Are transitional insurance plans driving losses on state exchanges?

The ACA's insurance exchanges have generally been successful in helping to reduce the ranks of uninsured in the United States, while providing reasonable quality insurance to its policyholders.

However, several challenges still remain. Most notably, a good number of insurers have had problems with breaking even on the exchange products (though some are making money). That stems from several problems, notably Marco Rubio blowing up the risk corridor reinsurance program that undermined many insurer's business plans, especially non-profit co-ops trying to break into the market. Other insurers, like United Health, have just tended to be bad at competitively designing and pricing plans.

However, the risk pools for the exchanges have also proven to be somewhat older and sicker than predicted, which has also tended to drive up prices in 2016 and will likely do so more in 2017. I haven't seen a great explanation for this other than "predicting new risk pools is hard," -- which it undoubtedly is.

However, the New England Journal of Medicine last week featured a very interesting Perspectives piece (gated unfortunately) by John Hsu that fingers grandmothered plans as the culprit.


So what are "grandmothered" plans and how might they be partially undermining the exchanges?

Back in the brouhaha preceding the roll-out of Healthcare.gov, millions of Americans getting insurance through individual plans received scary cancellation notices informing them that their plan wasn't compliant with the ACA and would not be renewed. The insurance companies then offered to sell the a compliant plan -- for a much higher price.

Of course, the plans being cancelled were so cheap (for young healthy people) because they were generally lousy insurance -- they excluded essential services, had obscenely low pay-out caps and any number of other consumer-unfriendly features that the ACA banned.

And naturally, the companies didn't inform the consumers that a similar or better plan, much cheaper -- and often eligible for government subsidies would be available on the state exchange.

But Obama had promised that "you could keep your plan" and a huge political firestorm ensued. So, the administration allowed states to continue (or in Charles Gaba's term "grandmother" in non-compliant individual plans as long as they didn't change.  39 states permitted the grandfathered plans to continue, 11 and Washington DC didn't. Gaba reports now only 35 states do.

Hsu split out the states into two groups based on this variable and found that states that banned grandfathered plans lost an average of 55 percent less per enrollee on the individual market ($493 to $222) than those states that permitted them in 2014.

This makes  sense: the transitional plans likely covered healthy people cheaply, keeping them out of the exchanges where they would cover the claims for sicker people. As a result, prices will have to increase. Government subsidies and the steady ending of the transitional plans will hopefully prevent a death spiral (knock on wood) but prices jumped in 2016  and will again in 2017 as insurers adjust.

Another point in favor of Hsu's argument is that the 2016 benchmark Silver plan (on which exchange subsidies are based) increased by 12 percent in states that permitted transitional plans -- and only 5 percent in those that didn't.

Hsu's piece isn't particularly methodologically rigorous: I suspect the states that have phased out grandfathered plans are blue states also have attempted to promote the exchanges much more vigorously and do other things to generally manage their insurance markets more effectively so I'd want to see at least some simple regression analysis to weed out potential confounders.

Also, though he does look at per-beneficiary losses as a key variable, some of his stats, are, well, bunk (he notes that 90 percent of the losses occurred in the states that permitted grandfathering to continue, but doesn't explicitly note that roughly 80 percent of the states permitted plans, nor does he control for state population, though fortunately his per-enrollee numbers do.)

Still though even if I don't think he has enough evidence to get a conviction, Hsu definitely has probable cause to get a search warrant for a more rigorous follow-up study.

Incidentally, Gaba has been performing his usual yeoman work trying to get a handle on how many of these plans are out there and how many people they cover. I and my half-dozen loyal readers (Hi Mom!) would be interested to see him extend Hsu's analysis (since he clearly has nothing else more important to do with his time.)