What's up with the ACA's state-based marketplaces?

Enrollment in the 18 SBEs appears on track to decline by about 3% in 2023

Attendance is down in the SBM League

Update, 1/24: Please see the follow-on post that compares the SBMs’ pre-pandemic uninsured rates and marketplace takeup rates, along changes in unemployment in 2022 and other factors, compared to FFM states.

While enrollment in the ACA marketplace as a whole in the Open Enrollment Period for 2023 is on pace to finish about 13% higher than in OEP 2022, enrollment in the eighteen states that run state-based marketplaces (SBMs) is on course to come in about 3% below the OEP 2022 total.

As my last post emphasized, enrollment growth throughout the pandemic has been overwhelmingly concentrated in states that have refused to enact the ACA Medicaid expansion. In those states, about 40% of enrollees would be eligible for Medicaid if their states had enacted the expansion, and the American Rescue Plan Act made a benchmark silver plan free to almost all of them. As of this past week’s enrollment snapshot, enrollment in the twelve current nonexpansion states is up 23% year-over-year.

But enrollment in twenty-one expansion states that use the federal HealthCare.gov platform is also up by 10%. That throws the apparent enrollment decrease (barring last-minute surges or Week 9 reporting lags) in the SBMs into sharp relief and has marketplace watchers scratching their heads. (See the bottom of this post by Charles Gaba for breakouts of 2023 enrollment to date by marketplace type and expansions status.)

In fact, throughout the pandemic years, enrollment growth in the states currently running their own marketplaces has collectively lagged far behind growth in the overall market. Enrollment in these eighteen states is up 6% since OEP 2020, compared to 43% for the marketplace as a whole.

---

Six new SBMs have launched since 2020. Enrollment totals marked in red below denote that the state was using HealthCare.gov in that year.

Table 1
Enrollment in Current State-based Marketplaces, 2020-2023

Source: CMS Marketplace Open Enrollment Public Use Files
Totals marked in red denote that the state was using HealthCare.gov in that year.
See note at bottom regarding projections of end-of-OEP 2023 totals.

The slower growth in the SBMs in the pandemic years represents something of a reversal. From OEP 2017 to OEP 2020 — the first three years of the Trump administration, before the pandemic scrambled the calculus — enrollment in HealthCare.gov states dropped 9%, while enrollment in the 12 SBMs then in operation ticked up by 1%. (Those totals are as of the end of each OEP, which is not the only measure or the best, as discussed below.)

Many attributed the declining enrollment totals in HealthCare.gov states in part to the Trump administration’s gutting of funding for enrollment assistance (reduced by 84%,) and marketing (cut by 90%) in those states. The SBMs were insulated from these cuts, as enrollment assistance and outreach in the ACA marketplace is funded by a user fee assessed to insurers participating in the exchanges.

Progressives in HealthCare.gov states imprinted the lesson that states with SBMs controlled their own fate. As shown below, enrollment decreases during OEPs from 2017-2020 were particularly acute in five of the six states* that have launched their own marketplaces since 2020 (labeled “Future SBEs” below).

Table 2
Enrollment in SBEs and HealthCare.gov States, 2017-2020

Source: CMS Marketplace Open Enrollment Public Use Files
Nevada launched its SBE in time for OEP 2020 (total marked in green)

It’s tempting to point out that the six “new” SBMs taken together mostly recouped their losses from 2017-2020 in 2021-2023. In fact, though, enrollment patterns within those six states differ radically, as they do in the 50 states generally. Myriad factors affect enrollment swings, including swings in unemployment, the pandemic moratorium on Medicaid disenrollments, spending on outreach and enrollment, broker participation, insurer departures and new entries (which can affect premiums net of subsidy in unpredictable ways), newly added state “wraparound subsidies,” silver loading (a pricing practice that can produce steep discounts in bronze and/or gold plans), the varying quality of state websites, and state culture (though Republican-based hostility to the marketplace as such may finally be fading). Sustained enrollment drops in Massachusetts, Kentucky and Nevada call for, shall we say, local study.

All of which is to say, proceed with caution in venturing causal generalizations about enrollment patterns in broad categories of states. In fact, even the enrollment drops in HealthCare.gov states during the Trump years were something of a mirage. Retention, at the point of first premium payment and throughout the coverage year, improved during those years, to the extent that average monthly enrollment nationally was actually higher in 2020, when 11.4 million people selected plans in OEP, than in 2016, when 12.7 million did. Possible reasons: 1) the Trump administration shortened OEP and cut advertising and enrollment outreach, possibly culling more marginal enrollees likelier to drop coverage; 2) large premium increases in 2017 and 2018 reduced unsubsidized enrollment, and those faced with paying full freight may have been more likely to forgo enrolling; and 3) ironically, Trump’s October 2017 cutoff** of direct reimbursement of insurers for the value of Cost Sharing Reduction (CSR) subsidies, which attach only to silver plans, led to “silver loading,” — the pricing of CSR directly into silver plans, which created discounts in bronze and gold plans and dramatically raised the proportion of enrollees paying nothing or next to nothing for coverage.

Caveats notwithstanding, the negative growth in thirteen out of eighteen SBMs in a year of strong enrollment growth in almost all HealthCare.gov states does ask for some explanation. A few possible factors:

• The SBMs’ marketing advantage in the Trump years may have effectively been reversed as the Biden administration raised spending on enrollment assistance and outreach to new levels and rebuilt the navigator program for HealthCare.gov states.

• Most if not all SBM websites present information less clearly than HealthCare.gov.

• Commercial Direct Enrollment (DE) and Enhanced Direct Enrollment (EDE) platforms, widely used by brokers, can enroll people only in HealthCare.gov states. The main story here is HealthSherpa, which is faster and more user-friendly than all the government exchanges and is used by thousands of brokers. HealthSherpa processed more than 4 million applications by January 1, accounting for almost half of enrollment in HealthCare.gov states to that point. Many insurers also provide Enhanced Direct Enrollment on their websites. EDE — a program begun in the Obama administration but eagerly developed and promoted by the Trump administration — may be giving a boost to broker participation. And HealthSherpa, which enrollees can also access directly, may be smoothing enrollment for the low income enrollees in nonexpansion states who make up its core client base (43% of HealthSherpa’s enrollees are paying zero premium). (Thanks to Charles Gaba for fingering HealthSherpa as a possible factor.)

A few possible explanations that occurred to me don’t seem borne out by evidence:

• I might have thought that the SBMs had less room to grow because they’d enrolled more of the potential market in prior years. But the Kaiser Family Foundation’s estimates of the percentage of potentially subsidy-eligible people enrolled in each state as of 2020 don’t really bear that out, though more current SBM states were above the median takeup rate in that year than below it. [Update, 1/19: I may have been too dismissive on this core question. Uninsured rates as of 2021 are suggestive: They are 6.1% for SBMs, 7.5% for expansion states using HealthCare.gov, and 10.9% for nonexpansion states. While the higher uninsured rate in nonexpansion states is perennial, the difference in rates between SBMs and the expansion states on HealthCare.gov is suggestive. As for KFF’s estimates of takeup rates in the marketplace: nonexpansion states have a natural advantage on this front because eligibility starts at 100% FPL compared to 138% FPL in the expansion states, and benchmark silver coverage with strong CSR is free at incomes up to 150% FPL. Those penetration rates bear a closer look, which I plan to do in a followup post.][Update 2, 1/24: please see the follow-on post.]

• The pandemic’s moratorium on Medicaid disenrollments, which has been in effect since March 2020. Normally there is a lot of churn out of Medicaid, and some of those disenrolled end up in the marketplace, particularly in expansion states, which have enrolled some 20 million nondisabled adults in Medicaid via the expansion. But expansion states using HealthCare.gov did show strong enrollment growth this year.

• Full employment shrinks the pool of people who need ACA marketplace coverage and are eligible for subsidies. But unemployment, low nationwide, is not particularly low in SBE states.

OEP ends tomorrow in most states, though it’s open through Jan. 31 in four SBMs, (CA, DC, NJ, NY) and through Jan. 23 in Massachusetts. Final tallies, and breakouts of enrollment by income and other factors published in the spring, may provide more clues as to what’s going on in the SBMs. States have good reason to move to SBMs, which enable or facilitate state wraparound subsidies, standardized plans, public options and other potential innovations. SBEs are scheduled to come online next year in Oregon and Virginia. But the data indicates that the federal exchange may also offer significant advantages.

Update: I neglected to note that Charles Gaba has pointed out that if you add enrollment in New York and Minnesota’s Basic Health Programs into the mix, the SBM enrollment reduction year-over-year shrinks to 0.7%. The BHPs are Medicaid-like programs that serve enrollees with income up to 200% FPL; in New York, the coverage is free. New York’s BHP, the Essential Plan, has more than a million enrollees, and all the enrollment growth is there; enrollment in MinnesotaCare is flat. The BHPs are a different animal from the marketplaces; coverage is standardized, out-of-pocket costs are much lower, and in New York, again, there is no premium.

- - - - -
Note to Table 1: Projected final enrollment totals for each state in 2023 are derived by increasing the Week 9 total by the percentage increase in that state that occurred in 2022 from Week 9 to the end of OEP. For the projected all-state total, I added 3% to the current total, in line with the 2022 increase from Week 10 to end of OEP.

* Kentucky, under the leadership of Democratic Governor Steve Beshear, began the ACA years with an acclaimed SBM, Kynect. Beshear’s Republican successor, Matt Bevin, took Kynect offline in advance of OEP 2017. Bevin’s successor, Democrat Andy Beshear, relaunched Kynect as an integrated benefits application site in time for OEP 2022.

* *When Trump abruptly cut off direct CSR reimbursement he boasted that he’d killed the ACA, despite longstanding forecasts ( see CMS 2015,, CBO 2017) that the resulting silver loading would actually boost enrollment, which it appears to have done. Whether Trump thought that stiffing insurers for the last three months of 2017 (silver loading began in 2018) would drive them out of the marketplace, or was completely ignorant of the likely effects of silver loading (though HHS leaders could not have been), or actually was okay with bolstering markets while assuring his followers he’d destroyed them, or never consciously sorted any of these possibilities, who knows? But the cutoff of direct CSR payments did strengthen the marketplace, as did the Trump administration’s encouragement of Enhanced Direct Enrollment development and receptivity to state “innovation waiver” proposals establishing reinsurance programs with partial federal funding. Even the effective repeal of the federal individual mandate (requiring those with access to affordable insurance to obtain it or pay a penalty) may have ultimately bolstered the marketplace by defusing a major source of hostility to it. On the other hand, the Trump administration did batter the marketplace by reducing the OEP period, gutting funding for enrollment assistance and marketing, and relentless encouragement of lightly regulated, ACA-noncompliant insurance markets.

“Oops” note: the terms “state-based marketplace” and “state-based exchange” are used interchangeably. The Trump administration favored “exchange,” while the Obama and Biden administrations favored/favor “marketplace” — e.g., in the Public Use Files — but not consistently. I toggled heedlessly between the two in this post and have now corrected to SBM throughout — except in the tables, which I daren’t swap out for fear of formatting troubles.

Photo by Digital Buggu