Is Higher Education a Good Investment?--Posner's Comment
Becker marshals convincing evidence that people who have more education have on average higher earnings and that the spread has been growing. But it is a bit of a leap to conclude that there are high (and increasing) returns to education. Correlation is not causation. Suppose what are increasing are not the returns to education but the returns to intelligence, and suppose that people with high IQs both enjoy education more than other people do and are more likely to be admitted to college or a graduate or professional school because teachers prefer teaching (and learning from!) them and because good students are more likely (because they are more intelligent, not because they are good students) to be affluent, and therefore generous, alumni.
Now if this is correct, one might expect many intelligent people to bypass college, because it is so costly; but few do. However, colleges and graduate (including professional) schools provide a screening and certifying function. Someone who graduates with good grades from a good college demonstrates intelligence more convincingly than if he simply tells a potential employer that he's smart; and he also demonstrates a degree of discipline and docility, valuable to employers, that a good performance on an IQ test would not demonstrate. (This is an important point; if all colleges did was separate the smart from the less smart, college would be an inefficient alternative to simple testing.) An apprentice system would be a substitute (and there is evidence that in Germany it is a highly efficient substitute), but employers naturally prefer to shift a portion of the cost of screening potential employees to colleges and universities. Because those institutions are supported by taxpayers and alumni as well as by students, employers do not bear the full cost of screening.
These points are consistent with higher education being a good private investment, but do not suggest that it is either a particularly good social investment (it does improve matching of employees to employers, but at great cost) or that its value has much to do with the institution's educational program.
Another good that higher (or for that matter lower) education provides is the creation of social networks, consisting of the students who get to know each other. They learn something from associating with other intelligent kids and they form friendships with them that may carry over into adult life and become business or professional relationships that enhance the graduate's income. Again that would be a benefit having little or no connection with the school's educational program. There may be little value added by the program to the contribution that attending college makes to a person‚Äôs income.
Orley Ashenfelter, Alan Krueger, and many other economists have worried about the possibility that the correlation between education and earnings is not causal and have tried a variety of ingenious methods for correcting for differences in ability, such as comparing the earnings of twins who have different amounts of education, on the theory that they have similar native ability, or comparing the earnings of people who have different years of schooling just as a function of the arbitrary age cutoffs that determine when one starts school, or seeing whether an increase in the age at which students are permitted to drop out is associated with an increase in the earnings of that cohort compared with its predecessors. Some studies correct for performance on standardized tests assumed to measure intelligence rather than knowledge. Most studies find that education has a substantial effect on earnings independent of native ability, and the convergence is impressive. However, the studies are convincing mainly about the benefits of precollege education.
I am skeptical that it should be a national priority, or perhaps any concern at all, to increase the number of people who attend or graduate from college. Presumably the college drop-outs, and the kids who don't go to college at all, do not expect further education to create benefits commensurate with the cost, including the foregone earnings from starting work earlier. This would be an entirely rational decision for someone who was not particularly intelligent and who did not anticipate network benefits from continued schooling because the students with whom he would associate would not form a valuable network of which he would be a part, either because he could not get into a good school, in the sense of one populated by highly promising students, or because if he did get into a good school the other students in the school would not consider him worth networking with.
This assumes that enticing the unwilling or the unmotivated to attend or complete college would not confer social benefits in excess of the private benefits (which I suggested in the preceding paragraph would probably exceed the private costs). But the marginal students are unlikely to be kids who, with a little more education, would make the kind of contribution to society that a worker is unable to capture in his wage. Nor are these marginal students likely to be educated into an interest in political and societal matters that will make them more conscientious voters or otherwise better citizens.