A few points regarding "Obama Dentata"

(Yes, I realize that title may be too clever by half. Anyone know if this is actually an offensive reference?)

OK, I wasn't going to post anything else about this today, really. I was planning on holding off until tomorrow. However, it's turned into a Thing (of course), and Things have a way of sucking the oxygen out of the room. So...

• First of all, to reiterate: In terms of how the ACA is functioning, the fact that the net attrition rate has turned out to be 2.7% (6.7 million on October 15th / 8.0M on April 19th = 83.75% retention over a 6 month period) was normal and expected. I thought it was about 3% per month until the "7.3M in August" announcement caused me to drop it down to 2%...and it turns out I was right in the first place, both for August as well as October, when I suggested that the number may have slipped to as low as 6.8 million.

• This afternoon, HHS Secretary Burwell issued a full apology:

• As for how it happened, the HHS Dept. is claiming that it was a genuine mistake for both the August and October numbers; Jonathan Cohn has the official response:

Administration officials told me that, during the spring, they used an internal system for tracking how many people were shopping and selecting new insurance plans. That system distinguished between health and dental policies, and didn’t count the latter towards the sign-up totals. But, as you may recall, that data didn’t provide a critical piece of information: How many people were actually paying for their plans. To get that information, HHS got information directly from insurers and used it to produce the autumn numbers. When they did so, they didn’t think to break out dental plans—and ended up including them accidentally.

Huh. Um...OK. This would explain how it happened the first time...but they did it again with the October "7.1 million" estimate. The HHS Dept's defense here is basically "We aren't dishonest, we just screwed up royally" which doesn't exactly fill my heart with joy either.

On the other hand, that actually is a plausible explanation, given that the only alternative is that they deliberately fudged the number by a whopping...4.3%, for no particular reason.

So, do I believe them? Honestly, I don't know. I'm just very, very tired right now; it's been a very long day, and I've already been accused by at least one "supporter" of being "the reason why the ACA polls badly":

My response, of course, is that if that's why he feels that way, why is he following me in the first place? "Publicly aggregating and reporting accurate ACA enrollment data" is kind of my whole thing, after all.

• So, what actually happened to the 1.3 million (using the corrected October 6.7M figure, not the corrected August 7.0M figure) who were dropped from the 8 million April figure?
  • About 960K of them never paid their first month's premium (12% didn't pay).
  • Another 112K of them were given the boot on September 30th because their legal citizenship/immigration status couldn't be verified.
  • Assuming 2.5% of the remaining 6.95 million were 64 years old (25% of the 8M total were between 55-64), that means around 170K 64-year olds. About half of those should have turned 65 by now (all of the original 8M policies had started by May 1st). That's up to 85K who've moved to Medicare (I don't know how long it takes to go through that process).
  • The remaining 143K or so dropped their coverage for a variety of reasons, including:
    • Some of them were hired to a job with benefits and are now on ESI. That's a good thing.
    • Some of them fell on hard times and moved to Medicaid. That's not a good thing, but at least they're still covered.
    • Some of them may have left the country, married someone else with ESI, etc.
    • Some of them...well, died, frankly (it happens).
      Yes, all of these last 4 items have also been cancelled out during the off-season by people being added for similar life changes, but the point is that there's nothing unseemly or unusual about any of them.
    • Finally, some unknown number (as of yet) fell into the "3 month grace period" gap: That is, they paid for one or two months and then coasted for 90 days until their insurance company was finally able to give them the boot.
  • In other words, there's nothing nefarious going on here. Some people just don't pay their friggin' bills, and yes, some people simply can no longer afford the premiums for various reasons, whether due to their own actions or not. That sucks, but let me ask you this: What was the attrition rate in the private insurance market before the ACA exchanges went into effect? I'd imagine it was the same 2-3% per month:

After analyzing patterns in the nongroup health coverage market from 2008 to 2011, Harvard professor Benjamin Sommers found that the market was characterized by high turnover before the Affordable Care Act was implemented. Additionally, Summers found that most people who left the nongroup market acquired other insurance within, suggesting that the cancellations attributed to Obamacare are unlikely to have a significant impact on overall coverage rates.

Relying on data from the Census Bureau's Survey of Income and Program Participation for his study, Sommers used a sample of 4,199 respondents under 65 years old with nongroup coverage in their first month in the survey.

Over one-third of those respondents no longer had nongroup insurance coverage after just four months, Sommers found. Just 42 percent had stable nongroup coverage after one year. After two years, just 27 percent had stable coverage.

Let's see: That's 58% of individual policy enrollees changing to some other form of coverage over a 12 month period...or 4.8% per month. Yes, many of these people should be expected to switch from one ACA policy to another due to life changes (which would mean no net attrition on the exchange total), but many of them wouldn't be expected to anyway. In short, a 2.7% net attrition rate is nothing to be upset or worried about.

In other words, consider this a simple lesson learned: Whatever the official number is as of February 15th, 2015 (which I still think will end up being 12 million in spite of today's revelation), about 12% of them probably won't pay their first month's premium (which is why I have the paying estimate down as 10.6 million). Then, knock off another 4% over the next 9 months for various reasons and you'll probably end up with about 84% of the original number (which, going by my projection, should mean about 10.1 million by next December).

On the other hand, these numbers should improve this year (and in the future) as the payment systems are streamlined, the immigration/citizenship databases are updated and so on. I'm guessing that instead of 12% non-payments and 1.4% being dropped for "legal status" issues, these will be more like 10% and 1% next year, and hopefully even better after that.