Chris Conover's Truth Twisting on the ACA: Part Two

The other day, Paul Krugman posted a short blog entry over at the New York Times in which he laid out 6 of the biggest anti-ACA attack points which have been, in his view, completely obliterated by reality:

1) "No one will sign up!"
2) "But how many have PAID????"
3) "OMG!! 5MM POLICIES CANCELLED!!!"
4) Rate Shock
5) "Not enough YOUNG INVINCIBLES to avoid a DEATH SPIRAL!"
6) "Overall healthcare costs will skyrocket!!"

In response, Chris Conover has posted a piece over at Forbes which claims to rebut Prof. Krugman point by point, basically calling him full of beans on 4 points and "generously" giving him a split decision on 2 others.

Now, I'm not going to comment on the last three of these because they really aren't my area of expertise (although I will note that even my old sparring buddy Avik Roy--also of Forbes--has flatly admitted that the "death spiral" hand-wringing was proven false...and that was way back in February).

In addition, I've already addressed the third point ("cancelled policies") in a post way back in early April, where I provided solid evidence that the "5 million" claim is vastly overstated; the number of policies actually cancelled in the end was no more than 1-2 million at most.

However, on the subject of the first two, I think I'm in a pretty good position to respond.

I addressed the first point regarding the total number of QHP enrollments in Part One.

Point 2: But How Many Have PAID??? Conover writes:

Purported Incorrect Prediction #2: Even if people sign up, they won’t pay their premiums.

And in fairness, this was backed up by a McKinsey survey conducted April 7-April 16 showing that 83% of those previously uninsured and 89% of those previously insured had paid their first month’s premium for new coverage.  A recent Kaiser Family Foundation survey shows 57% of Exchange enrollments were previously uninsured. That implies a weighted average payment rate of 85.6%; so even if we generously credit 8 million sign-ups, 85.6% x 8 million = 6.85 million, meaning enrollments still fell short of the 7 million benchmark.

Not surprisingly, Obamacare enthusiasts have argued that many of the late sign-ups would not have had to pay their first-month’s premiums at the time McKinsey conducted its survey. In short, the payment rate is underestimated. This is a legitimate point, but a) the historical payment rate in the nongroup market was only 90%; b) it ignores evidence that 2-5% of people who paid their first month’s premium on the Exchanges failed to pay their second month’s premium. Thus, it’s no surprise whatsoever to learn that the administration has no plans to release further monthly enrollment reports. April enrollment may well have represented the apogee of sign-ups in 2014: why broadcast the steady erosion of that progress month after month?

OK, let's unravel this onion: A study conducted in mid-April gives an 85.6% payment rate of the first month's premiums...and that, again, only included about 5 million people whose 1st premium was even due yet (the other 3 million had policies which didn't even start until May 1st...in fact, several hundred thousand people had just enrolled a day or two before the poll was conducted; they hadn't had time to even enter their credit card number or pull out a checkbook yet.

He "generously" uses 8 million as the high-end number, even though, again, it's up another half-million or so over 2 months later. He does correctly acknowledge that "many" of the enrollees payments weren't due yet ("many" actually means "over 37%"), and acknowledges that it's "a legitimate point", and then states that a payment rate of "only" 90% is a historical average in the individual private market to begin with...which doesn't sound bad to me at all for a subscription service (I have no idea what the average payment rate is for, say, Cable TV or Gold's Gym memberships; that would be interesting to know).

Now, Conover does provide another interesting tidbit: He claims that the 2nd month drop-off of those who paid their first premium is 2-5%. This is actually a useful stat to know, if true, since it helps understand the overall "drop-off" rate as time goes on. Of course, as I've noted before, just because someone "fails to pay their second premium" doesn't mean that they're a deadbeat; in many cases it means that they got a job with benefits, got married to someone who did, joined the military, fell on hard times and now qualifies for Medicaid, turned 65 and now qualify for Medicare, or any of a dozen other very good reasons. In some cases, tragically, they might have, you know...died. It does happen to everyone sooner or later, after all. That's the "steady erosion" bit that he's talking about, although Conover seems to be suggesting that all of those people are simply deadbeat/moochers who refuse to pay after grudgingly ponying up the first month. I'll return to the "2-5%" bit in a moment.

Conover then makes possibly the dumbest, least-shocking claim possible: April enrollment may have represented the high point for the year. Um...yes, I'm quite certain that it did. The HHS Dept. expected it to be, just as it was in Massachusetts the first year of Romneycare. In other breaking news, states with early voting report that the vast majority of people still cast their ballots on election day instead of weeks beforehand. SHOCKING!!

Now, let's look at the payment data we do have; remember, in some cases these include people who have enrolled since open enrollment ended...and pay special attention to the dates:

  • Alabama reported 82% as of 5/08
  • California reported 85% as of 4/27
  • Connecticut reported 100% as of 5/22
  • Maine reported 90% as of 5/08
  • Masssachusetts reported 100% as of 5/01 (MA requires payment to even report the enrollment to HHS)
  • Minnesota reported 95% as of 5/27 (and continuously since)
  • Nevada reported 76% as of 5/28
  • North Carolina reported 85% as of 5/10
  • Oregon reported 91% as of 6/24
  • Rhode Island reported 91% as of 5/03
  • South Carolina reported 71% as of 5/07
  • Vermont reported 77% as of 6/16
  • Washington reported 100% as of 5/27 (like MA, WA requires payment to report the enrollment to HHS)
  • West Virginia reported 94% as of 5/03
  • Wyoming reported 92% as of 5/20

Kind of a mixed bag, right? Well, when we tally 'em up, we get the following:

2,254,978 paid out of 2,590,988, or 87% overall...and over half of the total (54%) was from California, which reported their number way back on April 27 (again, 3 days before many of the policies started). An additional 39% were last reported in the first third of May (many insurance companies have grace periods of up to 30 days). These 15 states, combined, represent about 1/3 of the total enrollments nationally.

I think it's more than reasonable to assume that exchange enrollments nationally are indeed matching the same 90% rate that was (according to Conover's own statement) typical in the industry outside of the ACA. Earlier this spring I was using a 93% "paid or will pay" rule of thumb, but I've since dropped it back to a flat 90% for this very reason, and am extremely confident about this.

Next, let's consider Conover's "2-5%" attrition rate. He uses this to refer to the 2nd month drop-off, but doesn't mention anything about the 3rd, 4th and so on; I can only assume that he's suggesting this will be typical each month. Let's split the difference at 3.5%.

However, note that he isn't saying 2-5% of the full 100% of enrollees--his exact statement is "2-5% of people who paid their first month’s premium". I've already shown in the previous entry that between 9.4 - 10.5 million people will have enrolled by the time the 2nd open enrollment period comes around; again, let's split the difference at an even 10 million and see what we get, using Conover's own attrition rate claim and assuming that it holds true at exactly 3.5% every month.

  • 10.0 Million enrolled
  • 9.00 Million paid 1st month's premium (90%)
  • 8.69 Million paid 2nd month (96.5% of 9M)
  • 8.38 Million paid 3rd month (96.5% of 8.69M)
  • 8.09 Million paid 4th month (etc...)
  • 7.80 Million paid 5th month
  • 7.53 Million paid 6th month
  • 7.27 Million paid 7th month
  • 7.01 Million paid 8th month

After the 8th month it does indeed drop below the 7 million mark...except that again, it's a rolling average, and 50% of the 10M total enrolled in policies which didn't even start until May or later (5 million enrolled as of 3/15, another 3 million between 3/16 - 4/15, and anther 2 million between 4/16 - 11/15). Most additional drop-offs wouldn't apply until 2015 anyway, at which point we're into an entirely new enrollment period where the first year numbers are no longer relevant.

In other words, using Conover's own attrition rate claim, we're very likely to end the 2014 calendar year with almost exactly 7 million fully-paid exchange-based QHP enrollments...precisely the number that the CBO was projecting in the first place.

Imagine that.