Disruption and Adverse Selection

This one gets a little wonky, but the subject matters.

In reflecting on last week’s Baltimore conference, I keep circling back to a couple of points. One is about “white-labeling,” which I will address another day. The other is adverse selection.

My first encounter with the concept came in reading about the politics of health insurance. Insurance works by pooling risk; everybody pays in a set amount, and payments go out based on events that are predictable in the aggregate but often unpredictable for an individual. Any given person can get hit by a bus on any given day, but over time and large populations, it’s possible to figure out the typical number and cost of bus accidents. Anybody can die on any given day, but over time and large numbers, the likelihood of death on a given day is higher for the average 90-year-old than the average 30-year-old.  

Insurance companies need enough 30 year olds in the pool to cover the 90-year-olds. Or, if you look at it from a profit-maximizing perspective, they want as many 30-year-olds as they can get, and as few 90-year-olds as they can manage. Loading up the pool with low-risk folks and excluding high-risk folks makes it much easier to turn a profit.

To the extent that it makes sense to refer to an American health care “system,” our system is a mix of public and private. We have Medicaid and Medicare, the VA, and various state-level coverages for the poor on the “public” side; on the “private” side, we have health-insurance companies and various for-profit providers. (As with higher ed, there’s also a significant cluster of private but not-for-profit players who draw on aspects of both.)  

In a purely public system, such as single-payer, everybody is in one risk pool. In a hypothetical purely private system, everybody would be in what they pay for, or not.  We have a mix, which creates a serious distortion. Private providers can try to cherry-pick the folks they’ll cover, maximizing the low-risk pool (i.e. the 30-year-olds) and dodging the high-risk folks. They can do that through how they market themselves, what they cover or don’t, how they deal with pre-existing conditions, and the like. They can get away with that because the public systems exist as fallbacks.  

But when the private providers pick off the healthiest members of the risk pool, they artificially raise the average cost to the public providers. If everybody in your risk pool is 90, good luck staying solvent. Then, folks bankrolled in part by the profits of the private providers run campaigns against the inefficiency of the public system, thereby doubling down on the adverse selection death spiral. Over time, as a society, we pay progressively more for progressively less or worse coverage, while people actually get sicker. The alternative -- single-payer -- is so obvious that it requires a sustained ideological assault to keep it at bay. And that happens.

I bring this up because some of the “disruptors” at the conference seem to be positioning themselves in the same way that private insurers do. They’re trying to cherry-pick the students who are the least expensive to teach. To the extent that they succeed, they’ll artificially inflate the per-student cost in the public system. Their business models depend on sloughing off the students who need more time, or more help, or more human contact.  

That’s an existential threat to the public providers, since we employ internal cross-subsidies to cover the costs of the most expensive students. Lose those cross-subsidies, and we lose the ability to provide for the students who need the most. The decline of operating funding is bad enough; pick off enrollments, and the fiscal decline is even worse. And that’s without even considering the history of what happens to public funding when programs are identified as being for the poor.

In the case of providers that pick off courses, rather than groups of students, the damage is both subtler and more radical. StraighterLine, for instance, offers cheap gen eds, leaving the more expensive lab and specialized courses to the publics. But we use the revenues from those gen eds to fund the more expensive courses. Siphon off that revenue, and we won’t even be able to grow our way out. That will lead to price increases and performance declines, which, in turn, justify even greater "disruptions."

This isn’t an argument against innovation at all. It's an argument against mistaking piracy for reinvention. If public institutions are going to thrive, they need large risk pools. That is to say, the closer they get to universalism, the better. That means resisting being cast as “fallbacks.” Public institutions need to be good enough to attract people who have choices. That may mean trying some things we haven’t tried in the past, and rethinking some longstanding "givens."

We need to innovate, but we need to do it in the direction of inclusiveness. Cherry-picking is easy; universalism is harder. But it’s our best shot at survival.