Actuarial Value: a Healthcare Reform Backgrounder
Actuarial value is frequently defined as the percentage of medical expenses paid by the insurer , . Back in 2009, the Congressional Research Service estimated actuarial values for a variety of plans; notably Medicare (including a Part D plan) was around 76%, and the Federal Employee Plan (FEP) was around 87%. Generally, actuarial values are determined for a standard population, meaning that is is generally difficult to directly compare a retiree population like Medicare to an employed population like FEP. They may or may not be adjusted for utilization differences caused by the benefit itself.
On the Exchanges, there are going to be at least four tiers of benefits, the so-called “precious metal” plans, defined as plans with 60%-70%-80%-90% actuarial values for Bronze-Silver-Gold-Platinum plans. Applying the definition of “actuarial value” literally, this means that one would expect Bronze plans to cover 60% of the medical expenses for those on that plan. This is only true in the abstract, because the population that enrolls in the 60% plan is not expected to be standard. Specifically, it is expected that sicker people will enroll in Gold and Platinum plans (unless they are very low income, in which case their Silver plan is enriched to have an actuarial value as high as 94%). Healthier people will enroll in the bronze plan. Therefore, when we see the plan-by-plan data for 2014, actuaries would not be surprised to see that the 60% Bronze plan resulted in only 50% of charges paid by the plan, on average, while the Gold plan may have 85% of charge paid, on average. Further, the average medical expense paid by either the insured or insurer is likely to be much lower than on the Gold plan. These effects are sometimes referred to generally as “adverse selection”.
This sort of counter-intuitive result is why it is important for as many people as possible to understand the underlying mechanics of actuarial value calculations, discussed below. Further, it is likely that these calculations will be dictated by either federal or state government, meaning that a sophisticated understanding of actuarial values is important for the political classes.
For simplicity, assume a standard population of just two people, Abby and Bob. Both go to the doctor exactly once a year, at a cost of $100. One of the two will go to the hospital, which charges $9,800. The average medical expense of this population is expected to be $5,000. If the plan has no deductible and no cost-sharing, then the plan will need to charge $5,000 in premiums to cover the expected cost. This plan would have a 100% actuarial value.
What’s the actuarial value of a plan than has a $1,000 deductible and no other cost sharing? The expected cost for the person who goes to the hospital is $9,900, with $8,900 paid by the insurer and $1,000 paid by the person as part of the deductible. The expected cost for the other person is $0 to the plan, and $100 to the insured. This means that the plan will need to charge a premium of $4,450 ($8,900 / 2) to cover claim costs. The actuarial value is $4,450 / $5,000, or 89%.
Actuarial models usually include a utilization component. Whether utilization effects will be allowed when evaluating Exchange plans is unclear. There is usually no way to verify what the exact utilization impact of a particular plan design is, making it difficult to regulate the usage of these factors.
Here’s how a utilization factor may impact the Abby/Bob example. If there is no deductible, actuarial models may suggest that rather than both Abby and Bob going to the doctor once, they may end up going to the doctor twice. This would increase the total cost to $10,200, with an average cost of $5,100. The no-deductible plan still has a 100% actuarial value. In contrast, the $1,000 deductible plan still has an expected cost of $4,450. This means that in a utilization-adjustment world, the “slimmer” $1,000 deductible plan has a lower actuarial value of 87.25% ($4,450/$5,100).
As mentioned above the fold, the choice of the underlying population is very important. As a general rule, the older and/or sicker the population, the higher the actuarial value. Again, whether the models used to score Exchange plans will be allowed to adjust for this is unclear. If the “precious metal” plan actuarial values are calculated only on a standard population, then the lower actuarial value plans, like Bronze, will be overstated in reality. When this occurs, the relative premiums between metal plans will match their relative actuarial values, but all premiums will need to rise in tandem to compensate for the impact of adverse selection (proof of this is beyond the scope of this post).
In general, the more types of claims that are covered by the plan, the higher the actuarial value will be. This phenomenon results in a counter-intuitive effect: the larger and more inclusive the Exchange’s Essential Health Benefits are, the higher the deductibles and out-of-pocket payments are likely to be. For example, consider what will happen if the Essential Health Benefits increase the quantity of preventive services paid at 100% by including, for example, all versions of contraceptive devices and drugs without any cost-sharing. In this case, the deductibles and cost-sharing on the remaining services will need to go up to keep the Bronze actuarial value at 60%.