NEW Feature Added: Potential Enrollment Pool Spreadsheet

For months now, I've been pointing out that while the CBO's original 7 million overall Private QHP projection seemed reasonable at the time, the CMS's state-by-state projections to achieve that 7 million goal never made any sense, and in some cases were flat-out ludicrious. States like New York and Kentucky were assigned the same enrollment target even though NY's population is 4.5x the size of KY's, and so on.

Even within individual states this made no sense--Vermont was expected to have 57,000 people enroll in QHPs even though they only have around 47,000 uninsured residents. Sure, there'll be some overlap due to the "5 million policies cancelled!!" debacle (ie, people switching from a non-ACA compliant plan to a compliant one, including my own family), but it still made little sense, and in fact the odds are very high that someone simply copied & pasted Utah's projection numbers onto Vermont, right below it.

Anyway, a couple of months ago a contributor going by the handle Brutus56 came up with a more logical formula for reasonable state-by-state targets:

  1. Find the total pool of uninsured people who qualify for enrolling in a Private QHP
  2. Divide 7 million into that number to find the fraction to multiply by for each state.
  3. Multiply that fraction by the uninsured residents per state to get a reasonable number that they should expect to get enrolled by March 31st.

At the time, I used the total number of uninsured people in the country, which I had down as roughly 48.3 million. I took 7 million / 48.3 million, which gave me around 0.145 or 14.5%.

I then took 14.5% of the uninsured residents of each state to give my "% of 7M Proportional to % Uninsured" targets.

For example, CMS had Alabama down with a target of 82,000 people. Alabama had about 666,000 total uninsured. 14.5% of that is about 96,600 people (AL's fair share of 7 million uninsured), so their target was set too low by CMS. Conversely, CMS had Kentucky down with a target of 220,000 people. Kentucky had a similar number of uninsured to AL, around 650,000 people. 14.5% of that is 94,300; CMS set KY's target far too high. And so on...

However, I forgot about several important points:

  • First, undocumented residents (or illegal immigrants, depending on your political leanings) can't be counted as part of the total pool, since they can't legally use either the Private QHP exchanges nor do they qualify for Medicaid or CHIP. So, you have to subtract them from the total.
  • Second, as much as it stinks that half the states didn't expand Medicaid, the fact remains that there's about 4.8 million people who don't qualify for either enrollment target. They don't qualify for Medicaid but don't make enough to qualify for QHP subsidies either, which effectively means they're kind of screwed. This sucks, and hopefully those states will all come around, but in the meantime, I have to subtract them from the Medicaid pool.
  • Finaly (and this is the biggest one), I didn't separate out the Medicaid/CHIP pool from the Private QHP pool! That is, just because 48 million people are uninsured doesn't mean that all 48 can sign up for one or the other; those who qualify for Medicaid don't get any subsidies on the exchanges (which pretty much means they can't afford private QHPs anyway), while those who qualify for QHP enrollments don't qualify for Medicaid by definition. In order to get a fair enrollment target for each, they have to be separated out.

To resolve all of these issues, I'm introducing a new, third spreadsheet: Potential Enrollment Pools!

Here's how I achieved these numbers, which rely almost exclusively on the Kaiser Family Foundation's excellent report on the impact of Medicaid NON-expansion in those states which haven't gone through with it:

  1. First, I've updated the estimated number of Total Uninsured, both State-by-State and Nationally; KFF has it around 700K lower than their earlier estimate, to around 47.6 million
  2. Then I took KFF's breakdown of the status of the total uninsured per state, separating out Undocumented Immigrants, those in the Medicaid Gap, those who qualify for Medicaid (either Expansion or Woodworkers) and those who can legally enroll via the ACA Exchanges for QHPs.
  3. KFF separated those who qualify for QHP Subsidies from those who don't; I combined these two (but did separate out Undocumented Immigrants)
  4. For the 13 states which didn't have uninsured Undocumented Immigrant numbers, I used a guestimate of around 60%, based on the average of the other 37 states.

Doing all of the above breaks the total out as follows:

  • Total Uninsured: 47.6 million
  • Uninsured Undocumented Immigrants: 6.4 million
  • Medicaid Gap in Non-Expansion States: 4.8 million
  • Total Uninsured Eligible for Medicaid: 14.0 million
  • Total Uninsured Eligible for Exchange QHPs: 22.5 million

OK, so there's about 14 million uninsured who could potentially be added to Medicaid, and about 22.5 million who could potentially enroll on the ACA Exchanges (with or without subsidies).

HOWEVER, obviously not all of these folks will do so in the first year, which is where the CBO's original 7 Million QHPs / 9 Million Medicaid numbers came in (later revised downward to 6 million / 8 million).

So, to achieve the original goal (reasonable state-by-state goals), I went back to my original formula, which has been corrected for the lower total pool:

  • 7 million / 22.5 million = around 31.1%
  • 6 million / 22.5 million = around 26.7%

Applying this to the total number of uninsured QHP-eligible residents per state gives me the new targets for each, depending on whether you go with the 7M or 6M totals.

As a final step, I then rounded off these numbers to the nearest 1,000 to simplify them on the actual QHP and Medicaid spreadsheets, still adding up to an even 7M / 6M.

To revisit the Alabama/Kentucky examples above, after removing undocumented immigrants, Medicaid Gap people (Alabama only) and so on, the newly corrected QHP targets for these states become 90,000 and 78,000 respectively.

And with that, I'm hereby demoting the absurd original CMS "targets" to the right edge of the spreadsheet where they'll hopefully be forgotten.