The ACA had three risk stabilization measures from 2014-2016. Two were designed to be temporary and one is permanent.
The first, temporary measure was a three year risk corridor. It was intended to claw back excessive profits and pay for some excessive losses of insurers. Midway through implementation, it became a revenue neutral requirement and now the insurers are litigating the payment of multiple billions of dollars before the Supreme Court. The second was three years of federally funded reinsurance. The final stabilization mechanism was permanent, revenue neutral, risk adjustment.
Why does the ACA have three mechanisms that fundamentally poured money into the claims payment pools from sources other than any given insurer’s premium revenue? Why three instead of one?
The three mechanisms are trying to solve different problems even if the solution is fundamentally moving money around.
Risk corridors provide protection for insurers and the government against fundamental actuarial uncertainty. In 2013 when the 2014 rates were being set, no one had a clue what the ACA individual market would look like. Would there be 30,000,000 people crushing the first sign-up period with half the pool being under the age of 35 and healthy as a horse? Would the market be fundamentally adversely selected and be wicked heavy on cancer patients and diabetics? Depending on where any insurer fell on that question, the rates would vary dramatically differently. Risk corridors were supposed to provide protection against insurers being too optimistic and setting rates too low to cover some of the excess losses. Risk Corridors were also set up to provide the government some financial protection in case all of the insurers were systemically too optimistic and set rates way too high for the actual claims incurred. As reality turned out, insurers were systemically too optimistic on their projections about the health of the population.
Reinsurance is fundamentally protection against catastrophic claims costs. The ACA model of reinsurance from 2014-2016 would eat a big chunk of large but not massive and rare claims. The set up was that CMS would pay the entire portion of claims that fell between $45,000 and $250,000 per year in 2014. This is a large portion of the high cost claims but it did not do much to help an insurer that had individuals with multi-million dollar claims. The new catastrophic risk reinsurance program run since 2018 has CMS cover 60% of any claim year for an individaul that runs over a million dollars. That program design is specifically aimed at extreme tail risk while the original reinsurance program was designed to eat up typical right hand risk. The state 1332 waivers that provide reinsurance also eat right hand risk.
Finally, risk adjustment in the ACA moves money between insurers that cover a population that is coded as less chronically ill than average to insurers that cover a population more chronically ill than average. These flows can be significant. Some insurers will send more than a fifth of their premiums out the door for risk adjustment. The objective of risk adjustment is to make insurers fundamentally risk agnostic so that there is no race to the bottom in marketing, network design, benefit design and hassle based underwriting to avoid high risk/high cost populations even as insurers are obligated to offer guaranteed issued, community rated policies to anyone who signs up.
These three programs solve different problems in any guaranteed issue insurance market.
In a COVID-19 world, I think we need to apply different tools to different problems.
For the 2020 policy year, the shape and size of the claims curve is going to be different. A lot of claims that would normally have occurred are not occuring. One insurance nerd I spoke with recently said that they have not seen a knee replacement claim in three weeks when they typically pay for a dozen per day. At the same time some insurers are getting slammed with high cost, high acuity COVID hospitalization claims. Here reinsurance for high cost claims to eat unanticipated high cost claims risk may be a reasonable response. Sherry Glied and Kathryn Schwartz proposed this in a recent Health Affairs blog:
The Federal government would promise to reimburse payers at the minimum of Medicare or negotiated rates (in the rare case that negotiated rates were below Medicare) for all COVID-19 claims through the reinsurance system. Because all reinsurance payments would be made at Medicare rates and negotiated rates are typically much higher, this system would provide partial, but not complete, protection to private insurers and self-insured employers.
This type of proposal would eat a significant amount of unanticipated right hand risk.
2021 has a different problem. 2021 premiums are being priced in an informational void. Actuaries can not look at 2019 experience and readily assume 2021 will look a lot like that year with predictable adjustments. We have numerous known unknowns that won’t be well bounded by the time final rates are set. In the ACA market, the most optimistic set of projections of the known unknowns will have the lowest premium in a market and get most of the price sensitive enrollees. If we assume that the winner’s curse is in effect on what is effectively a first price quasi-blind auction with a single rebid round, we can assume that the premium level has a good chance of being too low. From a policy point of view, we really don’t want insurers to go broke mid-policy tear in 2021. In this case, the appropriate tool to use might be a risk corridor where the federal government takes on significant losses once the insurer has paid out 100% of premiums in claims. Most insurers can survive a year of horrendous mispricing using reserves. If insurers systemically overprice, the Medical Loss Ratio (MLR) requirements already act as a weird de facto one way risk corridor with a weird pay-out mechanism.