Humana recently announced that next year it is withdrawing from 88% of the counties where it sold Affordable Care Act (ACA) exchange plans this year. United Healthcare forecasts higher earnings in 2017, stemming in part from its decision to shut down most of its exchange business. Aetna has cancelled plans to expand its ACA market footprint and is instead reevaluating its current participation. At least four states, Alaska, Alabama, Oklahoma and Wyoming will likely have only one exchange insurer this coming year. Sixteen of the 23 co-ops initiated with ACA funding have collapsed. And researchers supportive of the ACA estimate that insurers are requesting average gross premium increases of 23% next year These data points suggest the ACA’s individual market changes are faring poorly thus far.

In an attempt to understand the ACA’s initial effects, we published two working papers—one in April and one in June. In our April paper, we found that net reinsurance subsidies amounted to additional revenues equal to about 20% of premium revenue for the 289 individual Qualified Health Plans (QHPs)—ACA-compliant plans that CMS determined to be the same, or substantially the same, as those certified to be sold on exchanges. Despite that, insurers still suffered substantial losses in the aggregate. Reinsurance subsidies are payments to insurers to compensate them for much of the expense of their highest cost enrollees.

If insurers had also received a taxpayer bailout through the ACA’s risk corridor program, they would have booked $8.6 billion in corporate subsidies between the two programs or additional funding equal to 24% of premium revenue – and still lost money on their individual QHPs in the aggregate. We concluded that the larger losses incurred by insurers on these plans after 2015 combined with the expiration of the reinsurance program would cause premiums to significantly increase and that ACA exchange plans were most strongly attracting lower-income people who receive large subsidies and people who expected to file big medical claims.

A new study published by the Commonwealth Fund authored by Wake Forest law professor Mark Hall and Virginia Commonwealth professor Michael J. McCue also looks at 2014 data. They find that median losses for insurers that sold individual policies were 4.4% of premiums, and that there was substantial variation among plans. Comparing the performance of 144 insurers that sold individual policies in both 2013 and 2014, they found that about a third were profitable and nearly two-thirds lost money in 2014.After acknowledging insurers’ initial difficulties adjusting to the ACA, they argue that, after premiums are increased to reflect the more expensive than expected risk pool, better days are ahead for the ACA. However, Hall and McCue believe that insurers will probably need billions more in corporate subsidies, likely paid for through a regressive tax on other people’s health insurance, until the market “has matured.”

Although there are certainly differences in tone and inferences drawn from Commonwealth’s report and the ones that we authored for the Mercatus Center, there are important points of agreement, including:

•  The Obama administration increased reinsurance subsidies to insurers beyond what insurers expected when they set premiums, and this increase was critical in reducing insurers’ otherwise sizeable losses.

Commonwealth: “Ultimately, the reinsurance credits to insurers (net of fees that insurers paid) were almost 50 percent higher than insurers had originally estimated ($43 vs. $29 [per member per month]). These greater payments resulted in part because enrollees in ACA-compliant plans had more high-cost claims than first anticipated, but also because the federal government modified the reinsurance payment formula in mid-2014 to be more favorable to insurers.”

Mercatus: “As a result of HHS changes to the formula, the reinsurance program was made about 40 percent more generous on a per-enrollee basis than insurers expected when setting 2014 premiums.”

•  For a significant number of insurers, enrollees in 2014 were far more expensive than projected.

Commonwealth: “The quartile of insurers with the highest claims (75th percentile) underestimated their claims by an average of 35 percent, whereas the lowest-claim quartile projected their claims much more accurately, within 4 percent, on average, similar to the average claims underestimate of 6 percent marketwide.”

Mercatus: “Per-enrollee premium income for individual QHPs was 45 percent higher than for individual non-QHPs ($4,519 versus $3,126). However, the much larger per-enrollee premium income on individual QHPs was still not enough to cover medical claims that were 93 percent higher for their individual QHPs ($4,973 versus $2,581)…. The substantially higher average premiums were still not high enough to cover the cost of insuring enrollees overall, as too few relatively healthy people were part of the individual QHP risk pool in 2014.”

The differences between the Commonwealth study and the ones we authored for Mercatus appear larger-than-actual because Hall and McCue redefined a conventional term (loss ratio). The loss ratio is a common measure of insurer performance and is conventionally defined as the ratio of claims to premium revenue. For instance, an insurer that paid out $800,000 in claims and collected $1 million in premiums, for example, would have a loss ratio of 80%. A loss ratio around 80-85% is generally considered an indicator of a financially sustainable individual market insurance plan.

In Commonwealth’s study the term “loss ratio” is redefined to mean claims minus reinsurance receipts. Thus, by subtracting the approximately 20% of actual claims that were reimbursed by reinsurance payments, Hall and McCue computed an adjusted loss ratio of 89.4%. From the perspective of how insurers performed that ratio is not good, but it is not terrible either. In our studies, we used the traditional loss ratio definition and found that the average insurer loss ratio was 110% in the individual QHP market in 2014—a very bad number for any health insurance plan.

Which metric is better? It depends. If one wants to calculate insurers total performance for a line of business – that is, to use “loss ratio” as a rough proxy for financial profitability – the Commonwealth statistic has much to be said for it. However, if one wants to consider the sustainability of the ACA, our use of the traditional definition has the advantage since the substantial reinsurance subsidy ends this year. Thus, if one wants to see how insurers are likely to perform under the ACA once the training wheels come off, seeing how they would have done in 2014 without training wheels is highly relevant. Moreover, our Mercatus reports provided data from which both a standard loss ratio and Commonwealth’s adjusted loss ratio could be derived.

Ultimately, the Commonwealth study’s methodology and its rhetoric that insurance performance was not so bad cannot mask the import of the underlying facts it presents. Moreover, the Commonwealth study seems to neglect that insurers’ losses selling ACA plans likely doubled between 2014 and 2015. Since many insurers are already hemorrhaging cash, even with huge back-end support through the reinsurance program, it is not surprising for the law’s supporters to propose extending that program.

We are skeptical about extending reinsurance to address the ever-more apparent woes of the ACA. First, reinsurance is paid for by what amounts to a regressive tax on, among others, lower-income employees who struggle to afford health insurance through their jobs. The tax is hidden because it is levied on the plans or administrators but there is no doubt that it is passed on to workers. Second, reinsurance subsidies end up disproportionately helping those who can afford policies with lower cost sharing. As both the government’s Risk Adjustment program and Actuarial Value Calculator confirm, people who purchase policies with lower cost sharing tend to have higher medical expenses and are less restrained in their use of medical services. These are often the wealthy. So, not only does reinsurance tax the poor disproportionately, it sometimes provides them fewer benefits. Third, government reinsurance subsidies distort the proper pricing of coverage and give the artificial appearance of cheaper coverage. Fourth, extended reinsurance coupled with other heightened subsidy mechanisms now being proposed are likely to increase the costs of the law.

Helping people purchase insurance through hidden corporate welfare is an idea of which people across a broad spectrum of political ideologies should be wary. Given the dysfunction in the ACA individual market, policymakers should instead seek to understand what’s gone wrong and fix those problems before spending billions propping up a program that appears to be failing. That requires a full understanding of all of the factors in play, precisely what our studies for the Mercatus Center have sought to provide.

Acknowledgement

We thank the authors of the Commonwealth study, Mark Hall and Michael McCue, for retracting earlier mischaracterizations of our work. The first version of its report and a related blog entry mistakenly stated that we had deliberately excluded reinsurance and other government programs from consideration in its report. At our request and our presentation of evidence showing that we had repeatedly considered and extensively presented this data, Commonwealth issued a correction. This correction  facilitates debate of health care finance issues in good faith based on a common set of facts and understandings.

The views expressed herein are in my own and do not necessarily reflect those of the University of Houston.