuances.
I’ve tried to stress both Leonhardt (skills) and Krugman/Mishel (power), though I’ve probably leaned more heavily on the latter because a) education has many more allies than power, and b) power, for me, is in part of function of labor demand and its absence is intimately related to the persistent absence of full employment in the US labor market since the 1980s, a realization that is too often underappreciated. The education premium can be larger than the unemployment premium for a given person, but the full-employment wage premium—the wage boost to middle and low-wage workers from very low unemployment—reaches tens of millions of workers…and vice versa. Moreover, the two are of course related: weak labor markets hurt the earnings power of college-educated workers too.
And to be clear, and fair, for all their emphases, all of the people I’m citing—Paul, Larry, David–appreciate the importance of the other thing. Let’s avoid the DC thing where someone says “X is important, but don’t forget Y,” everyone jumps on them for not caring about X.
I was talking with a colleague the other day about ways to get wages up.
“Productivity” he said.
“Bargaining power,” I responded.
“Skills!” he said, getting excited.
“Unions!” I returned, matching his urgency.
“Raise the overtime salary threshold!” he shouted.
“Absolutely!” I shouted back, “and raise the minimum wage!”
“Tax breaks for job creators!” he yelled.
“Bullsh__!” I yelled back.
Well, except for that last part, it’s all of the above. Wages for most workers have been stagnant for decades, and we’re going to have to tirelessly fight for everything helpful to change that.