Friday, May 19, 2017

CMS's warmup for AHCA-world: Guidance on ACA innovation waivers

CMS recently issued guidance to states seeking ACA Section 1332 "innovation waivers," by which states can apply to alter or even thoroughly redesign their ACA marketplaces.

The checklist offers explicit encouragement to states to seek federal funding for reinsurance programs or high risk pools:
In particular we welcome the opportunity to work with states to pursue Section 1332 waivers incorporating a high-risk pool/state-operated reinsurance program. State-operated reinsurance programs have a demonstrated ability to help lower premiums, and if the state shows a reduction in federal spending on premium tax credits a state could receive Federal pass-through funding to help fund the state’s reinsurance program. 
Encouragement to states to implement a reinsurance program is good news, as reinsurance does keep down premiums, and the expiration of the ACA's federal reinsurance program after 2016 contributed to the 2017 premium spike. Including the high risk pool (HRP) option is a bit of a mystery, since 1332 waivers cannot be used to waive the ACA's ban on medical underwriting or guaranteed issue. (The waivers can be used to alter the ACA's Essential Health Benefits required of all qualified health plans, as well as subsidy formulas and the individual and employer mandates.)

It seems, then, that the HRPs encouraged here would have to be something like the "invisible risk pools" for which funding was tacked onto the AHCA prior to House passage. In an invisible risk program, insurers earmark enrollees in their ordinary offerings who have certain pre-existing conditions, forfeit the bulk of their premiums and shed responsibility for their healthcare costs above a certain threshold, the tab picked up by state and/or federal funding. High risk enrollees obtain their insurance on the same terms as other enrollees.

The limitation to invisible risk programs in the waiver checklist is not spelled out, but it seems required by the ACA. Nicholas Bagley, Professor of Law at U. Mich, concurs, writing to me, "to make a high-risk pool option work [under an ACA 1332 waiver], you'd have to work out the risk-sharing on the back end, not the front end. No waiver could allow an insurer to say 'no' or to charge more on account of someone's health status, so it's hard to see how you could shunt sicker people to a high-risk pool."

That's under the ACA, of course; the AHCA, if enacted, would allow states to opt out of the ban on medical underwriting. In general, CMS's encouragement of 1332 waiver applications may be regarded as a kind of warm-up for a post-ACA world -- witness the waiver scheme that Oklahoma is deliberating.

There's one more mystery in the checklist. While the ACA's ban on medical underwriting  does not fall within the scope of provisions that can be waived under Section 1332, the requirement that insurers maintain a single risk pool for all their qualified health plans (Sect. 1312(c)(1)) can be waived. In what Timothy Jost, Professor of Law at Washington and Lee, calls "the enigmatic footnote 5," the checklist more or less encourages states to do so:
5 For example, a state could waive Section 1312(c)(1) related to the individual market single risk pool in connection with implementation of a state-operated reinsurance program. Section 1312(c)(1) requires a health insurance issuer to consider “all enrollees in all health plans….offered by such issuer in the individual market…to be members of a single risk pool.” In its waiver application, the State would be required to explain how waiver of the single risk pool provision would facilitate the operations of and/or requirements for participation in the State’s reinsurance program or high risk pool and/or mechanism for a high risk pool in its individual insurance market in terms of its decision to implement its reinsurance program. For example, a state might explain how in order to maximize the rate-lowering impact of their proposal, the state would like to waive the single risk pool provision at 45 CFR 156.80 to the extent it would otherwise require excluding total expected state reinsurance payments when establishing the market wide index rate. 
Neither Jost nor Bagley can make head or tail of this -- and if they can't, who can? I have a query into CMS and will report if I get clarification. I asked Jost if this might envision a program in which insurers effectively set up their own invisible risk pool,s offering earmarked enrollees coverage on the same terms as everyone else, but seeking state/federal funding for the tagged invisibles. His response: "That isn’t exactly what it says, but might be what they were thinking."

Our Kremlinology is in its early stages. Standing by for CMS clarification...
Update, 5/19, 4:15 p.m.: I have been offered some clarification from Timothy Jost. Basically, the state would seek to waive the requirements of the single risk pool regulation so that each insurer can exclude state payments for reinsurance or invisible risk programs from the calculation of its index rate. That is, payments that cover the claims of the most expensive enrollees are excluded from the calculation of how much revenue the insurer needs per enrollee. In effect, those covered by state invisible risk or reinsurance payments are in a risk pool of their own for the purposes of rate-setting, though they receive insurance on the same terms as everyone else in the plans in which they enroll. That's my translation of this from Jost re Footnote 5:
 A state has to request a waiver of something in the ACA to get a 1332 waiver, and without a 1332 waiver a state cannot get access to federal pass through funding for a reinsurance program.  The single risk pool requirement under 45 cfr 156.80 allows for adjustment of the index rate based on federal risk adjustment payments and user fees.  Under footnote 5, a 1332 waiver could also allow the index rate to be based on adjustments for state reinsurance payments.   So a state can ask for a waiver to the 156.80 requirements to allow it to take state reinsurance payments into account in setting the index rate, and thus qualify for a 1332 waiver and APTC and CSR pass throughs.
Update, 9/25/17: Back in May, I neglected to add that CMS confirmed this interpretation on May 22. It's newly relevant, as ambiguously worded allowance for multiple risk pools in yesterday's updates to Graham-Cassidy may be for this purpose.


  1. A reinsurance fund reduces the mean cost of enrollees by subsidies, and the variance in the cost because the insurance company has off-loaded some of the risk of high costs to someone else, is that correct? Reducing either of those should reduce insurance premiums.

    How much of the premium increases we've seen is because enrollees are sicker than originally expected, and how much is because they're riskier than expected (ignoring for the moment CSRs and non-enforcement of mandate)?

    If the main problem is the enrollees are predictably expensive, then merely handing money to insurance companies is simpler than either reinsurance or a high risk pool, and just as good.

    1. Gee, I don't know, Anne: can "sicker-than-expected" be more or less evenly spread out? I gather that some risk pools can be badly disfigured by a handful (or less) of very expensive cases.

      It does seem crazy when you step back to be distributing money in all these complex ways. Why not just pay insurers to administer programs, like TPAs, and pay them say 2% over their operating costs?

    2. It depends on how many people you're covering, right?. If I'm covering two hundred young women, I could be wiped out by one or two "extra" premature births. If I'm covering two hundred thousand young women, I'm less worried about premature births because I think I can predict the number to good accuracy. The little insurer (or the self-insuring company) wants the reinsurance to protect against risk. The big insurer wants a pot of money to lower premiums.

    3. Of course using them as a TPA has another name, called single-payer...

      Re Anne's comment, existing ACA (subsidies tied to benchmark plan) would seem to handle the sicker but smooth case. Re reinsurance vs high-risk pool, at the risk of stating the obvious, they can both work as long as adequately funded. I think reinsurance administratively simpler but if the money is there I could live with either (note the money has never been there in the past).

    4. If you're the GOP (or, for that matter, if you're me) you want to lower premiums for everyone, including unsubsidized people. The subsidies don't do that very much.

    5. Fair - it depends on your goals. Reinsurance and risk pools are meant to smooth costs to average - if you want to lower the average cost per person for everyone, that's a different question. Nothing in the AHCA would do that. It may give the appearance it does by allowing the sale of plans with insufficient benefits (less benefits means lower premiums) but on an apples to apples basis premiums would only increase.

    6. What's the - whatever it is, $10 billion a year- going to do, except lower premiums? Either it's a giveaway to insurers and they just stick it in their pockets, or it lowers premiums/deductibles/copays/coinsurance. There is no third choice.

  2. Re: the update: The index rate says how much revenue is needed from each enrollee, which then allows the insurer to set premiums and deductibles. With a waiver, the insurer would be allowed to count the expected reinsurance payments. This makes sense; the insurer doesn't need to charge the subscribers for cost reimbursements it will be getting from the reinsurance payments.

    But what is the benefit of putting high risk people in a high risk pool, then? If Jost's interpretation is right, and I think it is, the people in the high risk pool still can't be charged higher premiums than the people in the regular pool. I suppose the high risk pool could have lower deductibles because most high risk people would be paying the deductibles, but then wouldn't low risk people try to get in the high risk pool because it had the same premiums and lower deductibles?

    1. In an invisible risk pool, if I understand right, no one "gets in" -- you enroll in an ordinary plan on the same terms as other enrollees, and the insurer earmarks you, ceding premiums that would pay for care above a certain threshold, and getting reimbursed for medical costs above that threshold.

    2. In a big risk pool, if it costs the government $X for all the people in the invisible risk pool, how is that different in practice from the government handing over $X to the insurer as a subsidy? When the insurer cedes the expensive people, it knows pretty well how much they were going to cost, doesn't it?

      All these baroque ways for the government to pay some health care costs for people in the private insurance market elide the simple truth: Health insurance costs a lot of money because a minority of people incur a lot of health care costs. To make costs for everyone go down 10%, the government can pay 10% of the costs in a simple way (give the insurance company the money) or in a complicated way (underwrite people, find the expensive ones, get subsidized for them). If all the enrollees are paying the same premiums, the simple way and the complicated way come out the same, except the simple way is simple and the complicated way is complicated.