Do subsidies protect Obamacare against the adverse selection death spiral?

Jonathan Cohn writes:

What you may not realize (because few people do) is that the subsidies, by design, protect people from rising premiums. The law basically dictates what these folks pay for the typical, “silver-level” Obamacare plan, no matter what the insurer charges. This is critical. It means that rising premiums won’t affect the willingness of those people to enroll—which means, in turn, they’d still have incentive to sign up next year, as long as the technological bugs were gone and Obamacare online was working. (Subsides were a missing element of those ill-fated reform experiments in New Jersey and elsewhere.)

The economics here are tricky.  Insurance companies set prices both for those who receive subsidies and for those who do not.  Furthermore, the subsidy — when there is a subsidy — is determined by a process akin to a second-price auction, rather than matching the highest price in the market.  (How much collusion is there anyway, once all these prices are posted?)

One question is this: pre- “pressures for adverse selection death spiral,” where is the price sitting?  I don’t see an a priori answer to this query, so let’s work through two possibilities.

One option is that, at the margin, the price is already high enough that further price hikes would lower insurer profits and subsidies won’t make up for enough of that difference.  So the scenario goes like this.  A smaller number of “invincibles” sign up early on than initially had been expected, in part because of negative publicity about the exchanges.  Providers respond by lowering service quality rather than by raising posted prices.

Think of this as the “price stickiness scenario.” One big reason for holding back on the price, and instead lowering network quality, is the adverse selection problem itself.  If some of the invincibles are paying part of the price hike, higher prices will put them off.  Yet, if indeed they are invincibles, vague rumors about inferior network access may not put them off much at all.  They don’t expect to be using the network anyway.  But of course for sick people this change in quality and access will be a problem.

On top of that, might there be some stickiness in the posted price?  Many macroeconomists stress price stickiness even under normal circumstances.  And that means very often sticky in the upwards direction too, for fear of alienating customers.  Prices are all the more sticky in heavily regulated industries and in sectors which are under a good deal of policy debate and media scrutiny and where suppliers are not all that politically popular.  Prices are even stickier when there are easy ways of raising “true net price” by lowering quality, raising wait times, restricting access and delivery speed, and so on.  Those dimensions of the problem are harder for customers, regulators, and also media-mongers to monitor.

Access restrictions are also a way of checking ultimate financial risk in a way that price increases cannot be.  We can thank Joseph Stiglitz for this insight, as Joe pointed out that not only are prices sticky, but quantities can be sticky too, and risk-averse firms may wish to limit how much they are on the hook for.

I see a good chance that the price stickiness scenario holds.  And in that case the problem is not so much a price spiral but rather network quality moves to a much lower level and sits there.

I call the second and simpler scenario the “subsidies make up the difference” scenario.

In that scenario, the initial prices are low enough that they can be raised and the subsidies pick up the difference.  The companies don’t try to game the adverse selection problem with quality decreases and most would-be buyers are covered by the subsidies at the relevant margin.  The thought experiment then runs like this.  The initial quality of pool applicants suddenly worsens, but this time the main effect is that posted prices go up.  Because of the subsidies the real net prices to potential purchasers do not change very much and all still seems OK.

We do not know which scenario will occur, or to what extent, or in which states.  But I hardly think the law is in the clear in this regard.