You are here
Medicaid/CHIP Compromise: Now a Range from 2.40M - 7.60M
OK, after discussing the Medicaid/CHIP situation with Caroline Pearson of Avalere Health some more (see her more detailed response to our discussion below), I've concluded there's just too damned much uncertainty and too many variables on the Medicaid side of things to try and pin it down to a solid number. This is why Avalere's own estimates ranged greatly from 1.1 million to 1.8 million people. In my case, both the low and high ranges are higher because I'm including more types and more recent data than they are.
I was thinking of 4 different types of Medicaid/CHIP enrollments: "ACA Expansion Only", "Out of the Woodwork", "Bulk Transfers" and "Renewals/Redeterminations". I've prided myself on successfully eliminating the fourth category, which has no business being included under any definition.
However, as Ms. Pearson points out, I forgot about a fifth category: "Baseline Churn"...that is, people who just happened to become eligible for Medicaid without expansion after October 1st and did enroll. These are not "woodworkers" since they weren't previously eligible; they just happened to fall on hard times after the exchanges launched. Avalere has concluded that the bare-bones, absolute low-end number remaining may be as low as 1.1 million people through the end of December.
On the flip side, Ms. Pearson does admit that Avalere didn't include three other important numbers in their analysis either:
- The 985,000 people who were automatically transferred over to Medicaid from an existing state-run program of some sort on January 1st, specifically due to provisions in the ACA;
- The 224,000 extra individuals in AK, CT, NV, NC and OR who were only listed as households;
- Some additional scattered data which has come in from January, which isn't included by Avalere at all since their study only ran through the end of December.
In other words, I did include a couple million "Baseline Churn" enrollees who I shouldn't have, while Avalere didn't include about 1.3 million people who they should have.
All of this being said, the Graph (and Spreadsheet) for Medicaid now shows a range:
- 2.40 million: Absolute bare minimum using the strictest definition of "Enrolled in Medicaid/CHIP due to the Affordable Care Act". Doesn't include the 25 non-expansion states at all, nor does this include the "woodworkers" or "normal churn" enrollees for the other 25 states. Includes Bulk Transfers, Household Individuals and another 90K to cover post-12/31 additions.
- 5.50 million: Doesn't include the 25 non-expansion states, but does include "woodworkers" and "normal churn" for the remaining 25 states.
- 7.60 million: Most generous definition. Includes all 50 states (+DC), including "churn" & "woodworkers".
Using this modified "range" system, the Grand Total for both Private and Public enrollments is now listed as anywhere between 8.85 Million - 14.06 Million.
Ironically, I'm doing this just before the January HHS report comes out, which will most likely cause me to have to recalibrate the whole thing yet again to keep up. Sigh.
Ms. Pearson's response, published with her permission (emphasis mine):
I appreciate your continued engagement on this important issue. I have had several days of client meetings out of the office, so I haven’t been able to get back to you sooner.
The way that I think about the Medicaid numbers is that they include 4 groups of people:
- New eligibles in expansion states
- Woodwork – current eligibles who join as a result of ACA
- Redeterminations – only in some states
- “Baseline churn” – These are the people who regularly come on and off of the Medicaid rolls absent ACA. They account for the fact that total annual enrollment (people enrolled at any time of the year) is much higher than average monthly enrollment in Medicaid.
Our analysis sought estimate the number of beneficiaries in 1 and 2. We believe this most directly approximates the CBO estimates of the number of people who will newly enroll in Medicaid as a result of ACA (both woodwork and new eligibles).
In Medicaid, it’s hard to extrapolate one states’ experience to another. So, it is difficult to know whether WA’s percentages would be representative in other states. However, more importantly, I think the WA numbers conflate #2 and #4 above. They are separating out redeterminations, but any new enrollees who are previously eligible are lumped into a single bucket. Whereas, we believe that a significant number of those new eligibles likely would have enrolled in Medicaid without ACA. We felt it was important to compare new application/determination rates to baseline application rates, and I believe this is the most important difference in our methods.
Since the July-Sept baseline was all that is available as a 50-state comparator, that is what we had to use. You correctly point out that there may be seasonal variation in application rates, which could impact our results. I am not aware of any 50-state data that would enable us to adjust seasonally, and I hypothesize that the seasonal variation may be inconsistent across states based on characteristics of the labor market (e.g., agricultural) and the program administration. However, any seasonal variation certainly could impact the results one way or the other.
On your other notes, the households point is a very good one, and I will certainly look into adjusting that in the future. I am going to explore using Current Population Survey data to make state-specific adjustments on household size among Medicaid eligible populations in the relevant states. Our current methodology would underestimate enrollment as a result of that issue. On the 984k and new updates, we chose to consistently rely on the HHS data and are not using any state-specific reporting post-January 1. Certainly, we expect that enrollment will accelerate in Q1, so all states will presumably gain significant enrollment once we have January data. Like you, we are tracking state reporting real time and can see the numbers increasing as more states report.
Hope this helps clarify. We look forward to refining the numbers as the data becomes more complete and consistent across states.