Surprise! ACA marketplace sabotage hit the most vulnerable hard

I fear I buried the lead somewhat in yesterday's look at how ACA enrollment losses were distributed among income groups in states that refused the ACA Medicaid expansion.

The distribution in the 18 nonexpansion states that use HealthCare.gov looked a lot like the distribution in all 39 HealthCare.gov states -- not surprising, since 69% of HealthCare.gov enrollees are concentrated in those states -- a fact kind of astonishing in itself. But enrollment is inflated in nonexpansion states because, thanks to a lucky ACA drafting error, eligibility for subsidies in those states begins at 100% of the Federal Poverty Level (FPL), whereas in expansion states it begins at 138% FPL. Those below that threshold in expansion states are eligible for Medicaid.

As I did note in the prior post, fully 36%* of enrollees in nonexpansion states, about 2.2 million as of the end of Open Enrollment, have incomes that would qualify them for Medicaid if their states accepted the expansion (as Virginia has for 2019, and as Maine has, though still blocked by Governor LePage's obstruction). Within that population, or rather, within the somewhat wider 100-150% FPL band broken out by CMS, enrollment dropped 6% in those states in 2018. That's a loss mainly among people who should be in Medicaid, and who are likely to be uninsured if they pass up marketplace enrollment, and for whom the marketplace is a relatively affordable deal -- premiums capped at 2% of income for CSR-enhanced plans with a 94% actuarial value. (Ironically, the Medicaid work requirements imposed by some expansion states will likely uninsure more people than the cuts in enrollment assistance and outreach in the marketplace.)

More broadly, the relatively modest top line of ACA on-exchange enrollment loss in 2018 -- about 4% -- shouldn't blind us to the fact that the loss was close to double that among people with incomes in the 100-200% FPL range (7.5% in HealthCare.gov as a whole). For that population, the core marketplace offering hasn't changed much, although fixed actuarial values mean a bit more out-of-pocket expense every year, and reduced competition may reduce quality of choice in many markets.

But let's not forget that the Trump administration assault on the enrollment infrastructure, coupled with toxic, misleading public pronouncements, took its toll on the most vulnerable marketplace population -- while sabotage more broadly, including the CSR funding cutoff and pending mandate penalty repeal took an even larger toll on the unsubsidized, who were fully exposed to huge premium increases driven largely by Republican sabotage. Those in the middle, with incomes in the 200-400% FPL, took advantage of the accidental compensation wrought by silver loading.  Those at the bottom, not so much.

Here again are enrollment changes by income in 2018.

Enrollment by income level, 2017 vs. 2018

HealthCare.gov 

Year	Total enrollment	100% to 150% FPL	150% to  200% FPL	200% to 250% FPL	250% to 300%  FPL	300%- 400%  FPL	Other FPL*
2017	9,201,805	3,208,242	2,050,555	1,312,520	752,403	786,678	1,091,407
2018	8,743,642	2,979,236	1,885,778	1,277,488	747,165	867,198	986,777
Change	-5.0%	-7%	-8%	-3%	-1%	+10%	-10%

* "Other FPL" is comprised mostly of unsubsidized enrollees. About 20% are likely enrollees with incomes under 100% FPL, most of whom are likely legally present noncitizens time-barred from Medicaid, who are subsidy-eligible.



The nonexpansion states show basically the same pattern - not surprising, perhaps, as they account for more than two thirds of HC.gov enrollment. Here's how enrollment shifted there:

Non-expansion states, excluding Idaho

Year	Total  enrollment	100-150% FPL	150%-200% FPL	>200% to ≤250% FPL	>250% to ≤300%  FPL	>300%- ≤400%  FPL	Other FPL
2017	6,320,036	2,750,502	1,281,503	792,247	433,904	435,790	626,090
2018	6,060,844	2,587,081	1,202,256	786,543	438,500	489,249	557,215
Change	-4%	-6%	-6%	-1%	+1%	+12%	-11%

----
* In HealthCare.gov states, 2,587,081 out of 6,060,884 enrollees had incomes in the 100-150% FPL range. In 2016, we learned that about 85% of enrollees in that somewhat larger category in nonexpansion states have incomes in the 100-138% FPL range, and so would be eligible for Medicaid if their states had accepted the expansion. It's possible that the enrollment drop would prove to be slightly lower at 100-138% FPL than at 138-150% FPL if we had those breakouts, since at 138-150% FPL a benchmark silver plan costs 3-4% of income rather than 2%. But with 85% of the cohort under 139% FPL, the difference is unlikely to be much.

Updated 8/15 8:40 with minor clarifications.