I wrote yesterday that the administration is wooing insurers to stay the course on the Patient Protection and Affordable Care Act. It’s been reported that the administration was looking at ways to tweak the risk corridor and reinsurance program -- a system of side payments meant to help defray the transition costs as insurers sold policies to an unfamiliar market -- in order to give the insurers more money. I was skeptical that it would be able to do much with the risk corridors, because the amounts they’re allowed to pay are written into the law. But it may have found more fertile ground in risk-adjustment payments, another reinsurance mechanism:

The U.S. government has issued a proposal that would likely increase risk payments in 2014 to health insurers offering plans on the Obamacare exchanges after the companies complained a recent policy change allowing people to keep their insurance policies had changed the financial equation.

The rule, published on Monday in the Federal Register, lowered the threshold at which risk payments kick in for the sickest health plan members. The government proposed paying insurers 80 percent of claims greater than $45,000 in 2014. Previously the lower limit was $60,000.

Here’s the new rule in Federal Register-ese:

In the 2014 Payment Notice, we finalized the following uniform reinsurance payment parameters for the 2014 benefit year -- a $60,000 attachment point, a $250,000 reinsurance cap, and an 80 percent coinsurance rate. However, updated information, including the actual premiums for reinsurance-eligible plans, as well as recent policy changes, suggest that our prior estimates of the payment parameters may overestimate the total covered claims costs of individuals enrolled in reinsurance-eligible plans in 2014. To account for this, we propose to decrease the 2014 attachment point to $45,000. We seek comment on this proposal.

So how much is this change going to cost? I dug into data from the Medical Expenditure Panel Survey, one of the best sources on who spends what on health care.* The good news is that more than 98 percent of people have health expenditures under $45,000 a year, and more than 99 percent of the claims paid by private insurance are under that amount. (At least in the unweighted survey data, 33 percent of expenditures over $45,000 were paid by private insurance, about 30 percent were paid by Medicare, 22 percent were paid by Medicaid, and the rest were paid through some combination of the patients and other state or local programs.)

The bad news is that the fewer than 2 percent of patients with high medical costs accounted for about 26 percent of total expenditures. If you just look at people whose private-insurance bills exceed $45,000, you’re talking about 0.5 percent of customers but 21 percent of total insurance payouts. So even though we’re looking at a small number of patients, we’re looking at a relatively high fraction of total health-care spending, for which the risk-adjustment program will be picking up 80 percent of the tab.

OK, but that was already going to happen for folks whose bills exceeded $60,000. How much will it cost us to add the folks whose bills are $45,000 to $60,000?

Fun fact: Because high medical bills are pretty rare, and the higher they get, the less common they are, there are almost as many people with total expenditures in the narrow band of $45,000 to $60,000 as there are people with total expenditures of any amount above $60,000. MEPS estimates that there were about 4.1 million people with total health-care expenditures above $45,000 in 2011: 2 million people at $45,000 to $60,000, and 2 million above that number. About 1.4 million private-insurance patients had claims of at least $45,000 in that year, of which roughly 45 percent fell in the $45,000 to $60,000 range and the rest somewhere above $60,000. So this represents a substantial increase in the total number of people who will have most of their bill picked up by Uncle Sam.

But while the number of claims covered by the program will almost double, that doesn’t mean the cost to the federal government will also double. In the data I looked at, patients in the expansion group had an average expenditure of about $50,000, while patients in the original group had an average expenditure of more than twice that. If the population in the exchanges roughly mirrors the population at large, it looks to me as if the cost of the risk-adjustment payments will increase by about 40 percent thanks to this rule change.

What we don’t know -- what we can’t know -- is what that cost will be, because that all depends on how many very sick people end up on exchange policies. Private insurers spend more than $45,000 on only about 0.5 percent of their current customers. To know how much this program will cost, we need two pieces of data:

  1. How will the insurance pool differ from the current pool?
  2. How many people will buy policies on the exchanges?

Obviously, we don’t know those things. But I've made a table laying out the costs under various assumptions:

To be sure, this assumes that the folks whose claims will be defrayed under the new rule still have an average claim of about $50,000, which may or may not be true. And some of the combinations on the matrix aren’t very likely: If 8 million people buy insurance on the exchanges, then it seems unlikely that 5 percent of them will be very sick. But under some reasonably conservative assumptions -- 4 million new users on the exchange, with three times as many very sick people as normal -- then by my calculation, the new rule could end up costing as much as $1 billion a year.

That’s not a huge amount in the context of a health-care bill that spends hundreds of billions of dollars. And, as the Federal Register notes, it takes place in the context of a broader proposal to mandate that all reinsurance collections are paid out every year:

For example, for 2014, if HHS collects $9 billion for the reinsurance payments pool and $10 billion in reinsurance payments are requested, HHS and each applicable reinsurance entity would reduce all reinsurance payments by 10 percent (effectively decreasing the coinsurance rate). If HHS collects $11 billion for the reinsurance payments pool and $10 billion in reinsurance payments are requested, HHS and each applicable reinsurance entity would increase all reinsurance payments by 10 percent (effectively increasing the coinsurance rate).

We seek comment on this payment proposal, including on whether any excess collections should be allocated to increasing coinsurance rates above 100 percent, or whether such funds should be used instead to change other reinsurance parameters or used for future benefit years.

Still, you know, “a billion here, a billion there, and pretty soon you’re talking real money,” as Senator Everett Dirksen may (or may not have) said. The insurers may think that $1 billion is very real indeed -- and that they’d rather have their reinsurance premiums paid back to them in full this year than have any overpayments used to reduce premiums in the future. It’s probably not a mega-giveaway to the insurers. But it probably does help the administration keep them on board.

*A few caveats: The data I have are unweighted, and they include uncovered expenditures such as dental care and plastic surgery. But these are a pretty small overall fraction of expenditures for people with very high bills, so this shouldn’t change the total number much.