Tuesday, May 25, 2021

The ACA as it should have been and may yet be: beyond premium subsidy boosts

Subscribe to xpostfactoid

The Wall Street Journal's Stephanie Armour succinctly captures where Democrats seem to be at on the healthcare reform front:

Many progressive Democrats and President Biden are facing the political reality that far-reaching healthcare overhauls aren’t likely to succeed in the short term, which means their hopes may rest instead on building on recent Affordable Care Act changes and reducing prescription drug costs.

That is, make the premium subsidy boosts in the American Rescue Plan Act (ARPA) permanent, and maybe maybe maybe pay for it by getting some tangible form of prescription drug cost control past Democratic moderates. 

Pretty plainly left behind is the centerpiece of candidate Biden's healthcare plan, which also happens to be the continually submerged Ur-template for Democratic healthcare reform:

Giving Americans a new choice, a public health insurance option like Medicare. If your insurance company isn’t doing right by you, you should have another, better choice. Whether you’re covered through your employer, buying your insurance on your own, or going without coverage altogether, Biden will give you the choice to purchase a public health insurance option like Medicare.

A public option truly "like Medicare" and offered on these terms would have the capacity to transform American healthcare by hitting three marks: 1) paying a modified version of Medicare rates to providers, 2) requiring providers that accept Medicare to accept it, and 3) being made available on an affordable basis to anyone who wants it, including those with access to employer-sponsored plans. 

Point three would also either transform or kill off employer-sponsored healthcare, which would have to reduce payment rates to providers in order to compete. For that reason, healthcare providers train massive firepower on any and all public option proposals, even those that remain confined to the relatively small individual and small group markets.  

The odds are vanishingly small that a 50-50 Senate and a 223-213 Democratic majority in the House will pass a strong public option of this sort. A weaker version that negotiates rates with providers market-by-(consolidated)-market and doesn't compel their participation is of dubious value, as Matt Friedman argues in detail. Perhaps its chief value would be as a vehicle for future transformation, when the fiscal pressures of subsidizing commercial plans in the marketplace become too much to bear, and/or when Democrats win larger majorities, if they ever do (and if we continue to have free and fair elections, which seems dubious).

That said, the subsidy increase provided through 2022 by ARPA, which Democrats all but have to make permanent, constitute a major boost to affordability and would bring the ACA much closer to fulfilling the aspiration inherent in its name. According to KFF estimates, subsidized coverage (Medicaid or marketplace) is now available for 63% of the estimated 28.9 million nonelderly uninsured. Those ineligible for subsidized coverage include the undocumented (3.9 million), people in the "coverage gap" created when the Supreme Court made the ACA Medicaid expansion optional for states (2.2 million), those with an offer of insurance from an employer (3.5 million) and those for whom an unsubsidized benchmark silver plan costs less than 8.5% of income (1.1 million)  

Those who are eligible for marketplace subsidies are likelier under ARPA's new subsidy schedule to find the subsidized coverage affordable -- if they find their way to a clear view of what's on offer. Within the narrow confines of the current ACA structure, Democrats can plausibly do much -- legislatively and administratively -- to reduce both uninsurance and underinsurance. Steps include:

1. Enrollment outreach. Ignorance of ACA offerings -- among the general population and the uninsured in particular -- remains astonishingly high, as veteran enrollment assisters attest. In 2020, less than a third of people surveyed by KFF knew that the ACA was still law. The Trump administration deliberately fostered such ignorance, cutting federal funding for enrollment assistance by nonprofit "navigators" by 84%, from a peak of $63 million in 2016 to $10 million by 2018, and advertising by 90%. The Biden administration has allocated $80 million for navigators in the 36 states using HealthCare.gov for 2022 (the funds are drawn from user fees charged to participating insurers, so the 15 states that run their own exchanges have their owning funding base for enrollment assistance). According to KFF, Trump administration underspending of the user fee revenue has left some $1.2 billion available to the Biden administration to boost enrollment efforts. 

The Biden administration opened an emergency Special Enrollment Period for the marketplace on February 15, running until August 15. After ARPA was enacted, boosting subsidies, the administration earmarked $50 million to advertise the insurance offerings. 

Meanwhile, the administration has proved to be pretty good at the logistics of getting tens of millions of people vaccinated. Question: has there been any effort to mount ACA outreach at vaccination sites? At a minimum, leaflets, texts and emails could be aimed at people getting shots. Appointments with enrollment assisters could also be on offer.

2. Enrollment streamlining. By KFF's estimate, 7.3 million uninsured people are eligible for Medicaid, and 6 million for zero-premium marketplace coverage. Maryland has pioneered an enrollment jump-start tied to tax filing, whereby the uninsured whose reported income and insurance status indicate they are eligible for subsidized coverage can check a box on their tax return and receive information about their likely eligibility for "free or low cost coverage." On a national level, aligning the annual Open Enrollment period with tax filing season and porting information on the tax return to a marketplace application could streamline the enrollment process. Cynthia Cox of KFF sees potential in enlisting tax preparers to encourage/assist enrollment. Integrating enrollment with tax preparation could also take some of the diceyness out of the income estimate that determines subsidy size.

3. Fix the family glitch and redefine "affordable" employer-sponsored insurance. ACA marketplace subsidies are not available to people who have access to an "affordable" offer of insurance from an employer -- defined as minimum essential coverage that costs less than 9.83% of income in 2021 (if you wonder why this formula varies by a few tenths of a percent annually, search under "idiotic technocratic mindset"). Unfortunately, under the statute as interpreted by the IRS, "affordable" means "affordable for the individual employee." If the employer offers a family plan that costs the employee more than 9.83% of income, but the individual plans is below that threshold, the whole family is subsidy-ineligible. That shuts millions of people out of the marketplace. Some of them remain uninsured (or the adults do; the children usually are eligible for CHIP); more pay well above 10% of income for family (or two-person) coverage. [Update, 5/30/21: There is some intimation that the IRS may lower the affordability threshold for an employer-sponsored plan to 8.5% of income, the new affordability cap for marketplace plans. But that hasn't happened yet.]

This "glitch" could and should be fixed by statute. If not, it can be fixed, by administrative action, according to health law experts including Timothy Jost. If Democrats take legislative action, they should also reduce the affordability standard for employer-sponsored plans to 8.5%, the in-marketplace affordability threshold for a benchmark silver plan. Alternatively, they might make marketplace subsidies available to anyone for whom benchmark silver (or benchmark gold, if we get there) costs less than a plan of equal value offered by the employer (though that flexible standard would add administrative complexity).

4. Reduce out-of-pocket cost exposure.  For ACA marketplace enrollees with incomes up to 200% of the Federal Poverty Level ($25,520 for an individual, $52,400 for a family of four), silver plans come with Cost Sharing Reduction (CSR) subsidies that reduce out-of-pocket costs to a level below the average for employer-sponsored plans - - though still high for many low income people. Above the 200% FPL threshold, benchmark silver coverage exposes enrollees to very high out-of-pocket costs. In 2021, silver plan deductibles average $3,385 for enrollees with incomes between 200% and 250% FPL (where weak CSR is available) and $4,816 for enrollees with incomes above 250% FPL (where no CSR is available). 

Candidate Biden proposed raising the value of a benchmark plan, against which income-adjusted subsidies are set, from a silver level (70% actuarial value, or AV) to gold (80% AV). The average gold plan deductible is $1,648 (and as in silver plans, many plans offer many services not subject to the deductible).  A bill introduced in April by Senator Jeanne Shaheen of New Hampshire would raise benchmark AV at every income level.

If Congress does not act to improve actuarial value in the marketplace, the Biden administration can do it administratively, in at least two ways: 

A. Mandate strict silver loading. In October 2017, Trump cut off direct reimbursement to insurers for  CSR (the ACA statute requires such direct reimbursement, but Republican Congresses refused to appropriate the funds). State regulators responded by allowing or requiring insurers to price CSR directly into the premiums for silver plans alone, since CSR attaches only to silver plans. That boosted premium subsidies, which are set against a silver benchmark, and created discounts in bronze and gold plans. Bronze and gold plans remain overpriced relative to silver plans, however. Silver plans (confusingly) provide platinum-level coverage at incomes up to 200% FPL, and most enrollees are below that income threshold. Gold plans should therefore be cheaper than silver, but they rarely are. CMS could require insurers to price plans in strict proportion to their actuarial value, as the ACA in fact requires by statute, and so make gold coverage available at below benchmark cost. More details here.

B. Change the formula by which AV is calculated. As I've noted before, retailing a proposal championed by Gabriel McGlamery of Florida Blue:

AV allegedly reflects the percentage of the average enrollee's annual medical costs that will be paid by the insurer. At present, the small minority of enrollees who rack up tens or hundreds of thousands of dollars in medical costs skew the averages, as Bill Gates walking into a bar skews the average income of those present. Thus a plan that by AV formula covers 70% of "the average enrollee's costs" may have a $5,000 deductible and require 40% coinsurance for most services. For 99% of enrollees, that plan doesn't come close to covering "70% of costs."

The current formula used to determine AV does exclude enrollees whose one-year costs top $1 million. The proposed action would cut far deeper, probably by using the "risk scores" that insurers assign to each enrollee (predicting the enrollee's medical costs compared to those of an average enrollee).  McGlamery estimates that cutting enrollees with a risk score above 20 out of the AV calculation (removing only about 1% of enrollees) would reduce the AV of a silver plan currently pegged at slightly above 70% to about 62%, and the AV of a bronze plan currently pegged at slightly above 60% down to about 50%.

Cutting the most (potentially) expensive enrollees out of the AV calculation would reduce out-of-pocket costs at every metal level.

5. Fight auto-enrollment inertia.  Beginning in 2018, silver loading made free bronze plans available to millions more enrollees than previously. Discounted bronze plans (free or not) induced millions of enrollees to switch out of silver*.  One ill side effect was a steady increase in bronze enrollment at incomes below 200% FPL, where strong CSR usually makes silver a much better value. Now, ARPA has made silver free at incomes up to 150% FPL and available for no more than 2% of income up to 200% FPL. Every effort should be made to induce current bronze plan enrollees  for whom silver plans are free or near-free to trade up, perhaps going as far as via autoenrollment, with an opt-out. In 2021, about a third of marketplace enrollees were auto-enrolled, meaning they were prior-year enrollees who did not log into an ACA marketplace and update their financial/personal information or actively choose a new plan but did not disenroll. As David Anderson has pointed out, passive autoenrollment will be a major problem if lawmakers change the benchmark plan to gold and autoenrollees remain in silver, as CSR will presumably attach to gold plans.

*          *          *

None of these actions would eliminate the ACA's administrative clunkiness, its characteristically narrow provider networks, the money wasted on commercial provider payment rates, or holes in coverage. But taken together they would reduce the uninsured population by millions, and the underinsured by millions more. For a host of further administrative actions that should be considered, see this roundup by Frederick Isasi and Stan Dorn of Families USA.

P.S. Commenter Norm Spier is absolutely right below: ending Medicaid Estate Recovery for over-55 Medicaid enrollees who are not receiving long-term care services should be on this to-do list.


* Silver enrollment for enrollees in the 200-400% FPL income range in HealthCare.gov states plummeted from 60% in 2017, the last year before silver loading began, to 23% in 2021. Below 200% FPL, the value of CSR, provided to enrollees at no cost, outstrips the value of bronze plan discounts, and the shift to bronze was more moderate, though still substantial, accelerating in 2021. Details here.


  1. Thanks for the valuable contribution--pointing out some of the remaining flaws in the ACA that need to be fixed. (This is the main focus of my own involvement in the public discourse--pointing out those defects so that they can be fixed.)

    A few notes:

    1) I don't know if your readers are aware, the extension of the removal of the 400% of Federal Poverty Level "subsidy cliff" (the most critical component of your "premium subsidy boosts in the American Rescue Plan Act (ARPA)") which are for two years only, and which you wish to be made permanent, would in fact be made permanent in the full proposed Biden "American Families Plan" (https://www.whitehouse.gov/briefing-room/statements-releases/2021/04/28/fact-sheet-the-american-families-plan/ .) Who knows, of course, what parts, if any, of this, will make it into law.

    2)In a recent Health Affairs "Following the ACA" blog post (https://www.healthaffairs.org/do/10.1377/hblog20210520.564880/full/ ), sharp lawyer Katie Keith has a pretty detailed look at aspects of the "family glitch", including how it came about, and a possible way to substantially fix it by Presidential executive action only.

    3)You missed a thing needing fixing that I know you know about, and of course which I comment about often in places, including xpostfactoid comments.

    That is, the Medicaid estate recovery that is done on ACA expanded Medicaid and other ordinary medical (non-LTSS) Medicaids for people 55-64 in at least 12 Medicaid expansion states, some of them blue, like NJ, MA, and MD.

    One of the things about this estate recovery is that states can recover either a capitation or all medical bills paid out. In the latter case, it's very bizarre and cruel -- the person has no insurance at all--just a loan for medical expenses. (As far as I can tell, this demanding of the money back for medical expenses is not done to anyone in the universal coverage systems in the rest of the developed world.)

    It's also a bizarre inequity of treatment under the ACA in those states that do the recovery. Above 138% of Federal Poverty Level (FPL), people get highly subsidized insurance--in fact currently free through 150% of the FPL. But below 138% FPL, in at least 12 states, people 55-64 get just a loan payable by their estate.

    Further, since it's an obscure ACA detail, with inconspicuous notification on many state exchanges, and no notice on the Federal ACA application, it's an accident waiting to happen to people with temporarily low incomes and with substantial assets to lose (since expanded Medicaid eligibility has no asset cap, and eligibility for it blocks eligibility for subsidized on-exchange coverage).

    For those unfamiliar, I supply the usual references:


    (which I wrote most of in 2019, after finding out about the issue. Note the assertions there can all be verified from the references, which are all online. Also, here is the backup in case someone at Wikipedia monkeys with the article:

    https://web.archive.org/web/20200701011813/https://en.wikipedia.org/wiki/Medicaid_Estate_Recovery_Program )

    And of course, you do know about the issue, and in fact, you discovered it the same year I did -- 2019 -- and wrote this:


    and more recently called again for the crux of the correction, by prohibiting Federally estate recovery on expanded Medicaid, here:

    https://xpostfactoid.blogspot.com/2021/01/the-117th-congress-should-end-medicaid.html and both of those xpostfactoid posts are other good references.

  2. Thanks for the excellent catalog. As always, I have a few quibbles:

    1. Passing a prescription price control proposal does not pay for diddly doo.
    This is classic Congressional timidity and deception: to propose a cut in Medicare so that we can spend more on the ACA.

    This was done a lot when the ACA passed, and it never worked. The Medicare "cuts" proposed in 2009 either never occurred or were balanced by extra provider benefits quietly added elsewhere.

    If you want to pay for something. raise taxes. If the public won't abide a raise in taxes, keep discussing the program until they do. Go bernie Sanders!

    2. I am glad you mentioned the user fees that could pay for more navigators.
    I suspect they could pay for many hundreds of navigators. Based on my reading, the ACA exchange user fees are enormous. They are about 3% on at least $50 billion of premiums I think.

    3. Also glad that you mentioned changing the AV formulas. This is long overdue.

    Keep up the great work!

    1. Bob, there's good evidence that the ACA's Medicare spending restraints -- on hospital cost growth, and Medicare Advantage reimbursement -- did work. Medicare cost growth has been pretty restrained since 2010, with costs growing much slower than in commercial insurance.

  3. Good comment. I am just being a public-choice grouch, I guess. I favor the German approach -- when the time came to pay for the integration of East Germany, they raised taxes. When they decided to pay publicly for long term care, they added about 2% to the Social Security payroll tax. It was all visible, on-budget.

    Whereas in the USA we have Medicare and Medicaid dipping deep into general revenue each year and helping drive the national deficit -- but we say that if we just reduce the annual growth, we can pay for something else.

    The devices used to pay for Romney's Massachusetts reforms in the 1990's are complex and ridiculous in some ways -- but they did help many people obtain coverage. The Minnesota Care program which I love is funded in similar convoluted and by-German-standards dishonest ways. I guess I have to accept that.

  4. Fabulous post.

    One thought: yes it would be helpful to cut the high-dollar claims out of the AV calculation.

    But better still would be developing an entirely new metric to help consumers determine the relative value of plans.

    Remember, the AV metric was developed by insurance companies, for their own internal use: they need to know AV to make sure they are collecting enough premium to stay solvent. That figure, however is meaningless to consumers, since as you point out the vast majority of patients have to pay something approaching 100% of their non-preventative expenses out of pocket, since most never hit their deductible. What the shopping consumers need to know is their likely all-in cost, that is, premiums plus out-of-pocket.

    The best way to do that is to scrap AV entirely in all consumer-facing documents, and instead develop a purpose-built, consumer-relevant metric that tries to illustrate likely all-in costs. My thought would be to follow the college admissions and college guidebook metrics whereby they no longer provide AVERAGE SAT/ACT scores and GPAs of successful applicants; they instead provide the interquartile range.

    That way the applicant can at a glance see where the middle fifty percent of successful applicants came from, as well as see at a glance the cut-offs for the 25th percentile (below which they are less likely to get in) and the 75th percentile (above which they are likely to get in).

    My office in a regulatory comment a couple of years ago provided CMS with detail around what such a new metric could look like and do. See Part II:

    Medium version:


    Original filed comment: