Buckle Up! A Summary of Adverse Selection Pressures on Health Insurance Exchanges

Print Friendly
Share on LinkedIn0Tweet about this on Twitter0Share on Facebook0

Richard Evans / Scott Hinds / Ryan Baum


203.901.1631 /.1632 / .1627

https://twitter.com/images/resources/twitter-bird-blue-on-white.png revans@/shinds@/rbaum@ssrllc.com


May 16, 2013

Buckle Up! A Summary of Adverse Selection Pressures on Health Insurance Exchanges

  • For most households, the marginal costs of acquiring health coverage on an exchange are within shouting distance of the annual cost of a new car (Exhibit 1); and, the odds of health costs (if uninsured) exceeding costs of coverage (if insured) are less than 50 pct (Exhibit 3). Among subsidy-eligible households, in many cases net premium costs are higher for younger (and presumably healthier) than for older (and presumably sicker) beneficiaries with similar incomes (Exhibit 4). It follows that many households – especially younger and healthier – may choose not to purchase coverage
  • Assuming households that do purchase coverage do so optimally, +/- 80 pct of households would buy the cheapest (Bronze) coverage and the remainder (+/- 20 pct) the most expensive (Platinum) coverage. By purchasing the cheapest plan, healthy households (the 80 pct on Bronze) minimize their payment of excess premiums (premiums above their costs of care), and thus minimize the extent to which they subsidize households (the 20 pct on Platinum) whose medical costs exceed premiums paid
  • Small employers with younger and healthier employees can almost certainly save by self-funding, i.e. by avoiding the exchanges entirely. Stop-loss policies to self-funding employers can be priced to reflect a specific employer’s health risks, where fully-insured policies on the public exchanges cannot. This is likely to result in small group exchanges consisting of significantly worse-than-average health risks
  • We believe that the Affordable Care Act’s (ACA’s) provisions for limiting adverse selection are too weak (e.g. penalties for being uninsured are too low), or even counter-productive (higher effective premiums for younger than for older subsidy-eligible beneficiaries at a given level of income). The Act’s provisions for risk sharing are relatively strong; however these only serve to ensure equivalent relative exposure to unsustainably high absolute risks
  • Conclusion: Enrollment in individual and small group exchanges will be much less than originally (and perhaps even currently) expected; sellers of alternatives to full risk policies (ASO services, medical stop-loss insurance) in the small group markets will see accelerating demand; and, additional legislation and/or regulatory rule-making (i.e. further reforms) may be necessary soon after the exchanges begin operating
  • Cigna (CI) is a notable beneficiary of rising demand for ASO services and medical stop-loss in the small group market. Health Net (HNT) and Aetna (AET) are relatively exposed to small group risk, and stand to lose from a shift by employers (with healthier workers) to self-funding. AET’s acquisition of Coventry, completed last week, increases their exposure
  • Risks to other insurers have less to do with adverse selection (you can see it coming and price for it; and, the ACA ensures the risks are more or less equally shared), and more to do with the high likelihood that adverse selection forces new – and potentially adverse – legislation and/or rule-making

Individual Market: Why the healthy might stay away

For any given person or family, the marginal cost of purchasing minimum essential health coverage[1] under the Affordable Care Act (ACA) is the premium for that coverage, less both applicable subsidies and penalties[2]

Using Milliman data[3], we calculated the marginal cost of purchasing a Bronze (minimum essential) health plan for 4 types of households (single age 35, single age 55, family of four with a head of household age 35, and family of four with a head of household age 55) at income levels ranging from 150 to 1000 percent of the federal poverty level (FPL). For context, we compare the marginal cost of obtaining health coverage to the annual cost of leasing various new cars[4] (Exhibit 1)

For a healthy individual (or household) that sees limited odds of high health costs in an upcoming year, it seems reasonable to believe many would choose the new car (or something of comparable value) over a similarly priced insurance policy that they don’t expect to use

Many potential beneficiaries will not expect their health costs to exceed their premiums, in large part because of the tendency of health costs to be so incredibly skewed (Exhibit 2). To illustrate this further, we calculated the probability of any of our four benchmark household types having out-of-pocket health costs in excess of their out-of-pocket costs for acquiring Bronze coverage (Exhibit 3). With the exception of families of 4 with a 55 y.o. head of household, the likelihood of health costs exceeding the costs of acquiring coverage is below 50 pct for our remaining household types at all income levels above 200 FPL

Crucially, out-of-pocket premium costs for subsidy-eligible persons or families are likely to be higher for younger beneficiaries, and lower for older beneficiaries – further increasing the odds that younger (healthier) beneficiaries choose not to participate. This is because subsidy amounts are calculated as a function of two variables: beneficiaries’ income, and the premium a given beneficiary would pay for the second lowest cost Silver plan sold in that beneficiary’s market. The subsidy is equivalent to the applicable Silver premium (higher for the older beneficiary) less the out-of-pocket maximum (calculated as a percent of income and thus the same for both beneficiaries), thus the older beneficiary receives a larger subsidy (Exhibit 4)

Individual Market: Why the healthy may buy cheap coverage

Health costs are VERY unevenly distributed (a very large percentage of costs is paid by a very small percentage of persons, Exhibit 2, again). And, individuals and households are on average reasonably capable of estimating their next year’s health costs. It follows that most enrollees will expect relatively low health costs in a given enrollment year, and that a small number of enrollees will anticipate relatively large health costs. On the assumption that beneficiaries will only buy as much coverage as they think they are likely to need, using empirically-based estimates of beneficiaries’ health cost expectations, we modeled which plans beneficiaries are most likely to buy. Not surprisingly, beneficiaries with relatively modest health cost expectations (by far the majority) are most likely to choose the least expensive coverage option (Bronze); correspondingly beneficiaries with higher health cost expectations are likely to take the most generous option (Platinum). We estimate that +/- 80 pct of persons buy Bronze coverage, and +/- 20 pct buy Platinum coverage, for a population-weighted average actuarial value[5] (AV) of coverage of +/- 65. For comparison, the average AV of employer-sponsored health coverage in the current market is +/- 82

This tendency of most beneficiaries to self-select into the least expensive option is magnified by the relatively high marginal costs of buying more than the cheapest (i.e. Bronze) option. Subsidies will cover much (and in some cases, all) of the cost of Bronze coverage, but will not cover the incremental costs of buying more generous (Silver, Gold, Platinum) levels of coverage. In other words once a person or household has decided to enroll in coverage, the marginal cost of choosing added coverage above the most basic (Bronze) level is paid entirely by the beneficiary, with no subsidy or tax leverage. For reference, Exhibit 5 estimates the marginal out-of-pocket cost of buying Gold coverage (defined as out-of-pocket cost of Gold less the out-of-pocket cost of Bronze) for households in various income brackets, and compares these marginal costs to the costs of other common household expenses. Generally speaking, the marginal cost of buying Gold over Bronze coverage is on par with households’ typical spending for electricity, food away from home, or telephone service

Individual Market: Why the healthy who buy may do so at the last minute

The initial open enrollment period for the exchanges runs for a full six months – from October 1, 2013 to March 31, 2014. And, the penalty for not being insured in 2014 is much lower than in subsequent years (generally the greater of $95 or 1 percent of income). It stands to reason that many younger and healthier persons who anticipate limited health costs will wait until the end of the open enrollment period to enroll, since they can always buy (relatively) immediate coverage if the need unexpectedly arises. There are some limits on this approach: coverage purchased before the 15th of any given month is not effective until the 1st of the following month, and coverage purchased after the 15th of any given month is not effective until the 1st of the second following month

Small Group: Why smaller employers with good health risks may avoid the exchanges

As a consequence of small group market reforms (e.g. community rating, narrowing of premium bands, minimum essential coverage requirements, premium tax), we expect small group (100 or fewer employees) premiums to increase in 2014 by roughly 20 percent, before medical trend (+/- 6 pct)

Small employers can avoid these requirements by self-funding[6]. We recently showed[7] that ACA-related inflation of small group risk premiums effectively makes self-funding a roughly cost-neutral alternative for a typical small employer. More to the point of adverse selection, self-funding is made particularly attractive for small employers with better-than-average health risks. Stop-loss coverage for self-funding employers will still be priced according to the specific risks of each small employer, whereas fully-insured coverage will be community rated. Thus for smaller employers with better-than-average health risks (younger, healthier employees) the all-in costs of self-funding are likely to be less than the cost of participating in the small-group fully-insured market

Smaller employers with younger and/or healthier populations benefit not only from lower costs in any given coverage year, they also benefit from greater optionality with respect to subsequent coverage years. In the past, self-funded employers faced the prospect of being unable to find affordable stop-loss coverage in a year following larger-than-expected claims. Now, self-funded employers experiencing adverse claims in a given year can simply opt into the fully-insured market (thus benefitting from community rating) in the year(s) following outsized claims

Sizing the Individual and Small Group Markets

Likely candidates for the individual market include persons aged 30 to 64 who are currently either uninsured, or individually insured (roughly 22M persons as of 2010, 71 pct of whom are subsidy-eligible Exhibit 6). Persons under age 30 are eligible for catastrophic coverage which is much cheaper than Bronze coverage, and are thus excluded

Firms with fewer than 100 employees employed roughly 43M persons as of 2006; we estimate that 19.6M of these persons were covered by employer-sponsored insurance (ESI). Employers with 100 or fewer employees accounted for roughly 28 pct of covered workers, and 23 pct of ESI premiums paid (Exhibit 7)

Equal Sharing of Misery? Adverse Selection Provisions in the ACA

The ACA contains a series of provisions intended to prevent the occurrence or limit the magnitude of adverse selection risks (Exhibit 8a); and, a series of provisions intended to equally distribute the risks that do occur (Exhibit 8b)

Our assertion is that the provisions in Exhibit 8a are simply too weak. As we’ve shown, healthy beneficiaries are likely to either not enroll (coverage is costly in general; and costs more for, and is less likely to be used by, those who are younger), or to buy the lowest cost plan available. This reduces both the number of market participants paying premiums in excess of health costs, and the dollar magnitude of premiums in excess of health costs for healthy participants that do enroll

The provisions in Exhibit 8b serve to equalize sharing of risks that occur, but do not serve to limit the presence or magnitude of risks that exist in either the individual or small group markets. We acknowledge that these provisions are important and positive in the sense that they reduce the tendency of underwriters to compete on the basis of risk selection, though we think it’s crucial to recognize the provisions do nothing to affect absolute risk levels across underwriters in either market

  1. ‘Bronze’ coverage for persons 30 and over, or catastrophic coverage for persons under 30
  2. Since a penalty applies for not purchasing coverage, the savings associated with being uninsured are not the full OOP cost of coverage, but rather the OOP costs less any applicable penalties
  3. Houchens, Paul R., “Measuring the Strength of the Individual Mandate”, Milliman Research Report, March 2012
  4. 36 month lease, $500 security deposit amortized, licenses and fees amortized, residual value = 55 pct of capitalized value at lease initiation. All examples are new 2013 models in base trim
  5. Actuarial value (AV) equals the percentage of allowable claims cost paid by the plan
  6. Self-funded plans are regulated under ERISA, and are exempt from the ACA’s small group provisions
  7. “Why Smaller Employers Will Shift to Self-Funding: Who Wins and Loses”, SSR Health, April 29, 2013
Print Friendly