A Polite Response to a Somewhat Polite Critique

A lengthy comment in response to my "5M as of Sunday" post from yesterday criticizes the work done at ACASignups.net for a variety of reasons. For the most part these are the same issues which I've already addressed repeatedly, but he's more polite about it than the prior critic and some of his points are new (or at least I haven't really talked about them before), so I've decided to respond:

Let me start by saying that the work you have done is truly impressive but it is beyond comprehension that we can not get this type of granular info from the people who are actually responsible for designing, implementing and running this program. They are either completely incompetent or intentionally withholding the detailed data and simply releasing the more favorable headline numbers.

First, thanks for the compliment. Second, I'm not going to try and judge HHS/CMS here; this is an incredibly complicated process, and there may be any number of factors involved which could explain why they haven't released more granular data yet. I get irritated when people who don't know anything about website development just say "can't you just..." when asking for a feature which they think should be a piece of cake, but which is actually far more involved than that. I don't know the full scope of HHS/CMS's job here, so I'm not in a position to judge them on the payment data situation. Anything ecommerce-related can get pretty involved since there's additional security considerations and you're talking about hooking into dozens if not hundreds of different insurance company billing systems.

Having said that, I do wish that they would have added a little Yes/No checkbox to the HC.gov application process which simply asks "Are you currently insured?" That would have gone a long way towards addressing that issue. I assume it got lost in the shuffle of the other technical problems which had to be fixed last fall, but it really should have been added by now if it hasn't.

The fact that the gov't can't/won't provide the details, many people have used your data to support their arguments that the ACA is a success but it is clear that you are a cheerleader for this law and you present your data in an extremely biased manner to make the law appear more successful. Of course, it is your work and you can show it any way you like but I believe you should try and be more neutral and objective.

I'm not sure "cheerleader" is the correct description for me. I'm a single-payer guy, and I frankly hate the thought of being required to pay private, for-profit insurance companies. However, I also see the positives of the law as outweighing the negatives--it paves the path towards single-payer (at least at the state level), and it also stops the uglier abuses by those companies (pre-existing condition denials, cancelations of policies the moment a policyholder actually gets sick, massive overbilling, discrimination against women and so forth).

For example, the website is aca SIGNUPS but you refer to the numbers as enrollments. You are NOT ENROLLED until you pay so the number you keep quoting is gross signups, which is not the same as net signups or enrolled, thus overstating the true number. There is no reason why the results aren't available at least for signups thru Dec. 31st as they would have to have paid by mid to late January. Odd that Sebelius said they don't have that data since they don't collect the money and said ask the insurers, yet the insurers have stated that they routinely send the info to HHS since they can't get paid the subsidy without the documentation. Again, either Sebelius and HHS are incompetent, liars, or withholding unfavorable results. I read you comments on the paid/unpaid and I don't disagree with you (although I think you 10% is low-probably closer to 15%+) but even using your 10%, there will NOT be 6.2 million people enrolled but a NET number of 5.6 million or lower. Of course, the 6.2 million will be the one used for all the media releases until it is quietly revised downward in 2-3 months. Even 10% of 6.2 million would be a pretty significant reduction but it would be the correct number.

The reason it's called "ACASignups.net" is because "ACAFullyPaidEnrollmentsButOnlyIfTheyWerentAlreadyInsured.net" would have been a wee bit lengthy, don't you think?

Seriously, though, I don't overstate the number at all. Both The Spreadsheet and The Graph (the single most popular and shared page on the site) clearly and prominently note the "Paid Enrollments" number, and I even use that as the baseline number for Private QHPs. Yes, I believe that if HHS reports an official enrollment number today, that number is likely to be 5.5 million. HOWEVER, I also clearly and openly state that of that 5.5 million, only around 4.97 million will eventually count. I've stated my rationale for this several times, in great detail.

This is why I provide a range of numbers with specific breakdowns to the best of my ability, so that I can be as transparent and unbiased as possible. If you want to include unpaid enrollments, they're listed. If you don't want to include them, that number is listed as well, and so on.

We were told over and over that the law was needed to help the 48 million uninsured and therefore the most important number should be the number of NEWLY INSURED. THE ESTIMATES WERE FOR 7 MILLION ON THE EXCHANGES AND 9 MILLION ON MEDICAID. Taking your 6.2 million less 10% for non payers and generously assuming that 75% are newly covered, we end up with LESS THAN 4.2 million compared to the original 7 million that Sebelius claimed would indicate the program was a success. Getting 4.2 mil vs. a 7 mil goal is an EPIC FAIL. 

Actually, the CBO report which gave the original 7 million QHPs / 9 million Medicaid numbers in the first place includes an extra line item of negative 2 million as well, from the private insurance market, for a net increase of not 16 million (7 + 9) but only 14 million (7 + 9 - 2). The CBO assumed that around 2 million people would move from an existing individual market policy to either an exchange-based QHP, Medicaid or something else.

For that matter, due to population growth, people aging into Medicare age, employment-based insurance changes and so forth, the total uninsured population at the end of 2014, according to the same CBO report, was actually only projected to decrease from 55 million to 44 million in the first year, a net reduction of 11 million, not 16 or even 14. Since this includes Medicaid expansion (and since Medicaid expansion was denied in half the states, screwing over about 4.8 million people), I'd say that if anything, you could argue that the ACA overall may have even exceeded expectations already, depending on your perspective.

The sub 26 crowd is irrelevant and I could not find any reference to their inclusion in the CBO data as being part of the previously uninsured population. Many college students are on their parents plan so not sure why letting them remain so until 26 does anything to change the equation.

Really? Everyone under 26 is irrelevant? I'm sure they're happy to hear that. However, the reason they don't appear in the CBO data as being "previously uninsured" is because they've been insured (in an increasing number) since 2010, when the law actually went into effect, while the CBO report only starts in 2013. My website doesn't just track people who have gained insurance specifically via the exchanges, it includes everyone who has gained insurance due to provisions in the Affordable Care Act. This is also reflected to a lesser extent on the Medicaid spreadsheet, which includes 81,000 people (30K in DC and 51K in Minnesota) who were moved onto Medicaid back in the first half of 2012 thanks to pre-exchange provisions (see those entries on the Medicaid spreadsheet for details). As for "many students being on their parents plans", yes, that's true...and the total number, according to the 2013 tracking study by the Commonwealth Fund, is actually closer to 15 million, of which they estimate 7.8 million are specifically due to the ACAHowever, since the HHS Dept. is only using the much more conservative 3.1 million figure, I'm going with that...and in the interest of giving the full benefit of the doubt, I've even reduced the low-end estimate further, to 2.5 million, based on Washington Post reporter Glenn Kessler's speculation that an average of the quarterly shifts should be used instead of the individual quarters themselves.

I also question your decision to include the bulk transfers of state plans over to Medicaid. Your rationale was that they met the strict definition of ACA enabled but having states simply dump existing participants onto Medicaid doesn't lower the number of uninsured but simply allows the states to transfer the costs over to the federal gov't. 
The difference in churn rates from the baseline is pretty substantial (85% vs. 30%) and I could not find the data you used to calculate this assumption. Given the current economic environment, I find it impossible that 60% of the current Medicaid population drops off in the states that are expanding while only 15% of enrollees drop outin the non expansion states. That difference makes up a huge contribution to the new enrollee population. Since you used the numbers, I assume you can explain why there is such a large difference and didn't just plug in the number because that was the number that was calculated.

Regarding the "bulk transfers to Medicaid", you're correct that these do not represent an actual reduction in the number of uninsured people, but they do represent the lifting of a massive burden off the shoulders of the states, which don't have the same sort of resources that the federal government has. Now, you're correct that the "bulk transfers" on the Medicaid side are the equivalent of the "canceled private insurance policies" being shifted over to exchange (or off-exchange) QHPs. However, these people are still new to Medicaid, which is what I'm primarily measuring. Nowhere on my site do I state that every one of the people on either the spreadsheets or the graph were previously uninsured, and I've discussed that issue several times quite openly and candidly in the past.

As for the "churn rates", you're making a classic mistake when calculating percentages, because you're forgetting that in the case of non-expansion states, there is no expansion to include in the percentages.

Example: Let's say that there's two states: North Expansionia and South Nomedforuia. Both states have the exact same socioeconomic conditions and the exact same low-income uninsured population.

In the absence of any hard data, I'm estimating that the breakdown in each is as follows:

  • North Expansionia: 30,000 baseline churn, 20,000 woodworkers, 50,000 strict expansion = 100,000 new Medicaid enrollees (30% / 20% / 50%)
  • South Nomedforuia: 30,000 baseline churn, 5,300 woodworkers, 0 strict expansion = 35,300 new Medicaid enrollees (85% / 15% / 0%)

In both cases, you have identical baseline churn (which makes sense if everything else was identical pre-ACA), but that 30K figure represents only 30% of the total in the first case, but 85% of the total in the second case, because it's 85% of a much smaller total

Now, you're correct that I don't have much to base these percentages on (the 20% expansion state woodworker number is a more conservative version of Washington State's 33%; the 15% non-expansion state woodworker number is based on the South Carolina estimate of 16%). However, on the spreadsheet, I also don't include the CMS report numbers at all for a dozen states, including large ones like Texas and Michigan, because those states mix renewals together with new enrollees. Since I don't know how many renewals to subtract, I don't include the totals for those states at all, which is why Michigan shows zero at the moment (our own Medicaid expansion doesn't start until April 1st). There are probably several thousand "woodworkers" in Michigan, but I'm not listing any of them, because with renewals being in the mix, I'm being as conservative as possible and disregarding them.

My guess is that there's actually a couple hundred thousand more "woodworkers" but I'm being very, very conservative in my estimates here.

While I am obviously not a fan of the law and believe that you are biased in some of your assumptions that paint a more favorable picture, You are focused on highlighting the GROSS SIGN-UPS, which does nothing to measure the number of newly insured, which was the whole point of the law. The relevant number to measure success should be the 16 million (7 from QHE 9 from Medicaid) uninsured that were supposed to gain coverage from the law. Using that metric, the signups are a complete failure and you are ignoring this fact by highlighting numbers which obscure this fact. I am NOT questioning your ethics or integrity as you stated was the case with another poster who disagreed with you. I think you have done a terrific job shedding some light on the situation, although it is highly disturbing that we can't get this from CMS or HHS.

Again, the CBO never said that it would be 16 million NEWLY INSURED in the first year. At most, they said the newly insured would be 14 million (this was before their revised projection, which dropped the totals down to 14 million / 12 million), and the actual reduction in the uninsured in the first year was only projected to be 11 million. By that standard, I'm pretty sure we've already hit the mark, although that remains to be seen. 

However, you've chosen to completely ignore the single largest missing piece of the puzzle here, which I've discussed many, many times: OFF-EXCHANGE QHP ENROLLMENTS.

I've documented nearly 560,000 of these, and that's from only 2 states (Washington and Wisconsin), plus a half-dozen other companies...out of the 200+ companies throughout the country. Even the WA and WI data for off-exchange enrollments are still out of date, since they only include through the end of January. My guess is that there are easily 4 million or more of these out there (in Washington State, as of the end of January there were about 84,000 exchange-based QHPs and 184,000 OFF-exchange QHPs since October...or 69% of the total! In Wisconsin the numbers were closer to what you'd expect: about 50,000 exchange-based and another 7,900 off-exchange, or around 14% of the total).

So, you're absolutely correct that both unpaid enrollments and previously insured enrollments probably should be subtracted from the total.

HOWEVER, if you're going to do so, you also have to ADD the total number of off-exchange enrollments, SHOP enrollments and Medicaid/CHIP enrollments first.

If you can provide me with the total number of outstanding off-exchange QHP enrollments from the remaining 47 states (Vermont, along with DC, requires all enrollments to be done via their exchanges), then I'll be happy to subtract the 4.8 million canceled policies (which includes my own) from the combined total, and will even subtract the Medicaid "bulk transfers" from the total as well, to get a true picture of people now covered by ACA-compliant policies or Medicaid who were previously uninsured.

My guess is that this would look like the following:

  • Current: 15.6 Million
  • ADD at least 4 million more off-exchange enrollments
  • SUBTRACT about 4.8 million canceled policies
  • SUBTRACT about 1.5 million "bulk Medicaid transfers"
  • SUBTRACT about 550 thousand unpaid enrollments

= NET gain of around 12.75 million newly-insured people.

However, until you (or someone) can provide that missing off-exchange QHP data, I'm not going to do so. Instead, I clearly and transparently list all of the possible numbers involved, including fully-linked sources and my rationale explanations for each.

That's the best I can do for now. If that's not good enough for you, so be it, but don't try to claim that I'm hiding anything and then claim that you're not questioning my integrity.

Having said all of this, I do agree that HHS/CMS should really have been providing a lot of this information all along. On that we agree 100%.