Generally, it is more efficient both socially and privately for the consumer, in this case the student, to pay the full cost of the goods and services that he buys than for the government to pick up any part of the tab. A student admitted to an elite college like Harvard and Yale has high expected lifetime earnings, and it seems absurd that the federal taxpayer should be required to defray a part of the cost of his education. This is not a point about distributive justice, for nowadays most federal income tax is paid by high-income individuals. It is a point about the inefficiency of using the federal tax and spending power to subsidize purchases by affluent (as measured by expected earnings over the recipient's lifetime) consumers. The inefficiency lies not only in the transaction costs associated with the subsidy (including lobbying expenses), but also and more importantly in distorting the allocation of resources. Suppose that for some marginal student the expected return from a college education, net of tuition and opportunity cost (the forgone income from working if the student attends college instead), is negative, but turns positive if his tuition expense is subsidized; then the subsidy is inducing a waste of resources. This would be obvious if the subsidy were for a course in automobile repair, but maybe there is something special about college or university education that distinguishes it from other services, including other educational services. There are two arguments. One is that higher education (lower also, presumably) confers social benefits (that is, benefits not captured by the student), whether by making people more informed voters, or by making them more productive workers (assuming they cannot capture their entire contribution to social output in their wage), or by reducing subsidized health costs by increasing health (Becker notes that educated people are healthier than uneducated people). This is not a good argument for a subsidy because it does not appear that many persons who would benefit from a college education fail to obtain one. As Becker points out, the private returns (higher earnings) from a college degree are very great and a student can borrow to finance the tuition and other costs of the degree. It seems unlikely, though it is not impossible, that kids who would not personally benefit from college nevertheless would, if paid to go to college, confer the social benefits of a college education that the students who do benefit personally might be thought to confer. But the points I have made so far really argue just against increasing the existing subsidies for college education, rather than against any subsidies. College education is already heavily subsidized, notably in the case of state and city colleges, where the taxpayer picks up a big share of the cost; but private colleges receive various tax breaks, so they are subsidized too. (I would not call alumni donations "subsidies," however, since they are voluntary and give value to the donor.) Since a worker usually cannot recapture in his earnings the full effect of his labor on output (because he produces some consumer as well as producer surplus), and college increases the productivity of those students who are intelligent enough to benefit from a college education, there is an argument for making college affordable by any qualified applicant. However, it is unclear to me whether this requires any subsidies; all that is required is that the boost in expected earnings from attending college exceed the cost of the loans, or other costs, that the student must incur for college to be a rumerative choice. For then the student will be motivated to attend college even though his doing so will produce social as well as private benefits. All that is important from the motivational standpoint is that the private benefits exceed the private costs. The second argument for subsidizing higher education is that its high cost nowadays, which for students who must borrow to pay tuition and living expenses forces them to go into debt, deflects students from nonremunerative jobs, such as (in the case of debt-written law students) public interest legal practice, or public school teaching, that (especially teaching) may confer substantial social benefits. (I doubt that public interest law practice does.) The students have too much debt to be able to pay it off without taking a high-paying job. However, a student loan subsidy is a clumsy device for channeling students into employment that is underpaid from a social standpoint, since every student gets the subsidy but only a handful are induced by it to enter the desired channel. It would be more efficient to raise the pay for the jobs that are thought to confer social benefits. A loan-forgiveness program, where forgiveness is conditioned on taking one of the favored jobs, is better tailored to the end of encouraging students to take such jobs than a loan subsidy; it operates to raise the full income of the job. To repeat an earlier point, which tends to be neglected in discussions of the student-loan issue, if I am right that very few persons who could benefit from a college education are deterred by its cost, the main effect of increasing the subsidy will be to attract applicants who would not benefit if they weren't being "paid" to attend college. That would be a misallocation of resources. I conclude that the case--for which I gather there will be support in the new Congress--for increasing the student-loan subsidy by having the federal government subsidize a larger part of the interest on the loan is a weak one.