Note: Free xpostfactoid subscription is available on Substack alone, though I will continue to cross-post on this site. If you're not subscribed, please visit xpostfactoid on Substack and sign up
 |
| Raise that subsidy cap... |
It may be a fool’s errand to write about reported features of a now-stayed Trump administration plan to extend the enhanced ACA premium subsidies for two years, with various conditions and haircuts imposed.
Since House hardliners have stayed the plan, it’s likely to get worse if it ever sees light of day. But as outlined in the various media reports, the pending plan is full of poison pills — or, shall we say, enrollment-inhibiting side effects. The features already floated raise a question we may well be confronted with: If Republicans do coalesce round a plan something like what’s been floated, would Democrats (most, some, a handful) get on board? Should they?
Let’s consider the constraints imposed on the enhanced subsidy schedule in the Trump outline.
Minimum premium. In 2025, some 8 million ACA marketplace enrollees — about one third of the total — paid no premium. Enrollees with income up to 150% FPL ($23,475/year for an individual in 2026) pay no premium for benchmark silver, and millions more at higher incomes access zero-premium bronze plans (and sometimes gold plans, in states where gold plans are available at premiums below that of benchmark silver). Imposing minimum premiums is purportedly aimed at quelling broker fraud, which metastasized in 2023-2024, partly enabled by the advent of universal zero-premium eligibility at low incomes.
Imposing minimum premiums is not necessary for quelling this fraud, however. We know this because of a natural experiment: the fraud was concentrated in the 30-33 states (varying by year) using the federal exchange, HealthCare.gov. That’s partly because low-income enrollees are concentrated in those states (in 2025, 84% of enrollees in the 100-150% FPL income bracket are in HealthCare.gov states). But it’s mainly because in HealthCare.gov states, brokers had easy, unfettered access to the accounts of all enrollees, via commercial e-broker platforms, needing only a name, location and date of birth to act on an existing account, and only slightly more information, gained through unscrupulous ads and pitches, to initiate an account. The state-based marketplaces require various forms of two-factor authorization, with verifiable input from the client, before a broker can act — and again, there is no evidence in those states of significant enrollment fraud.
At the same time, through other natural experiments, we know that imposing even the smallest of premiums on low-income enrollees significantly cuts enrollment (this appears to be more a matter of administrative friction than ability to pay). In one notable study, Harvard’s Adrianna McIntyre and co-authors examined the effect when the effect when a cohort of low-income enrollees paying zero premium for Massachusetts ConnectorCare (an ACA precursor and approved alternative to the marketplace) in 2016 were charged premiums under $10/month in 2017. About 14% dropped coverage. Surveying that study and others, Matthew Fiedler of the Brookings Institute estimates that imposing minimum premiums in the marketplace would cut enrollment by just shy of 1 million (about 12%).
Codifying marketplace “integrity” rule. This purported provision is reported in a single sentence in the Washington Times (a conservative outlet with presumably good access to the Trump administration). The so-called 2025 Marketplace Integrity and Affordability Rule, finalized in June 2025, threw up a thicket of administrative barriers to enrollment in the name of fighting fraud (in the same spirit as imposing minimum premiums, which could not be done without legislation). In August, a court stayed several of the most damaging of these provisions. Here I’ll focus on two (quoted from the CMS summary linked to above):
Income Verification When Data Sources Indicate Household Income Less than 100% of the FPL
CMS is finalizing the requirement that Marketplaces generate annual income inconsistencies in certain circumstances when a tax filer’s attested projected annual household income would qualify the taxpayer as an applicable taxpayer according to 26 CFR 1.36B-2(b), while the income data returned by the Internal Revenue Service reports that the tax filer’s income is less than 100% of the FPL.
This rule is aimed at marketplace applicants in danger of falling into the “coverage gap” that opened up when the Supreme Court in 2012 made the ACA Medicaid expansion optional for states. The ACA stipulates that a) adults with income below 138% FPL are eligible for Medicaid and b) that enrollees with income in the 100-400% FPL range who lack access to other affordable coverage (e.g., from an employer) are eligible for marketplace subsidies. (The mismatch in thresholds — 138% FPL for Medicaid, 100% FPL for the marketplace — is due to a drafting error that proved fortunate.) In the 10 remaining states that have refused to enact the ACA Medicaid expansion, 5.2 million enrollees in 2025 reported income in the 100-138% FPL range. The “integrity” rule aims to shut a good proportion of them out of the marketplace — those who can’t verify that their income is on pace to exceed 100% FPL ($15,650 for an individual, $26,650 for a family of three).
Start with the fact that the ACA drafters did not intend to leave people with income below the poverty line without access to affordable coverage (Medicaid). The coverage gap is a product of the recalcitrance, cruelty and stupidity of right-wing state legislators and governors. Continue with the fact that ACA subsidies are determined by a projection of future income - - and income for low-income people is decidedly variable and uncertain, often dependent on variable hours in multiple jobs, cash tips, self-employment, and seasonal work. Much of it is also presumably not reported to the IRS (tips, small payments for services from the self-employed). At present, low-income enrollees can attest to their next-year income without hard documentary evidence. The exchanges can ask for documentation, but they can’t end subsidy eligibility if said documentation (e.g., a spreadsheet of anticipated income from self-employed services) doesn’t match with “trusted sources,” beginning with the IRS. Codifying the currently stayed rule would require the exchanges to cut off subsidies to those who can’t verify income over the 100% FPL. According to CMS, “This policy will improve program integrity, reduce the burden of APTC on the federal taxpayer, and benefit consumers by ensuring subsidies are appropriately allocated and reducing their risk of improper tax liabilities.”
That is, this rule would kill coverage for some portion of the 5.2 million enrollees* in Florida, Texas, Georgia, and seven other nonexpansion states who estimated income in the 100-138% FPL range. Some of those who do turn out to earn more than 100% FPL would likely by stymied by the verification requirements, as documenting income can be tricky if it doesn’t come from one fixed-salary source. CMS estimates (p. 27200) that just 81,000 people would lose coverage as a result of this provision, based on results when the first Trump administration implemented this rule before a court stayed it, saving $191 million. That estimate seems low, especially as CMS elsewhere in the rule explicitly credits Brian Blase’s exaggerated estimates of fraudulent enrollment, which lean heavily on claims that millions of enrollees have income below 100% FPL. (Blase estimates that 6.4 million enrollees “fraudulently” estimated their income between 100% and 150% FPL, though some are alleged to have under-estimated to get into that range.)
Another potentially high-impact “integrity” rule provision would penalize auto-re-enrollment, again targeting lower-income enrollees specifically:
Requiring $5 Premium Responsibility
CMS is finalizing modifications to the annual eligibility redetermination process by requiring Marketplaces on the Federal platform to ensure that consumers who are automatically re-enrolled with no premium responsibility following application of APTC and without affirming or updating their eligibility information, are automatically re-enrolled with a $5 monthly premium beginning in plan year 2026. Once consumers confirm or update their information, the $5 monthly bill will be eliminated if they continue to be eligible for a $0 premium after application of APTC.
This rule would have an effect similar to imposing minimum premiums on those eligible for no-premium coverage. Most people with health insurance auto-re-enroll annually, including people in employer-sponsored plans, Medicare Advantage, and Medicare Part D. Imposing premiums, even short-term, would doubtless cause some people unaccustomed to paying premiums to drop coverage.
Funding CSR. Oh Lordy, here were go again… Cost Sharing Reduction (CSR) subsidies increase the actuarial value (AV) of a silver plan to a roughly platinum level for enrollees with income up to 200% FPL. (65% of all enrollees in 2025 are below that threshold.) The ACA statute stipulates that the federal government reimburse insurers separately for the value of CSR — that is, for raising the AV of a silver plan from a baseline of 70% to 94%, 87%, or 73%, depending on income. The statute did not make the funding for this mandatory, though, and Congress never appropriated the funds. The Obama administration funded CSR anyway; the Republican Congress sued; a court upheld the suit but stayed the order pending appeal; and Trump abruptly cut the payments off in October 2017. Since that move had been anticipated since his election, states responded by allowing insurers to price the value of CSR directly into premiums — in most states, into silver premiums only, since since CSR is available only with silver plans.
Since income-adjusted premium subsidies are set to a silver benchmark, this “silver loading” inflated subsidies along with silver premiums, creating net-of-subsidy discounts for plans priced below the benchmark (second cheapest silver plan). As a result, several million more people became eligible for zero-premium bronze plans in 2018 than previously. Further, since CSR raises the AV of a silver plan well above a gold plan’s 80% AV for most enrollees, gold plans became cheaper than silver plans in many states (some states took action to enforce the full pricing of CSR into silver plans). In 2019, silver loading boosted enrollment by an estimated 5%. At present, nationally, lowest-cost gold plans are on average priced below benchmark silver plans in 20 states, and nationally, lowest-cost gold is on average priced slightly below benchmark. In 2025, 13% of enrollment was in gold plans, up from about 3% in the pre-silver loading years. In Texas, where by statute gold plans are priced way below silver, 35% of enrollment in 2025 was in gold plans, as was 56% of enrollment at incomes above 200% FPL (where CSR is negligible (to 250% FPL) or not available (at incomes over 250% FPL).
Funding CSR would save the federal government tens of billions of dollars and eliminate these discounts. If the enhanced ARPA subsidies were made permanent without further infringements, some might see funding CSR as a reasonable pay-for. In concert with the other provisions floated by Trump, it would be one more substantial cut at enrollment. While silver loading probably does not currently boost enrollment by the 5% estimated in the pre-ARPA era (since the enhanced subsidies make coverage more affordable at all incomes), a 2% impact would mean about 500,000 fewer enrolled.
New subsidy cliff. The enhanced subsidies enacted in the American Rescue Plan Act (the ARPA subsidies) removed the infamous income cap on subsidy eligibility, instead setting the percentage of income required for the benchmark (second cheapest) silver plan at 8.5% of income for any applicant with household income above 300% FPL. The Trump plan reportedly restores the income cap on subsidy eligibility but raises it to 700% FPL. That’s considerably below the 935% FPL cap set in the Problem Solvers’ bill, the effects of which I looked at in detail in my previous post. In short: while 700% FPL sounds pretty affluent, a few hundred thousand people who need to buy their own insurance would be burned pretty badly — particularly at higher ages, as unsubsidized premiums rise with age. To adapt an example from my previous post:
The average unsubsidized benchmark premium for a pair of 62 year-olds in the marketplace in 2026 is $33,694/year, or $2,808/month. Lowest-cost bronze for this couple would average $24,612/year, or $2,051/month, with a per-person deductible north of $7,000.
As presently subsidized, if this pair of 62 year-olds have an annual income of 700% FPL ($148,050), they would pay $1,049/month for benchmark silver and just $292/month for the lowest-cost bronze plan. The full-year subsidy costs the federal government $21,109. Under the Trump plan, if just over the cliff, they would pay 23% of income for the benchmark silver plan or 17% of income for the lowest-cost bronze plan.
In 2025, 912,000 enrollees reported a household income over 500% FPL. That’s 3.7% of all marketplace enrollees. (Another 913,000 did not report income, most of them presumably recognizing that they did not qualify for subsidies.) Of those, how many would be newly subjected to the very steep new subsidy cliff? At a guess, 1-2% of current enrollees, or some hundreds of thousands. Based on JCT estimates of the cost of extending the ARPA subsidies at various income levels, imposing the 700% FPL income cap might save the Treasury about $2 billion per year.
If Republicans were to coalesce around an ARPA subsidy extension bill with these provisions (and no other changes to marketplace structure), the enhanced subsidy extension would come laced with a lot of ill side effects. Taken together, the provisions discussed here might cut enrollment by almost as much as expiration of the subsidies is forecast to in the first year — 3.8 million, by CBO’s December 2024 (in the absence of other factors such as the partially-stayed integrity rule). In fact, though, ending the ARPA subsidy enhancements would likely result in far deeper enrollment reductions over time, more in the range of 12 million, according to a Wakely analysis (or a simple comparison with enrollment in the pre-ARPA era, which topped out at average monthly enrollment of 10.3 million in 2020, compared to 21.0 million in 2024).
So, if Republicans coalesced around a bill that extended the ARPA subsidies up to 700% FPL for two years, with the provisions described here, would Democrats support it? (A requirement to further limit abortion, wrecking the ACA’s delicate abortion compromise (insurers must charge a separate premium for abortion coverage, not funded by the federal government), which might be a total dealbreaker, might be a real poison pill.) If Republicans did put the shadow Trump plan up for a vote, the pressure to stave off expiration of the enhanced subsidies might redound on Democrats. Many might calculate that when Democrats regain the majority in one or both houses, they could roll back objectionable provisions — and further extend the subsidies. The enhanced subsidies, if extended on these terms, might prompt a high-pressure renewal fight every year or two or three, as adjustments to Medicare payment rates for doctors used to do.
Then again, Republicans may well converge on a more toxic package of changes, or fail to agree to any package. That would be bad for public welfare, but good for Democrats’ electoral prospects.
- - -
* All enrollment stats cited here are derived from CMS’s State-level ACA enrollment public use files, sometimes with my calculations (e.g., tallying enrollment in several income brackets).
Photo: Smithsonian, “Abraham Lincoln’s Top Hat”
No comments:
Post a Comment