Tuesday, February 18, 2014

The (Fixable) Problem with Pay It Forward



Cross-posted from the Hamilton Project blog:

Michigan has now joined Oregon in proposing a “Pay It Forward” student lending system in which students pay no tuition up front and pay back a fixed percentage of their income after college. This sounds very similar to an income-contingent repayment (ICR) system, which I advocate in a recent Hamilton Project proposal. But there is a key difference between Pay It Forward and ICR, and it’s one that makes Pay it Forward unworkable as proposed.

In both the Michigan and Oregon versions of Pay It Forward a borrower in repayment contributes a fixed percentage of income for a fixed number of years. Her liability is not denominated in dollars, as in a standard loan, but as a fixed number of payments. 

An aside on vocabulary: economists call this a graduate tax –a tax on earnings for those who have gone to college.  It’s called a tax, rather than a payment, because a borrower can’t buy her way out of the liability. The borrower is taxed for 25 years, even if she has repaid the principal (plus interest) after a few years.

In the proposed Pay It Forward systems, a graduate who does extremely well in the labor market will end up repaying many times over the cost of her education, while one who does poorly will pay much less. There is therefore cross-subsidization in this system, with the “winners” paying some of the college costs of the “losers.”

Economic theory – and history – shows that loans funded by a graduate tax won’t work because those expecting high earnings won’t participate. Yale famously attempted a graduate tax in the 1970s, lending money to its undergraduates and then having them pay back a fixed percentage of their income for a fixed number of years. What happened? Yale students who expected high earnings (e.g., aspiring investment bankers) shunned the program, while those who expected low earnings (e.g., aspiring artists) embraced it. Yale’s program spiraled into insolvency.

This is a classic case of adverse selection – borrowers who would be subsidized participate while those who would subsidize stay away. This is unsustainable, as without the high earners the system does not get enough payments to cover tuition costs. Because of adverse selection graduate tax can work only if participation is mandatory, with everyone forced into the borrowing pool.  
Another interesting aside: this is similar to the dynamic in insurance markets, which collapse if sick people buy coverage and healthy ones go without. Adverse selection is why coverage is mandatory under the Affordable Care Act, and it’s why health policy wonks have so carefully tracked the enrollment of young, healthy people in the new insurance exchanges. The young, healthy participants cross-subsidize the older, sicker participants, just as high earners subsidize low earners in a (mandatory) graduate tax.

Neither the Oregon nor Michigan plan requires all borrowers to participate in their programs. I (and other economists) therefore predict that the programs, as currently proposed, will be brought down by adverse selection.

A minor tweak to Pay It Forward, as outlined in my Hamilton proposal, will maintain its positives (simplicity, insurance against bad draws in the labor market) while eliminating the negative (unsustainability). The change is this: denominate debt in dollars, and let borrowers pay their debt. If a student borrows $25,000 and (due to pluck and luck) earns enough that she has paid back the principal plus interest after just ten years, she will stop paying into the program. 

If a borrower instead runs into hard times and still owes money after 25 years, the balance will be forgiven. In this way, Pay It Forward and my income-contingent repayment proposal both subsidize low earners.  The key is that my modification keeps high earners from fleeing the program, transforming Pay It Forward it into a universally attractive program rather than one that appeals primarily to low earners. 
As pointed out by Theda Skocpol, universal programs are politically more resilient than those that primarily benefit poor people. Keeping high earners in Pay It Forward will give it the broad-based political support it needs to survive. 

My tweak also makes transparent the cost of Pay It Forward. Some borrowers cannot, given their low earnings, pay off their loans. Their forgiven debt is the cost of the program, and is borne by all taxpayers.  Without my tweak, the costs of Pay It Forward are disguised by the fairytale that it is self-funding. When high earners shun the program and it collapses, the taxpayers will be on the hook to bail it out. The financial cost, in the end, is the same, but the political cost is much higher, since the program will be deemed an expensive failure.  

With the political momentum behind Pay It Forward, we have a rare opportunity to restructure and reimagine how we pay for college. Get the design right, and we will have a financially and politically sustainable system for funding college that works for students and taxpayers. Get the design wrong, and we will have a spectacular flameout that sets back reform for another generation.