Tag Archives: enrollment

Stop reporting enrollment and start reporting policies in force

In theory, open enrollment ends tonight. No longer can individuals without any excuse whatsoever wait to purchase health insurance policies on the Exchanges that will be in effect during 2014.   I very much expect we will hear numbers over the next few days such as 6.5 million or 7 million bandied about as “enrollment figures.” Many supporters of the ideas behind the original ACA who have managed to tolerate its metamorphosis over the past year will herald those numbers as signs of success.  And, indeed, those numbers are considerably better than many had feared.  I, for one, am prepared to confess that I may have been too pessimistic in the past about first year enrollment in the ACA. My pessimism is all the more glaring because the enrollment numbers are apparently coming notwithstanding the Obama administration’s decision to tie its hands behind its back by creating a new opportunity to evade the individual mandate via an undocumented “hardship exemption” and to delay making a  purchase decision based on “honor system” claims of difficulties in accessing healthcare.gov or state enrollment systems.

That said, however, these enrollment figures are essentially irrelevant.  Those who persist in touting them as signs of success reveal themselves more as Obamacare fanboys than as credible advocates or serous scholars.  The aggregate enrollment figures are irrelevant for two reasons and of lesser value for another. In short, my predictions may have been on the low side, but the numbers that will be coming out tonight absolutely do not vindicate those who assured us that Obamacare would end up being just fine. Reciprocal honesty would be nice.

Nothing economically or legally in the ACA turns on “enrollment”

1. “Enrollment” is, as many have noted, simply checking boxes on a web site.  It is not the same thing as actually having a policy in force, of being covered by an insurer who will pay medical bills in the event the “insured” becomes sick. In terms of the real health of Americans, access to expensive medical care, enrollment numbers are as meaningful to the success of the ACA as knowing the number of individuals outfitting their fantasy Corvette Stingray on Chevrolet’s“build your own” web site is to the success of GM. The former is a leading indicator of interest in plans on the Exchanges just as the latter is a possible leading indicator for purchasing a snazzy car. But that, really, is all.  Responsible journalists and bloggers should stop bleating “enrollment” merely because it is the only number the Executive branch dispenses.

Friends of the Obama administration gain credibility and actually help their President by insisting that the Obama administration release the number of policies in force  This is so because there are only two and a half conclusions that can be drawn from silence: either the Obama administration has the number and is refusing to provide it or it just doesn’t have the number.  Not releasing a number or decent approximation is contrary to the transparency values that the Obama administration espouses.  My FOIA request on the topic is, like many others, I suspect, unanswered. Not having the number is perhaps even worse.  I have read much of the ACA and its regulations and I can not think of a single economic or legal matter that turns on the number of people who have “enrolled” in plans on the Exchange.  Insurers get paid, tax credits are advanced, and a whole host of other important financial and legal consequences depend instead  on the number of people who, for any given month, actually have coverage, have a policy in force.  I say “two and a half” conclusions because there is a variant of not having the information that is sometimes advanced — not having precise enough information to release.  But this is about as flimsy as it gets.  Governments release approximate numbers all the time.  I believe Americans are sophisticated to understand the word “about” and the concept of a “good faith estimate.” Any lack of precision down to a single individual does not excuse a failure to release relevant information. I am confident that the public would be well served by having a policies in force number that was accurate even to just two or three significant digits.

The difference between enrollment and purchase is not trivial. Suppose the fall off between enrollment and purchase is, as some have suggested, 20%.  And suppose further, as again some have said, that 3% of  the remaining policyholders don’t pay each month.  The graph below shows the fraction of policyholders persisting over 12 months.  The blue line shows the time series and the yellow dotted line shows the average level of policies in force. As one can see, by the end of 9 months, only 65% of the initial level of policies remain in force. The annual average is about 70% of the starting amount. Moreover, the healthy are the most likely to stop paying; those who are sick are most likely to persist in their policies. Thus, if 6.5 million enrollees start out, only 4.2 million will be left with policies at the end of 9 months and the average number of purchasers over the year will be 4.6 million.


Even if just 10% decline to pay their first month’s bill, only 70% of the policies remain in force at 9 months (see figure below on left); the annual average is about 79%. At 20% initial decay and 5% monthly decay thereafter, only about 55% remain after nine months; the annual average is 65% (see figure below on right).  And some, of course, have suggested the fall off between enrollment and retention of a policy is actually worse.


Note: it is possible that some of the losses due to non-payment will be offset by those purchasing policies via “special enrollment.”  These individuals, however, may be particularly high risk.

2. The state-by-state totals matter a lot; aggregate purchases matter little

Aggregate purchases or, worse yet, aggregate enrollment is largely irrelevant to the success of the ACA.  There are at least 51 markets for individual policies.  The fact that California may have exceeded expectations in terms of purchases does nothing to help Texas, Louisiana, Arizona and other states where enrollment has been low.  Insurers in states where enrollment is low or the demographics are particularly problematic — few young people, lots of middle age women — will not be compensated for their losses by the fact that enrollments are better in California, New York and Connecticut.  Even if it is the same parent company that makes money in one jurisdiction, that will not deter the subsidiaries in losing jurisdictions from either withdrawing or raising premiums. There will still be immense pressure on insurers in the less successful states to either drop out entirely — something the make-it-up-as-you-go-along implementation of Obamacare  fosters — or to raise prices substantially.

As I said in a story broadcast on National Public Radio and as the New York Times admits, we need to stop thinking about the ACA as a single narrative and start to think about it as multiple complex narratives.  Indeed, it’s probably more like 175 or more narratives (the number of issuers in all federal plans) because not only will the experience vary between states, they will also vary by insurer.  An insurer who charged a very low premium in a given state and attracted a good deal of business may react very differently next year in pricing policies than an insurer even in the same state who charged a high price and got less.

3. Experience matters most

But the life of the Affordable Care Act will not be enrollment or even purchases, it will be experience.  Did insurers set prices realistically or did they underestimate the medical problems and demand for services of enrollees?   Does the fact that many of the purchasers on the Exchange actually had policies already actually help insurers because many of those purchasers would have undergone at least some recent medical underwriting? Are the networks that have been created by insurers so narrow that they will lead to unsatisfactory medical care or will they in fact keep prices low? Will insurers regard uncertainty in the political environment  — like not knowing whether Obama will extend the inchoate hardship exemption into next year when, otherwise, the individual mandate/tax/penalty more than doubles — result in at least some insurers pulling out of the market? Will the diminishing reinsurance available to insurers writing in the Exchanges have the effect I predicted of increasing prices by about 7%?


Tonight, March 31, is a milestone for the ACA, but it is hardly the end of the challenge. We’ll see lots more important data start to dribble in over the next few months, including, critically, information on premiums insurers hope to charge in 2015. In the interim, though, could everyone please start to focus on statistics and data that matters rather than proxies such as “enrollment.”  Use of enrollment rather than policies in force may at one time have been a necessary evil. But persisting in doing so is a practice that is no longer useful other than as a vehicle for spreading political propaganda.

My suggestion, only slightly tongue in cheek, is that, just as the IRS imputes some awful income to you if you don’t file a tax return, journalists and bloggers start reporting appallingly low purchase numbers until the Obama administration releases the actual data. I’ll start:

“Tonight’s disclosure by the Obama administration that about 4.5 million individuals will have coverage via Exchanges under the Affordable Care Act during 2014 means that some big states such as New York and California have done reasonably well in making it likely that their insurance markets will be stable. It also means, however, that Exchanges could be under great stress in a number of states, most notably Texas.”




My rough estimate of the rate of decline in Obamacare enrollment appears to be vindicated by an article appearing here in Kaiser News titled  “Why Some Don’t Pay Their Obamacare Premium: It’s Not What You Think.”  Kaiser reports that Covered California, the Exchange for the nation’s largest state, has produced a report projecting a significant drop in the number of enrollees throughout the year:

According to the report between 53 and 58 percent of Covered California enrollees are expected to stay in a Covered California plan for 12 months. This analysis is consistent with a Kaiser Family Foundation study published earlier this year. It found that of people who enrolled in an individual insurance plan in 2010, years before the health law fully kicked in, only about 48 percent were still in the individual market two years later. (Kaiser Health News is an editorially independent program of the foundation.)

But most of these people dropping ACA Exchange coverage won’t become uninsured, the report says.  Instead, they will go on to the state’s expanded Medicaid program or find better/cheaper coverage elsewhere. It’s not clear from the Kaiser article or the Covered California report whether they expect those moving out of the Exchanges to be healthier than average.  In any event, though, the report further establishes why touting “enrollment” is ridiculous.

Share Button

Colorado enrollment numbers look bad

According to a news report today from the Associated Press, the enrollment numbers in Colorado are about half of what the state had projected as its “worst case scenario” for enrollment in its Exchange.  The  6,001 Colorodoans who have enrolled thus far (and that is enrolled, not yet paid) is also way less than the 20,186 that had served as the midpoint projection for this time in the open enrollment season. Colorado officials apparently believed that 136,000 people would enroll by the end of 2014. Enrollment  thus far has apparently been complicated, however, by a requirement that  consumers seeking a subsidy on the exchange first find out if they qualify for Medicaid, which may require a 12-page application.

Three facts make the low Colorado numbers yet more troubling for those concerned either about a reduction in the number of uninsured individuals or about the prospects that those getting enrollment in the Exchanges for 2014 will shortly find themselves sucked into an adverse selection death spiral.

First, these low figures are coming in a state where the web site has, for the most part, been operating respectably (further details here) and not in a state dependent on healthcare.gov.

Second, it comes in a state where between 106,000 and 250,000 individual health insurance policies have been cancelled. I have an inquiry in but have not yet been able to find from published sources whether Colorado will permit insurers to reinstate those policies and under what conditions. [LATE BREAKING NEWS at 13:52 Houston time: Colorado insurance official says Colorado expects to announce a decision on this point later today or tomorrow (per email from Vincent Plymell, Communications Director at the Colorado Department of Regulatory Agencies, who courteously responded to my inquiry)]

Third, Colorado already had 1,227 people enrolled in the Federal Pre-Existing Condition Insurance Plan (PCIP). One would think those people, who tend to have high medical expenses and are willing to pay full freight for insurance, would have been among the earliest to sign up under the Colorado exchange with the potential for subsidized rates.  Thus, some of the 6,001 already enrolled may not represent a reduction in people without insurance to people with health insurance but merely a substitution from one federally created plan to another.

Share Button