Thursday, April 17, 2014
Slate’s Jordan Weissman has an interesting story on CBO’s projections for the profitability of the three main federal direct student lending programs. The projections are reasonably big -- about $15 billion per year, of which $12 billion comes from graduate (including law school) loans. That’s about the same money we currently give up by allowing homeowners to deduct the cost of their mortgages (see here and scroll down to p. 35). Does that combination of borrowing rules make sense? Should government profit off of some people to give a tax break to others?
I think of the student loan programs as a really big bake sale.
Whether you’re using it to pay for schools or bombers, the bake sale is a way of replacing tax revenue with sales revenue. Taxes distort the economy, encouraging folks to do stuff they usually wouldn’t just to avoid the tax. In contrast, it’s often safe to assume that a rational consumer shelling out a few bucks for banana bread perceives herself as being better off as a result of the transaction. So bake sales or “user fees” look like a swell way to raise money along this dimension (unless they provoke an armed standoff, but let’s assume most folks aren’t up in arms about dessert).
But bake sales are really only at their tastiest when there is a private market failure. Presumably the bake sale crowds out private entrepreneurs who sell similar products. For lots of very familiar reasons, it’s likely that shift from private to public also reduces overall welfare -- maybe our bake sale agency is less likely to invent the cronut, for example. Probably those welfare losses are much larger than the gains we could achieve by cutting taxes. But since this crowding-out problem by definition doesn’t happen when the private market is failing to produce a given set of goods (or failing to produce them at a particular level of quality), the bake sale looks better by comparison in that circumstance. Maybe this describes the market for student loans, which most people think is usually crippled by asymmetric information between borrower and lender.
So perhaps government should try to turn a profit on lending, but how much of a profit? Since the government has the field to itself (unless it also subsidizes private or rival government alternatives), let’s think about it as a monopolist. Monopolists usually maximize revenues, not output. Should government charge the revenue-maximizing price, or the output-maximizing one? Imagine that each additional unit of output produces consumer surplus and/or positive externalities, at the cost of increased taxes. The optimal price, then, depends on a standard cost-benefit analysis: given the government’s marginal cost of substitute funds, is it worth it to produce the marginal unit beyond the revenue-maximizing point?
Putting this another way, whether government should turn a $15 billion profit from student lending depends on how much we’d value even more student lending. Probably we don’t value the lending as such, but instead value the education it buys. That’s tough to put a price on, of course. Even tougher: how much education does the government actually purchase when it offers cheap loans? Some of the value may be captured by educators (cough, cough) or by infra-marginal borrowers who could have self-financed their education. Maybe the government isn’t profiting enough right now.
On the other hand, the distribution of burdens under a tax and a bake sale may differ. The most obvious example is government-run lotteries, which strike a lot of observers as being highly regressive. I’ll return to these distributive questions in another post.