Massachusetts Individual Mandate Gaming, Continued
Apparently there are no studies that report the selection effects of gaming the Massachusetts individual mandate. This is according to Amy Lischko, Tufts University Assistant Professor of Public Health and Family Medicine and former Commissioner of Health Care Finance and Policy and Director of Health Care Policy under governor Mitt Romney. Lischko is also an investigator on a project studying risk selection in state health insurance programs.
She wrote in an e-mail that the Boston Globe’s investigation has produced all that is known on gaming the mandate. “The Division of Insurance is investigating this as well but they have not released results yet. There hasn’t been any academic research on this issue,” she wrote.
Also, in my prior post on Kay Lazar’s reporting in the Globe I didn’t note that it implicitly answered some of my earlier questions. Namely,
- Does Massachusetts have an open enrollment period? No.
- Does an exclusion period for pre-existing conditions exist in Massachusetts? No, though I thought the answer was yes. But Lazar wrote in the Globe, “[W]aiting periods for coverage … was effectively disallowed by the 2006 law … [Governor Patrick's proposal] would also bring back the rule allowing insurers to exclude coverage for preexisting conditions for six months …”
So, there doesn’t seem to be much to prevent gaming in Massachusetts. And yet the problem of gaming has not yet been established as a significant one. Why is there so much uncertainty about this? I suspect that access to data has been difficult. More should be known about Massachusetts’ health reform, especially considering we’re going to implement something like it nationally.
To channel Kevin Drum, “Get cracking, scientists.” (Of course, I’m one of them. Though I’m not funded to study anything about Massachusetts.)
Gaming the Individual Mandate in Massachusetts
Today Kay Lazar reported in the Boston Globe that thousands of Massachusetts residents are purchasing health insurance only when it is needed and dropping it and paying the relatively low penalty when it is not.
In 2009 alone, 936 people signed up for coverage with Blue Cross and Blue Shield of Massachusetts for three months or less and … [paid a monthly premium of] $400, but [had] average claims [that] exceeded $2,200 per month. …
Governor Deval Patrick recently filed legislation that state regulators believe will help fix the problem, by restricting insurance enrollment to twice a year for people who buy on the open market and allowing waiting periods before coverage kicks in. …
It would also bring back the rule allowing insurers to exclude coverage for preexisting conditions for six months, or impose a similar waiting period under certain conditions for people buying coverage on their own. …
Consumer advocates said they aren’t convinced that a lot of people are gaming the system, and they said that many of the individuals buying on the open market are likely those who are between jobs, new to the state, or have some other legitimate reason to buy coverage for a short period.
Implications for the nation’s individual mandate are obvious. But there is plenty of time to learn from Massachusetts’ experience and tweak the federal law, as there is to do more thorough analysis of the issue.
I have no doubt there is some gaming going on, as there would be with any mandate. What has not yet been shown by anybody is the total size and significance of it. That is, how much higher are premiums due to the selection experienced by insurers? Is this a mountain or a mole hill? Some of the reporting is based on sources that have a reason to inflate the significance of this problem. And it can’t be all that big when about 97% of the state’s population is covered (though it depends how that figure is measured).
All in all, I haven’t seen enough work on this that suggests there is a sound basis for policy. I hope somebody, somewhere can get their hands on some solid data and do some credible analysis.
Later: I have many other posts on this topic.
This Title Adequately Characterizes the Post
Sometimes readers get confused by the title of a post (or newspaper article, book, etc.). How accurate and complete can a brief statement be? Not very. Sometimes one has to read the content of the piece and apply a little thinking to assess the title’s scope. We know this, right?
Case in point: I titled a prior post “Individual Mandate Penalties Are Adequate.” Adequate for what? And in what sense? Adequate to rid the world of disease and famine? Adequate to confuse or annoy people? Of course the body of the post is crystal clear of the scope of adequacy, the means by which the assessment was made, and the limitations of analysis applied. As Reihan Salam quoted my conclusion,
Since there is little evidence of substantial gaming in Massachusetts, based on an analysis of penalty size alone there would seem to be little cause for concern over gaming under ACA, particularly for higher income individuals. This analysis ignores other differences between Massachusetts and its health reform law and the national population and ACA, respectively. Results are sensitive to assumptions, but I deliberately selected those conservatively as indicated above.
Salam also correctly points out that Massachusetts permits insurers to apply a six month pre-existing condition exclusion period. That is, they must guarantee issue of a policy but do not have to cover a pre-existing condition in the first six months. As far as I know, no such exclusion period is permitted under ACA rules. So, this is a difference between the Massachusetts individual mandate and the ACA’s mandate. I pointed out this difference in a prior post to which I linked in the “Individual Mandate Penalties Are Adequate” post.
For all that, Salam concludes, “Frakt titles his post, ‘Individual Mandate Penalties Are Adequate.’ I’m struck by his confidence, and impressed by it.” I’m delighted to be viewed as impressively confident, but what’s so impressive really? Does Salam think I meant that the penalties are adequate to overcome all the other differences of relevance between the Massachusetts and ACA laws? That would be an impressive assertion. Impressively dumb. If I meant that would I have pointed out that other important differences may exist and linked to a post that described one (the exclusion period)? No. Nor would I have written the conclusion Salam quoted, emphasizing that my analysis was based on “penalty size alone.” And I wouldn’t have started the piece with “Some have asserted that the individual mandate penalties under the Affordable Care Act (ACA) are lower than those imposed in Massachusetts.”
Thus, it should be clear to that I was documenting that there is no basis for an argument that the ACA penalties are lower than Massachusetts’ penalties. That’s not how one goes about supporting a claim that there will be gaming under ACA. Yet that’s the argument Salam had made in an earlier post in which he wrote, “My reading is that the penalties are considerably more onerous in Massachusetts than under the new federal legislation.” Since that is false, one has to base the argument on something else. That is, ACA penalties are adequate as far as penalties go, but other provisions of ACA may not be sufficient to prevent gaming.
I’m not saying there is no cause for concern about gaming. On that Salam and I may agree. I’m saying that penalty size differences aren’t the place to look for support of such a concern. Clear?
Lit Review: Health Insurance Benefits Mandates
Jason Shafrin is the only other health economist I’m aware of who routinely blogs. He deserves some credit for reviewing literature and posting references. His post today on the effect on premiums of health insurance benefits mandates is a good example. Here’s an excerpt.*
A recent paper by the Pacific Research Institute summarizes the findings of various studies of the impact of mandates on health insurance premiums.
- CBO (2000): 4 to 9 percent of premiums, all mandates aggregated
- Graham (2008): 5 to 23 percent of premiums, all mandates aggregated
- Bunce and Wieske (2009): 20 to 50 percent of premiums, all mandates aggregated
- New (2006): 15 percent of premiums, all mandates aggregated
- Congdon et al. (2006): 0.3 to 0.7 percent of premiums, per mandate above 20
- Wisconsin OCI (2002): 1 to 3 percent of premiums, five specific mandates aggregated
- GAO (2003): 3 to 5 percent of premiums, all mandates aggregated
- Krohm and Grossman (1990): 0.2 percent of claims, specific mandated benefits
- Maryland HCC (2006): 2 percent of premiums, all mandates aggregated
- Maryland HCC (2008): 0.01 to 1 percent of premiums per each of five specific mandates
… What one can conclude from the above studies is that mandates do increase cost. The degree to which health insurance premiums increase, however, is not a settled matter.
Of course the notion that mandates increase costs and premiums cannot possibly be controversial except in the case of a small subset of services the increase use of which might offset other, more expensive, health care utilization. An advantage of mandates, or standardization, is that it can decrease complexity and search costs for the consumer, making the market function more efficiently.
Note that costs are increased in two ways: (1) More benefits covered translates to higher insurer payout; (2) More benefits covered attracts enrollment from higher risk individuals. In a market with no standardization low-risk individuals could find less expensive insurance that covers fewer services. But such a market might segment risks so finely that the risk pooling mechanism of insurance ceases to function. That’s made all the more likely in a market with guaranteed issue and no pre-existing condition exclusion periods. Switching products to match needs to coverage is just an extension of the gaming problem I’ve been writing about lately.
Call for Guest Post or Reference
I’ve covered the extent to which ACA and Massachusetts individual mandate penalties differ in size. But there are so many other ways in which ACA and the Massachusetts law may or may not be similar that could have an impact on gaming and selection. Some of these are coming up in the comments.
- Does ACA have an open enrollment period? Does Massachusetts? (Restricting enrollment to certain periods increases the downside risk of waiting until one is sick to obtain insurance.)
- Does Massachusetts have stronger enforcement for failing to pay penalties than ACA seems to have?
- Does an exclusion period for pre-existing conditions exist in Massachusetts? (I think the answer is yes to this one.) I assume one does not exist under ACA, right?
If a credible expert on the Massachusetts individual mandate and penalties wishes to submit a guest post on those topics and how ACA differs for each, the door is open. Or, if readers can point to a summary that serves the same function, let me know. (Yes, I could Google about and figure all this out myself. But maybe someone out there with the expertise would like a little credit for saving me the trouble.)
Individual Mandate Penalties Are Adequate
This post has been cited in the 1 April 2010 edition of Health Wonk Review. See also my follow-up post on this topic.
Some have asserted that the individual mandate penalties under the Affordable Care Act (ACA) are lower than those imposed in Massachusetts. If that were the case then it would be one reason why one couldn’t generalize the experience in Massachusetts where guaranteed issue exists and near-universal coverage has been achieved with low penalties. If ACA penalties are lower than Massachusetts’ penalties then there is reason for concern that individuals might game the system–buying coverage only when sick, paying the low penalty when coverage isn’t needed–more than they appear to in Massachusetts.
So, are ACA penalties lower than those in Massachusetts? This is an empirical question, and I can answer it. The details are below, but to cut to the chase, the ACA penalty will be $674 for an average U.S. resident while the Massachusetts penalty would be $537 on average. That doesn’t mean the ACA penalty is higher for everyone. About 40% of the population would have a higher penalty under Massachusetts rules than under ACA rules. However, nearly half of those who would have a higher Massachusetts than ACA penalty are exempt from ACA penalties due to low income. Many such individuals are eligible for premium and cost sharing subsidies under ACA. Thus, the incentive for gaming is lower for this subset.
So, I don’t think it is fair to say the ACA penalties are lower than Massachusetts’ penalties. On average they’re higher, and they’re higher for 60% of of the population. If gaming is low in Massachusetts we cannot expect it to be higher under ACA based on a penalty-size argument. Hence, ACA penalties are not too low. However, the U.S. population may differ from the Massachusetts population, and other details of ACA differ from health reform in Massachusetts. For these reasons, gaming may still be an issue despite the evidence on penalty size. Keep reading if you want the details.
Let’s first look at the Massachusetts penalty schedule for 2010:
FPL = Federal Poverty Level. This table is copied from the Massachusetts Department of Revenue website. I believe the 18-26 age specification in the 250.1-300% band is in error, that the dollar figures in that band apply to all ages. The penalty figures shown are per adult (i.e., married couples pay double, kids are exempt).
When fully phased in (2016), the penalty under ACA will be $695 per person per year up to a maximum of three times that amount ($2,085) per family or 2.5% of household income, whichever is greater. For the penalty calculation, children under 18 count as half a person (i.e. lead to a penalty of $374.50–h/t reader Jacob Shmukler). Individuals with out-of-pocket (OOP) premium-to-income ratio above 8% are exempt from the penalty. Here, OOP premium is net of employer contribution or exchange subsidies (source for premiums: 2009 Kaiser/HRET Employer Health Benefits Survey; premiums for employer-sponsored plans not reduced by the tax subsidy, which is conservative).
With a nationally representative source of income data we can calculate what proportion of individuals would face lower penalties under ACA than in Massachusetts. To answer these questions I turned to the Medical Expenditure Panel Survey (MEPS) because I have it handy and am very familiar with it. One could also use the Current Population Survey or any number of other nationally representative surveys with income data. Because it is the latest available, I used the 2007 version of MEPS. I didn’t trend incomes forward to 2010, which is conservative to the extent incomes went up (but given the economy they likely have not).
I computed the penalty paid by each family in the MEPS sample under each set of rules, ACA and Massachusetts. I then assigned to each individual in each family an equal share of each penalty and averaged the penalties over the population, weighted appropriately to compute national means. I also computed the number of individuals for whom the ACA penalty would be greater than the Massachusetts penalty. Results:
- Mean ACA penalty: $674
- Mean Massachusetts penalty: $537
- Percent of population for whom ACA penalty > Massachusetts penalty: 60%
These results are qualitatively robust after stratifying according to exchange subsidy eligibility. Since there is little evidence of substantial gaming in Massachusetts, based on an analysis of penalty size alone there would seem to be little cause for concern over gaming under ACA, particularly for higher income individuals. This analysis ignores other differences between Massachusetts and its health reform law and the national population and ACA, respectively. Results are sensitive to assumptions, but I deliberately selected those conservatively as indicated above.
Gaming the Individual Mandate
Bloggers over on EconLog are anticipating some gaming of the individual mandate (Arnold Kling, Bryan Caplan). I agree with them that the individual mandate is sufficiently low that it could make financial sense for some folks to wait until they are sick to enroll. I don’t agree that there is enough incentive for employers to drop coverage, or certainly not in large numbers. Here’s why.
Kling and Caplan are ignoring the reason why employers offer coverage, to compete in the labor market. They compensate with health insurance at the expense of wage due to the employer tax subsidy. That subsidy is huge and will not be available for exchange-based plans for large firms (at least not initially, and ultimately at the discretion of states; the distant future is uncertain in this regard).
Workers in the labor market, and especially the older, experienced, and highly valued ones with families, want health insurance and, moreover, want it through their employer. Thus, if a firm doesn’t offer insurance it will lose access to the class of workers who value it. For larger firms that’s going to be a lot of people, some they can’t afford to lose. Doing so will put them at a competitive disadvantage in the labor market and the quality of their products will suffer. That doesn’t sound like a good business plan.
Smaller firms might rationally decide not to offer insurance. But many already make that choice. Health reform law now includes tax credits for those businesses to offer insurance. On net I think we’ll see an increase in small businesses that do so.
And finally, nobody has explained why we should ignore the experience in Massachusetts where guaranteed issue exists and near-universal coverage has been achieved even with low penalties. Maybe the six month exclusion of coverage for pre-existing conditions in Massachusetts is enough. That could be replicated nationally, though simply raising the penalty would eliminate the adverse selection problem Kling and Caplan point to. Either way, it is conceptually a small tweak to the system. Such a tweak may not be necessary, but if it is at least the structure is in place that can accommodate it. None of this is good reason to condemn that structure.
Or the answer could simply be that that nearly every healthy individual in Massachusetts is irrational. In that case, maybe you can’t believe anything I say.
Later: See my follow-up post that shows that the individual mandate penalties are not low, at least by Massachusetts standards.
Adverse Selection in Insurance Markets
A late 2009 NBER paper by Alma Cohen and Peter Siegelman reviews the literature on adverse selection in insurance markets. Adverse selection becomes a problem when individuals of higher risk than expected (specifically, than reflected in the premium) purchase coverage. Higher risks lead to higher costs and higher premiums, which induce lower risk individuals to forgo coverage, increasing per insured costs further–a classic death spiral.
Cohen and Siegelman’s paper “Testing for adverse selection in insurance markets” includes a section on health insurance, but covers many other types of insurance as well (auto, life, long term care, crop, etc.). As for health insurance, a large body of empirical work supports the claim that adverse selection exists in health insurance markets. Cohen and Siegelman (hereafter C&S) cite a review article by Cutler and Zeckhauser as one source for accessing this literature.
Some prior reviews of the literature on adverse selection conclude that the evidence is “mixed,” “inconclusive,” or “ambiguous.” C&S
argue that one should not expect the question of whether a coverage–risk correlation exists to be answered identically in all insurance markets or even in all pools within a market. Thus, one should not regard studies that reach opposite conclusions about the existence of a coverage–risk correlation as necessarily in conflict with each other. (© 2009 by Alma Cohen and Peter Siegelman.)
This is a crucial point because it suggests important limitations in our ability to generalize from one market or sub-market to another.
C&S make brief mention of risk adjustment, a means by which the effects of adverse selection can be mitigated by compensating the insurer for the level of risk of its insureds. There is widespread misunderstanding of the degree to which adjustment solves the adverse selection problem. It is never complete, often far from it. C&S tell us why and back it up with literature citations: insurers and their regulators do not always have access to or make use of all information relevant to risk.
The paper includes a section that describes the ways in which adverse selection can fail to appear. One class of ways essentially boils down to policyholders not having an information advantage over insurers. If insurers have sufficient information relative to policyholders they can price their products to account for the expected level of risk each will draw, eliminating the problem of adverse selection.
Another way in which adverse selection can be offset is if a subset of the population that purchases insurance is a source of favorable selection. For instance if cautious people are more likely to buy insurance and more likely to prevent claims they will provide a source of favorable selection. C&S provide other examples like this. It is tempting to view this whole line of reasoning as: selection can fail to be adverse if it isn’t. The value added is the explanation of why it is not adverse.
Selection into an insurance product can be driven by non-personal factors. Institutional or regulatory factors can be the dominant factor in selection. Examples include the employer-based health insurance market, or the high rate of subsidization of Medicare drug plans. One should not expect adverse selection in such cases.
Even when selection is not adverse, coverage can lead to higher utilization of covered services, an effect known as moral hazard. C&S review the literature that attempts to disentangle adverse selection from moral hazard.
Finally, C&S conclude with policy implications of their findings:
Policy discussions should try to tailor themselves to the specific insurance market under consideration, recognizing that adverse selection and coverage–risk correlations vary across insurance markets (and even among pools of risks within a market), and that they do so in ways that are at least somewhat predictable on the basis of existing research.
This is good advice, but no doubt hard to follow. Much of the utility of empirical research flows from its generalization. If such generalizations are shown to be misleading, the import of each piece of work is greatly circumscribed. In practice, judging the credibility of generalizations is as much art as science. Knowing more of the details of the relevant market matters in getting things right (and certainly in getting published), but policy messaging that actually makes a difference rarely relies on the nuances.
Can Risk Adjustment Save the Public Option?
Ezra Klein has a truly excellent post on adverse selection and the public option. He concludes with, “The most important factor here will be the strength of the risk adjustment in the exchanges, so keep an eye on that.”
I wonder how optimistic we can be about the degree of variation in spending predicted by risk adjustment models. I think the answer is “not very.” From the literature on health care risk adjustment (via this post):
Statistical models developed by scholars have relatively low predictive power. Predicting ten percent of the variation in [health] expenditure is considered good (e.g., Medicare Advantage’s risk adjustment model). That means ninety percent of the variation is unexplained by the model or chalked up to random error. An individual ought to be a better predictor of his or her health expenditures than a model that cannot include measures unobservable to the researcher. (How much better? I don’t know.)
Expenses for some specific services are more predictable. Drug expenses, for example, are persistent because individuals tend to use the same medications year after year. The best statistical models of drug spending can predict about 55% of the variation in next year’s drug expenses, leaving 45% to random error.
That puts a reasonable cap at 55%, but only for very persistent services, like drugs. Expect the best overall risk adjustment to be no worse than 10% and no where near as good as 55%.
Private insurers should not be so worried but taxpayers should. The public plan looks game-able.(*)
(*) A wonky note: It isn’t game-able because the risk adjustment model is of low power. It is game-able because insurers likely have access to information not observable to researchers and omitted from the risk adjustment model, which makes it lower power than it could otherwise be. The risk adjustment model was developed in a political environment in which the insurers were participants.
Subscribe
