How much will it cost to restabilize the Exchange insurance markets?

Short answer: My best order-of-magnitude estimate is between $500 million and $1 billion for the coming year of which a third to a half could be attributed to the President’s decision to honor his promise to let Americans keep their existing health insurance.

Insurers are rightfully complaining that the move by the President to fulfill a promise he made to permit Americans with “substandard” but previously grandfathered policies to keep their health insurance is going to destabilize insurance markets.  There were such complaints going in to a meeting on November 15 between President Obama and selected insurance leaders and there were somewhat muted complaints coming out of the meeting. Insurers are concerned because the people who are now being given access to another market in which insurance policies may be cheaper are likely to be precisely the healthy people that insurers who wrote policies in the Exchanges assumed would be in those Exchanges.  Their concerns are important because unhappy and unprofitable insurers have a tendency either to stop writing insurance or to raise rates.  That hurts policyholders and it also hurts politicians who assured the public that the rates would be affordable. (The insurers are also upset because it’s a little challenging to uncancel policies on short notice, but we’ll leave that grievance for others right now.)

The instrument by which some are proposing to pacify the insurance industry for the surprise deprivation of healthy insureds is the hitherto obscure “Risk Corridors” provision baked into the Affordable Care Act (section 1342, 42 U.S.C. § 18062, for those scoring at home). It provides that the government cover up to 80% of losses an insurer incurs on an Exchange. It was always assumed — foolishly in my opinion, but assumed nonetheless – that this backstop would be costless because the government would also effectively tax up to 80% of profits via the same provision. If the insurers systematically lose money, however, because many of the people they thought would improve the Exchange pools with their good health are being given an option to separate themselves out and keep their old often-less-expensive and often-less-generous insurance policies, the Risk Corridors provision could cost the government a fair amount of money.

So, the question is, how much money is Risk Corridors likely to cost? To use the language from my prior post, how much VOOM?  If it’s a relatively small amount, that would suggest that the President (and others’) proposal to honor a campaign commitment to let people who liked their health plans keep them is a better idea than if it’s a relatively large amount of previously unbudgeted money. I thought we might try a back of the envelope computation to see what’s involved.

Time to trot out some calculus.  The Risk Corridors provision basically creates a mathematical function between profitability (as defined in that provision) and the size of a positive or negative transfer payment from the government to insurers writing policies in the Exchange.  So, if we knew the distribution of profitability of insurers under the Exchange we would calculate the mean payment (an “expectation” for those with some statistics background) the government would make (or receive). Of course, we don’t know that distribution yet, but we can make some guesses and get some order-of-magnitude estimates.

If one assumes that the distribution of the ratio between claims and premiums has a mean value of one (i.e. that insurers on average break even), the the expected payment of the government is zero.  That’s the assumption on which the Congressional Budget Office worked when it asserted that Risk Corridors would cost nothing. But what if one assumes that the distribution of the ratio between claims and premiums has a mean value of 1.1, i.e. insurers on average lose 10%.  We’ll also assume for the moment that the distribution of the ratio is “log normal” and that 95% of insurers have a claims/premiums ratio of between 0.922 and 1.22. If we do the math — here’s the link to the Mathematica notebook that stands behind these computations — it turns out that the average payment of the government is about 3% of the average premium (before subsidies).  If the mean of the distribution were 0.5, i.e. insurers on average have claims 50% higher than profits, and we hold everything else the same, the average payment of the government is about 34% of the average premium (again, before subsidies). So if, just for the sake of discussion, one assumed there were 2 million people in the Exchanges and that the average gross premium was $3,500, the government would end up shelling out $210 million per year to provide insurers with some relief if they lose 10% on average and would end up shelling out $2.37 billion per year to provide insurers with similar relief if they lose 50% on average.

The graphic below shows the size of the government’s Risk Corridors obligation as a function of the mean of the claims/premiums ratio under the continued assumption that the distribution is log normal and that the spread of the distribution is similar to that described above. With a little wiggle when the mean of the claims/premium ratio is close to one, the relationship is pretty linear.

 

Relationship between mean insurer claims/premiums and risk corridor payments
Relationship between mean insurer claims/premiums and risk corridor payments

To get the total bill for the government, however, we not only have to calculate risk corridor payments in relation to a premium amount, we also have to make a guess about how many people will enroll in the Exchanges and what their premiums will be.  It’s complicated because, precisely because of adverse selection, there’s likely an inverse relationship between the number of people that enroll and the mean of the claims/premiums ratio.  But since all we are trying to do here is get some order of magnitude estimates — the discussion of this Act has been hurt all along by false claims of precision — we can try to make some reasonable guesses.

Suppose, for example, that the relationship between the mean of the claims/premium distribution and the number of people enrolling in the Exchanges looks something like this.

Hypothesized ratio between enrollment and mean of claims/premium distribution
Hypothesized ratio between enrollment and mean of claims/premium distribution

What we can now do is graph the government’s overall risk corridor payments as a function of enrollment.  I’m going to assume that the average premium is $4,000 per enrollee.  That’s roughly the average $328 per month that Kathleen Sebelius reported for a silver plan.  If people flock to the gold and platinum plans, the average could be somewhat higher. This graph is essentially the headline of this blog entry.

Hypothetical relationship between enrollment and risk corridor payments
Hypothetical relationship between enrollment and risk corridor payments

So, what we we see is that if, for example, enrollment for this year were to be 1 million, the total risk corridor payments might be somewhat in excess of $1 billion. If enrollment were 2 million, risk corridor payments might be $500 million.  One enrollment crosses 3 million, the government actually could gain money via the risk corridors program.

There are a lot of unknowns going in to the graphic above.  I do not pretend that it is precise.  I do not even contend that it is accurate.  Nonetheless, I believe it is useful.  I do believe it provides a plausible order-of-magnitude estimate of an unforseen cost of the Affordable Care Act.  If you asked for my best guess, I would tell you the Risk Corridor payments will likely be between $500 million and $1 billion this coming year as I would guess enrollment in the Exchanges will come out between 1 and 2 million (assuming they ever fix healthcare.gov).  This does not mean, by the way, that the cost of the President’s fix (or of the similar bills now in Congress) is the full amount of the Risk Corridor payments. Some of these risk corridor payments might have been made even without the Obamafix. That is so because enrollments in the Exchanges may always have been overestimated and may have been made considerably lower as a result of all the fallout from the debacle of the healthcare.gov website rollout.

In the end, then, I suspect that for the coming year the price tag for the President keeping his promise that “If you like your health plan, you can keep your health plan” is going to be somewhere in the $200 million to $400 million range for the coming year.  That’s about a third of the overall stabilization bill. And we’ll never know for sure because we won’t know how many of those that in fact do keep their health plan would have enrolled on the Exchange.  In one sense, the money cited above may be seen as a rather inexpensive price to pay to make good on an alluring promise.  On the other hand, it may also be seen as yet another unforeseen or unadvertised cost of a bill to transform American healthcare. It’s easy to make feel-good campaign promises when you aren’t fully honest about the cost.

Share Button