• Adverse selection at every turn

    Benjamin Handel’s paper “Adverse Selection and Inertia in Health Insurance Markets: When Nudging Hurts” (forthcoming in the American Economic Review) is fascinating. It’s about two things: (1) “inertia” or status quo bias in selection of health plans and (2) adverse selection that arises due to efforts to overcome that inertia, e.g., helping consumers better understand the cost and benefits of different benefit packages and cost sharing features.

    This paper highlights (i) that inertia can be substantial in health insurance markets and (ii) that efforts to improve choices in health insurance markets with consumer inertia should take into account the potential impact on incremental adverse selection. We cleanly identify the extent of inertia by leveraging a panel data set where all consumers must make an active plan choice from a new menu of health plans in one specific year but must deal with inertia in other years. Several model-free preliminary analyses reveal that inertia has a substantial impact on health plan enrollment as the choice environment evolves over time. We estimate a choice model of consumer decision-making under uncertainty to quantify inertia, ex ante health risk distributions, and risk preferences that yields clear evidence of large and heterogeneous inertia. We use this model to study the impact of counterfactual policies that reduce inertia, without differentiating between specific policies. While reducing inertia increases welfare in the naive setting where health plan prices are held fixed, in the setting where health plan premiums adjust as enrollees switch plans reduced inertia leads to incremental adverse selection and a welfare loss. When inertia is reduced by three-quarters, this incremental welfare loss effectively doubles the welfare loss from adverse selection in the observed environment (8.2% of consumer premiums). Though these results are specific to our setting, they illustrate that the interaction between inertia and adverse selection can be quite important, and that policies to improve consumer choices in health insurance markets should consider the potential for incremental risk- based plan selection. [Bold added.]

    In English: The difficulty consumers face in appreciating the value of plans and the resistance they exhibit in plan switching has a risk pooling effect. Think of it this way, if everyone could pick the plan that minimizes their cost, more sicker individuals would end up in generous plans with higher premiums and healthier individuals in less generous plans with lower premiums. This is adverse (favorable) selection into the more (less) generous plans. That consumers don’t perfectly identify their costs under each plan and have status quo bias (inertia) undoes some of this selection.

    That’s not to say you want to keep people in the dark just for the risk pooling effect. It’s just to point out that there a selection issue that arises as people are better informed and, perhaps, more willing to switch plans on that basis. To the extent that selection can be counteracted through risk-adjustment of subsidies or other mechanisms, it’s not a big deal. But one at least needs to know about the issue to address it. And, it’s a, perhaps surprising, reason why the efforts to inform consumers about health plan options (more here [PDF]) and provide plan selection guidance under the ACA could exacerbate risk selection.


    • I’ll state the obvious: why not choose a bronze plan when not sick and then a platinum plan when sick. The explosion in chronic illnesses (illnesses that were at one time death sentences but with advancements in medicine are now manageable and treatable) is the main source for the explosion in health care costs. Will all those individuals with chronic illnesses end up in platinum plans? [To be clear, I include old age as a chronic illness – there’s no cure for it. Seniors and non-seniors with chronic illnesses are essentially in the same predicament. We’ve chosen to manage the health care risk for seniors through public finance (Medicare). In the past, we’ve chosen to manage the health care risk for non-seniors with chronic illnesses first by imposing the risk on the individuals and private insurance and second on public finance (Medicaid) and the hospitals that must treat them. Beginning in 2014, the goal for managing the risk for non-seniors with chronic illnesses is a combination of private insurance and public subsidies. I suspect that what will become clear over time is that the risk for seniors and the risk for non-seniors with chronic illnesses are much the same and that, therefore, it would be appropriate to manage those risks in the same way. Will the choice for managing that risk be public finance (Medicare) or a combination of private insurance and public subsidies (Obamacare). Seniors who express concern about what will happen to their Medicare under Obamacare may not be able to articulate their concern (it’s not death panels, it’s managing risk), but that’s it.]

      • People with chronic illnesses would tend to gravitate to the bronze plan, and those with low expenses, would tend to gravitate to the bronze plan.
        Fortunately, only about 2.5% of the policyholders are estimated be part of the Exchanges.
        We do not need participants selecting against the insurers, and we do not need insurers selecting against the participants.
        We need insurers and particupants to work together in forming long-term relationships, not the constant churning we now see. Insurers typically turn over their clientele every 3-5 years.
        Can you imagine how expensive it is to create a whole new customer base every 3 to 5 years?
        Don Levit

    • Perhaps it’s just me, though I suspect not. When I read the English version of what the research shows, I get this: there’s a set of choices consumers would make if they were ideally informed and rational. And then there’s what they actually do, which differs from this partly because of less than ideal information, and resistance to making otherwise-rational changes.

      That’s where I start to follow less well. I *think* part of the point is that if we all selected our plans based on perfect information/rationality this would have unfortunate side-effects. But for those of your readers who aren’t health-policy wonks, it may not be clear that this is the correct reading, and if it is, it may not be clear just what the downside is. Is it that the well would not be sufficiently subsidizing the less-well? Or is it something else? Or am I just confused?

      • The problematic side-effect is that plans that draw a sicker risk pool cost more (higher premiums). This discourages the lesser sick among this sicker risk pool from enrolling, making the risk pool sicker still. If this continues, an “unraveling” or “death spiral” can occur whereby the plan collapses. This is a market failure in the sense of rendering the market less complete than it was. A plan that some people wanted could not be sustained.

        That being said, the solution isn’t to keep people in the dark so they make bad choices. The solution is to address risk selection through various means of risk adjustment.

        • Austin:
          I agree with you that risk adjustment is necessary to maintain a viable pool for the long run.
          In my vision of a health plan, everyone pays fully community-rated premiums. That way, no one has a financial advantage, art inception.
          If we have a family of four, the total premium, initially, is about $16,000 a year.
          What we have discovered through working with our actuary, Milliman, is that within this family of 4, we can actually provide risk adjustment, so that the healthier members (usually 3 of the 4) are paying lower premiums, and the one with the chronic illness continues to pay fully community-rated premiums, the highest of the 4.
          Due to the lower premiums of the low claimants within the family, the total family premium reduces significantly by year 3 – on a magnitude of 60-80%!
          Don Levit

    • If you’re really sick, you should choose a bronze. Same out of pocket maximum as platinum, but lower premiums

    • There’s no avoiding the realization that subscribing to insurance, qua insurance, is irrational behavior, since the return is below, often far below, 80% on the premium dollar. Insurance makes sense, one supposes, for those who are so seriously risk-averse that they don’t mind a huge financial hit.

      There are many folks who are not risk-averse–hitchhikers, mountain climbers, ocean sailors and sky-divers, etc. They would be utter fools to take out insurance that covered prostate exams and colonoscopies, of course, but not their dangerous activities. The irony is that neither Medicare nor Obamacare will cover them while hitchhiking in Chile, mountain climbing in Nepal or sailing around the world, though they are still liable for the premiums!

      Nowadays, of course, insurance is not insurance. Obamacare, for example, is just another means to transfer wealth. To the extent that it is not insurance, it makes sense for the old and sick, even those who are seriously risk averse, to participate, since the losses they would sustain with real insurance are compensated by the wealth stolen from the young and healthy.

      • @Jimbino

        1) Hedges always cost money and just because a bet is a financial win (or lose) doesn’t mean it is worth (or not worth) doing.

        2) travel health insurance is usually cheap because you aren’t going to be there long and the people doing it tend to be healthy. Tour operators in Russia, where many people don’t have coverage for overseas, require health insurance as part of the package. It costs something like $10/person for a week so you can’t decline, even if you have other coverage.

        3) I doubt that people are still going to be liable for Obamacare while getting overseas insurance. I live overseas and have expat insurance. Believe me, this would be a big issue in the America expat communities across the world because health insurance from American insurance companies for Americans living abroad is half the price of health insurance from the same company for the same Americans living in America. A worldwide policy excluding the US costs half as much again.

    • I’m a bit troubled by what I read in the article summary and comments. As best as I can tell from the summary, the original analysis is abstract and formalized, applicable perhaps only to the no longer in existence market for unsubsidized individual policies, or to some idiosyncratic private employer setting. In the real world there are three Federal exchanges that present many insurance choices–the FEHBP, Medicare Advantage, and Medicare Part B. They all have three properties of high importance: (1) every enrollee gets a massive subsidy, typically two thirds or more the premium cost of whatever plan is chosen (these formulas differ widely among the programs, however), (2) everyone is given the opportunity to use “expected cost” information to improve choices, and (3) even when given the opportunity and information to choose plans that are better buys, most enrollees ignore both that information and the existence of Open Season choices. There is massive aversion to comparing plans and massive inertia in enrollment. The welfare loss from bad plan choices can be huge–thousands of dollars a year–far more than the 8% cost of adverse selection found in the forthcoming article.

      In the case of the FEHBP, my publisher and I have been providing information on the expected cost of all plans (premiums plus OOP) for over 30 years in “CHECKBOOK’s Guide to Health Plans for Federal Employees.” Recent versions are online as well as in print and it takes only a minute or two for the user to see available plans ranked by an actuarial estimate of likely dollar cost (least costly first). Medicare has similar comparison tools for its enrollees.

      There are real world behaviors in these programs, such as the large (but not as large as purely rational choices) switch from Medicare plus Medigap to Medicare Advantage plans, and from Blue Cross Standard to Blue Cross Basic in the FEHBP, over the last half-dozen years. There is risk selection in these programs from plan choices, but that does not mean that there is welfare loss compared to the alternative. Moreover, there are methods to minimize risk selection, used intensively in the Medicare programs but ignored in the FEHBP.

      Using well-designed cost comparison tools overcomes virtually all of the cost comparison complexities that are impossible for ordinary consumers to calculate (we compare plans taking into account differences in deductibles, copays, coinsurance, and MOOP limits, along with information on costs by family size and age, for example). That said, very few enrollees in these three programs either know of or use these tools. Also, other complexities and tradeoffs remain, such as “which networks include my doctor” (a question CHECKBOOK answers for you on the summary results page, if you use a DC metro area zip code in your search). There are also big issues in designing the optimum premium subsidy amount and design (Flat amount? Capped at what level? Percentage amount?) to maximize rational choices based on marginal costs, as well as in designing contribution formulas that minimize risk selection. The crucial point I want to emphasize is that we need research and analysis focused on real world programs and potential changes in these, far more than abstract models.

      The tax subsidies in the ACA exchanges create huge and often irrational incentives and disincentives (like a marriage penalty that can reach $10,000 a year) and would be another useful target for real world analysis. So while I look forward to reading this article, what is most needed is future research that is targeted on real world programs and problems and that identifies ways to improve program designs. (A good starting point in rational design is laid out in Alain Enthoven’s vintage book “Health Plan,” a proposal for something like the ACA as it should have been designed, but was not.)

      p.s. It is not true as one commenter suggested that the exchange Bronze plans will typically be the best choice for enrollees expecting high costs. The best choice will depend both on OOP and on premiums net of subsidies, with due regard to “applicable percentage” limits on premiums based on income, plus enhanced benefit formulas. The interactions of these formulas are VERY complicated. But one of their highly predictable effects, coupled with the modified community rating used in the exchanges for the young to subsidize the older and sicker, will be to encourage older and sicker persons to enroll, and younger and healthier persons not to enroll, an adverse selection problem of potentially crippling consequences for the exchanges. Had more rational formulas been designed, along the lines of those used in Medicare Part D, these and other ill effects could have been avoided.