Last week, JAMA Health Forums published two short articles on reinsurance for the Affordable Care Act.
The first research letter by Polyakova, Bhatia and Bundorf concerned itself with the efficiency of the program design of the ACA reinsurance programs:
We examined 2 metrics for each program: (1) total government spending on reinsurance claims and average insurer liability (the sum of government spending and insurer liability is a constant), and (2) insurer risk exposure, measured by the coefficient of variation—the standard deviation divided by the mean of per-enrollee insurer liability…
Each program was associated with greater government spending, lower insurer liability, and greater insurer risk protection compared with no reinsurance (Figure). Large differences existed, however, across programs. State and federal reinsurance policies were far from the frontier, defined as a policy that achieves the maximum simulated risk protection for insurers for a given level of public funding.
Actuaries get ulcers when they think about variance. The more variability in a population, the more likely that insanely expensive events will occur and the harder it is to offload that catastrophic risk to another entity at a reasonable price. Polyakova et al define optimal reinsurance that eats all the tail risk of super high cost claims for a given level of expenditure. Insurers statistically know that there are several people who will have a multi-million dollar MRSA infection or significant 3rd degree burns in the next contract tear. These individuals are fundamentally unpredictable at an individual level, but there is a chance that any insurer could cover one or more of these folks. Insurers need to price that risk into their premiums. Reinsurance lowers premiums through both the immediate and direct payment of some claims from non-premium revenue and by reducing variance.
The big thing that Polyakova et al identify is that state policy makers were never optimizing program design for actuarial efficacy. Instead, they were seeking to lower gross premium levels to help make insurance more affordable for non-subsidized individuals.
Coleman Drake and I published a commentary on the need for reinsurance in the ACA/American Rescue Plan world.
The ARP’s enhanced premium subsidies undercut the logic of state-based reinsurance programs by capping benchmark marketplace premiums at 8.5% of household income for families with incomes above 400% of the federal poverty level.
Redesigned and more efficient state-based reinsurance programs may still be of use to policy makers under the ARP, as illustrated by Polyakova and colleagues.3 Risk adjustment is imperfect; there will always be idiosyncratically costly enrollees whose claims are unlikely to be captured by risk-adjustment algorithms (eg, enrollees with rare genetic diseases or those whose treatment costs are far above average). Even with good but imperfect risk adjustment, insurers may have a strong incentive to avoid geographic areas where such enrollees live.5 Reinsurance can reduce insurers’ risk of participating in markets with idiosyncratic, residual costs.6 Reinsurance can thus act as a complement to risk adjustment in that it helps reduce variance associated with catastrophic costs, whereas risk adjustment reduces risks associated with predictable costs.
We have long argued that states had three fundamental choices to improve consumer affordability in the US individual health insurance marketplaces.
- Direct subsidization
- Indirect subsidization via reinsurance
- Encouragement of parallel, underwritten markets
The ARP fundamentally chose to double down on direct subsidization by both increasing the richness of the subsidy and broadening the eligibility criteria for the subsidy. Indirect subsidization via state 1332 reinsurance waivers no longer makes sense as the federal ARP subsidies are far richer than any state’s reinsurance program.
Reinsurance can make sense as a corrective measure for idiosyncratic issues within a state’s insurance market. However, the policy challenge is that state lawmakers see reinsurance as a means of lowering gross premiums. That policy objective is now barely relevant for 2021 and 2022. Depending on the outcomes of the reconciliation bill, extended ARP-like subsidies may continue to render gross premium reduction redundant as a policy goal for several more years or permanently. Reinsurance as a state policy needs to be rethought to either be for smoothing of both actuarial risk (multi-million dollar burns) and likely to be poorly risk adjusted chronic conditions that can generate predictable massive claims where risk adjustment significantly underpays like inhibitor resistant hemophilia.
The challenge from a policy perspective is that the programs that are designed to smooth the market look very different than current 1332 reinsurance waivers. I think that the ARP has removed a significant pain point for upper-middle class families who had effectively mobilized for the Novocain of indirect subsidization from reinsurance by directly subsidizing these families. Through this pain removal, I think state policymakers will want to divert their attention to any number of other problems and we’re going to get zombie reinsurance programs that really don’t do much unless policymakers devote significant attention to reshaping these waivers.
Brad F
David
Helpful post.
A curious thing, however. Katie Keith in a HA blog this week had below to say on the Colorado reinsurance program. Even with ARP boost, as per outside Colorado analysis, their current reinsurance regime still optimal path:
As outlined in its application, Colorado did not propose changes to its current program except for the extension of time. Colorado did consult with an actuarial firm to assess the effects of the American Rescue Plan Act and reinsurance on its market and weighed alternatives in light of enhanced subsidies under the new law (such as using the reinsurance program to reduce premiums less or only applying reinsurance to certain metal tiers). But neither approach reduced premiums as much as the current program, so Colorado opted against changes.
Why?
Thanks
Brad
David Anderson
@Brad F: The question is which premium — gross or net…
Brad F
@David Anderson:
“The actuarial analysis identified insurer pricing
decisions as the primary cause of subsidized
consumers’ net rate increases.”
pg. 8
Reinsurnace Subsidized Enrollee Impact Study – March 2021.pdf – Google Drive
Stan Dorn
From my perspective, the chief advantage of waivers measures that lower gross premiums is to generate federal pass-through payments that states can use to further improve affordability — for example, by providing assistance to currently ineligible people (e.g., undocumented immigrants and people in the family glitch) or by funding supplemental CSRs. So long as federal pass-through payments exceed state contributions to a 1332 waiver, that goal can still be achieved.
But that is a quibble. The main point you make is 100% right. It will probably take time for state policymakers to internalize the huge change wrought by the ARP, especially if it continues beyond 2022. Reducing gross premiums no longer makes sense as a state policy priority, generally speaking.
-Stan Dorn, National Center for Coverage Innovation at Families USA, [email protected].
David Anderson
@Stan Dorn: I am in complete agreement with you. If the policy goal of reinsurance is to generate a pool of pass-through dollars to do other things that are good and useful, then that makes perfect sense. But if the goal is to reduce gross premiums for families earning 600+% FPL or 23 year olds in highly competitive markets, then does it really make sense as a public program?