Adverse Selection on the HIEs – It’s Not (Just) About the Federal Website

Print Friendly, PDF & Email

Richard Evans / Scott Hinds / Ryan Baum


203.901.1631 /.1632 / .1627 richard@ / hinds@ /


January 20, 2014

Adverse Selection on the HIEs – It’s Not (Just) About the Federal Website

  • Please see the SSR Health YouTube Channel for a detailed video walk-through of this research note
  • Through December 28, enrollment in states operating their own health insurance exchanges (HIEs) is roughly 2x (in terms of the percent of potential beneficiaries enrolled) the enrollment rate for states relying on the federally run HIE (a.k.a. Total enrollment thru December 28 was 2.2M persons; if enrollment rates in the federally managed states were as high as those in the states that manage their own exchanges, overall enrollment would be roughly 2.9M to 3.4M
  • Beyond reducing overall enrollment, there is some question as to whether the federal HIE’s problems will foster adverse selection – i.e. that less healthy potential enrollees will find it worthwhile to persist thru a difficult enrollment process, where healthier enrollees will not
  • We can’t see the health of enrollees, but can see their ages. Persons aged 0-34 represent 32.0% of enrollees on the state-based HIEs, as compared to 28.7% of enrollees on the federal HIE. Thus in absolute terms younger (and presumably healthier) enrollees are in fact less prevalent on the more difficult to navigate federal HIE
  • That said, comparing the age-mix of state and federal HIE enrollees is on par with comparing the rate at which two ships are sinking – neither appears to have enough young / healthy enrollees to avoid adverse selection, so the small difference in age-mix between the two exchange types doesn’t really matter
  • Across all of the HIE’s – state and federal – 18 – 34 y.o.’s are about 24% of enrollees (24.7% on the states; 23.3% on the federal) – far short of the 40% enrollment share we believe is necessary to avoid adverse selection
  • Barring an unlikely improvement in enrollee age-mix over the remainder of the open enrollment period (ends March 31st), we expect adverse selection in the first year of the HIE’s operation. This implies a high likelihood of further regulatory and/or legislative reforms in the relatively near term
  • We have no realistic means of forecasting specific changes to policy, but believe policy options are limited to the following: allowing greater premium differences based on age, allowing higher out-of-pocket maximums, reducing the scope of benefits contained within the definition of minimum essential coverage, increasing subsidies, merging these poor risks into a pool containing better risks; and/or increasing penalties

Health insurance exchange (HIE) enrollment levels through December 28 remain quite low. States operating their own exchanges enrolled 1.0M persons, equivalent to 7.4% of these states’ adjusted uninsured populations (Exhibit 1), and to 9.4% of these states non-group beneficiaries (many of whom presumably will now have to purchase coverage on the HIE, Exhibit 2)

Enrollment rates in states relying on the Federally-facilitated exchange are lower, presumably because the Federal enrollment website tends to be more difficult for beneficiaries to traverse. 1.2M persons enrolled in the Federal-exchange states, equivalent to 3.6% of these states’ uninsured (Exhibit 1), and to 5.9% of these states’ non-group beneficiaries (Exhibit 2)

Thus it remains clear that the Federal website’s problems have reduced enrollment levels to date. If enrollment in states relying on the Federal website was on par with enrollment in states managing their own exchanges, overall enrollment would be between 0.7M and 1.2M persons greater

Adverse selection is a key risk to the health of the HIEs, and the question has been raised whether the challenges faced by beneficiaries on the Federal site would amplify these risks. Specifically, the question is whether sicker enrollees are more likely than healthier enrollees to persist in enrolling when the enrollment process is so challenging

We can’t yet see the relative health of the persons enrolling, but we can see their ages. A close cousin of the tough enrollment process / adverse selection hypothesis holds that younger / healthier enrollees are less likely to persist through a difficult enrollment process than older / less healthy enrollees. If this were true, then the age mix of enrollees on the state and Federally-facilitated exchanges should be different – specifically the young should be under-represented on the federal exchange as compared to the state exchanges

Persons aged less than 35 constitute 32.0% (range 18.0% to 51.0%) of state-exchange enrollees, v. 28.7% (range 21.0% to 44.0%, by state) of federal-exchange enrollees (Exhibit 3). So in precise mathematical terms there are fewer young enrollees on the less functional federal exchange, though it’s not clear the difference will persist as the federal exchange improves

As near as we can tell, it’s not the small (3.3%) difference in 0 – 34 y.o. participation between state and federal exchanges that sets the stage for adverse selection, it’s the large extent to which age-mix is skewed to older beneficiaries across the combined (state + federal) base of enrollees

Last December we offered an analysis[1] of adverse selection risks on the HIEs, and as part of that analysis calculated the 18 – 64 age-mix we believed would have to be met for adverse selection to be avoided. Exhibit 4 compares what we believe is a ‘balanced’ age-mix that avoids adverse selection (grey columns) to the current age-mix of HIE enrollees (green columns) – plainly the age-mix of the current enrollees skews much older than our estimate of ‘balanced’ age-mix

To be fair, we can’t really know if the current enrollee mix will result in adverse selection without also knowing the relative health status of these enrollees – i.e. estimating adverse selection risks with these limited variables is an incredibly noisy process. That said, we would tend to think that if anything, these early adopters are less healthy than the average potential beneficiary, which of course feeds our expectation of adverse selection

As a final argument that the current skew in age-mix has relatively little to do with problems on the federal exchange, we would point out that the relative odds of enrolling by age correspond fairly well with what we know about the relative risk aversion of potential enrollees by age (Exhibit 5). A note of caution – the y-axes in Exhibits 4 and 5 are measuring very different things. In Exhibit 4, we’re measuring the percent of total enrollees by age category (i.e. “age-mix”); in Exhibit 5 we’re measuring the percent of potential enrollees in a given age group who actually enroll. Across the nation as a whole, the young are enrolling at substantially lower rates than the old – only 2% of potential enrollees[2] aged 18-25 have enrolled, as compared to more than 7% of potential enrollees aged 55-64 (green columns, Exhibit 5). And, these relative odds of enrolling by age category correspond fairly well to the relative levels of risk aversion[3] shown by each age category. This suggests the odds of enrolling by age group are a function of relative risk aversion, rather than any difference in willingness to confront the difficulties of enrollment

  1. “Why Adverse Selection is Likely on the HIEs”; SSR Health LLC; December 15, 2013
  2. Uninsured less newly eligible Medicaid beneficiaries, plus 2013 non-group beneficiaries
  3. As measured by survey responses in the context of the Medical Expenditure Panel Survey (MEPS)
Print Friendly, PDF & Email