UPDATE: I'm now inclined to believe HHS screwed up instead of deliberately misleading...not that that helps their image much.

So, it's been 3 days since the Great #DentalGate story broke (yes, I'm abandoning my efforts to make #ObamaDentata the Officially Accepted Stupid Media Label for this incident...no one else seems to be running with it, so it seems that slapping "-Gate" onto every scandal/alleged scandal in politics is inescapable these days).

When Alex Wayne of Bloomberg News asked for my response before he broke the story publicly, I said that, assuming the story was accurate, I was "appalled". I stand by that response, because I am appalled regardless of whether it was done deliberately or by mistake.

If it was deliberate, it means that HHS/CMS padded their enrollment data. Not only would this be dishonest, it would be stupid because a) the true numbers are still pretty damned impressive and worth being proud of, and b) because they know that there's no other policy in the Obama administration which has been the subject of as much hatred and as many attacks as the Affordable Care Act, whether valid or not...and they had to know that the Republicans, regardless of how many lies about "death panels" and "socialized medicine!!" they spew, would go over any number HHS put out with a fine-toothed comb.

In other words, in addition to the honesty factor, this scenario would mean that they padded the true August number by a whopping...4.6%. Just to nudge the official number over an arbitrary line which was never actually necessary for the law to be considered successful in the first place. It would be like getting a B+ on your final exam and then faking the grade & risking being suspended just to make people think you scored an A. That has to be one of the worst risk/reward ratios I've ever heard of, and is actually a strong argument in favor of believing HHS when they say that it was just a big mistake on their part.

In fact, I'll go you one further: According to Wayne, the actual "total" number (ie, the one you get by double-counting the 393,000 dental plans) was 7.37 million, which is closer to 7.4 million than 7.3 million. And yet, when providing testimony to Congress, Tavenner chose to knock the extra 70K off, saying "7.3 million" instead of "nearly 7.4 million" which would have been "accurate" under the circumstances. So, why would she underplay the number, if only by 70,000 people? Well, ironically, my guess is that she and her people at CMS were indeed deeply concerned about the Congressional committee claiming that they were "padding the number" if they rounded up to 7.4M, so they decided to err on what they thought was the side of caution. Again, this makes the "we just screwed up!" defense much more plausible.

Taking that one step further, the actual math was: 7,370,000 minus 393,000 double-counted dental plans equals 6,977,000 actual QHPs as of August 15th.

Even the Republicans wouldn't be stupid enough to try and make hay out of rounding up by 0.3%. If Tavenner had simply said "7 million even as of 8/15" the GOP would have gotten no more politcal mileage out of "dropped by 1 million!" than they did by actually screaming about the number having "dropped by 700,000!!"

On the other hand, if it wasn't deliberate, then it means that HHS/CMS was sloppy/lackadaisical about tracking the enrollment data.

For heaven's sake, they were providing Congressional testimony to the House Oversight Committee, which was being run by Darrell Issa. Yes, the whole Benghazi thing has now been proven (by the Republican-led House Intelligence Committee) to be a big fat nothingburger...but even so, as I noted above, they knew that he'd be putting the screws on as tight as possible...which was all the more reason to double-check their own math carefully before giving "7.3 million" as the August number.

Here's the official explanation they provided to Jonathan Cohn of The New Republic:

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.

OK. That could have happened, and certainly would make more sense than fudging the number by less than 5% for no particular reason.

The problems I have with this scenario are:

  • They did this not once, but twice (in September, for the August number and in November, for the October number...although if they truly didn't think about the dental plan factor the first time, it actually makes sense that they wouldn't think about it the second time either, so I suppose that's not such a big thing).
  • According to Wayne, he gave HHS ample time to respond prior to publishing the story at Noon on Friday: From the previous evening until the story went live. I don't know what the normal procedure is for this sort of thing, but if I was whoever he spoke with at HHS, I'd be having a "Oh, crap..did we screw up the numbers??" moment and at the very least give a "we'll have to look into this" response immediately, followed by scrambling to verify/confirm the situation.
  • The wording of the press releases for the August and October numbers ("enrolled in Marketplace plans" as opposed to "enrolled in Qualified Health Plans" which had been used previously).

UPDATE 10:04pm: OK, this last point appears to be less clear; as Andrew Sprung has noted, they also used the term "Marketplace Plans" in the official reports from last spring, although not necessarily in the accompanying press releases. It looks like the "QHP" vs. "Marketplace Plan" thing may not be that telling after all.

So...do I believe HHS/CMS when they say that this was a case of screwing up royally rather than a case of giving a deliberately misleading number? (sigh) I honestly don't know.

UPDATE 10:04pm: OK, Ken Kelly has pointed out another interesting tidbit: The 393,000 dental plans, which I had assumed was a huge drop from 1.1 million nationally back in April, actually represents only the dental plans from the state-based exchanges. That is, the May report stated (on page 19) that there were 1.13 million standalone dental plans on the federal exchange (Healthcare.Gov), but an unknown number from the state-based exchanges at the time:

In other words, in April they had the 1.13 million dental plans through HC.gov but no dental data from the other 15 state-run exchanges. Then, when the insurance companies from all 50 states (plus DC) sent in the payment data for August, it included the dental data from those 15 states in the mix. Instead of 1.13 million total dental plans, it was really more like 1.13M + 393K = 1.52 million dental plans in all (perhaps closer to 1.6 million if you assume the 393K was down slightly due to attrition).

My guess is that by the time September came around, the HHS folks basically had forgotten that they didn't previously have dental plan data for those 15 state exchanges. They presumably thought that they had already separated out 1.13M from the 7.37M tally and forgot that they still had to go in and subtract dental plans from those 15 exchanges as well.

Based on this missing piece of the puzzle, I'm now much more inclined to believe that this was an honest (if really, really bad) mistake as opposed to being deliberately misleading.

As Kelly put it:

@charles_gaba @xpostfactoid1 I never thought it was on purpose - rather, the kind of error we more easily miss if it cuts in our direction.

— Ken Kelly (@_KJKelly) November 24, 2014

What I do know is this:

  • Until the 7.3 million August figure was released, I had been estimating that the number of currently enrolled QHPs was likely around 7 million even, and I said exactly that at the time:

When the "correct" number of 7.3 million was given out by HHS/CMS, my response was pleasant surprise. I was quite happy to be "mistaken" at the time.

  • Then, on November 6th, I posted an exclusive entry regarding the current QHP enrollment figure and stated that I was pretty sure that regardless of the August number being 7.3 million (or so I thought) it had dropped significantly, 2 months later, to as low as 6.8 million:

Once again, when the "correct" October number of 7.1 million was given out by HHS/CMS, I was quite happy to be "proven wrong":

This, of course, is also why I got so angry with Investor's Business Daily for accusing me of "false outrage" the other day. Not only was I touting the correct numbers at the time in both cases, but some of their own data came from me in the first place.

In summary:​

  • Going forward, should I assume that my own estimates are accurate if they conflict with the Health & Human Services Department of the United States?

I realize how boastful this makes me sound, but...yes, I think I'll have to start assuming that.

UPDATE 11/23/14 10:25pm: Or, again, as Ken Kelly just put it:

@charles_gaba @xpostfactoid1 It seems to me that you made the mistake of thinking you made a mistake once before. Don't do that again!

— Ken Kelly (@_KJKelly) November 24, 2014

Oh, and it should also be reiterated: Regardless of how this plays out, in the end, at least 11 million people now have healthcare coverage who didn't have it previously, which is a good thing.

And with that, unless Issa's #DentalGate investigation manages to come up with more actual evidence of deliberate misconduct than the GOP did with Benghazi (which is to say, none whatsoever), that'll be the last I have to say on the subject.

Meanwhile, I'm pretty sure that at least 212,000 people have selected a Qualified Health Plan (not including standalone dental plans!) for 2015 so far.