David Anderson

The Affordable Care Act’s individual market is likely to have significant legislative policy changes this year.  However, if there is not a fundamental restructuring of how the subsidy system works, weird things are likely to happen that are counter-intuitive.   We need to understand how price-linked subsidies work and who those subsidies apply to.

The ACA subsidy system is based on filling a gap.  The gap is the difference between the benchmark gross premium and an individual household’s applicable percentage of income.  The federal premium tax credit (PTC) serves as a gap filler.  The PTC is designed so that two households that are identical in all aspects (age composition, number of household members, and family income as a function of countable federal poverty level (FPL)) pay the same amount out of pocket for their benchmark plan even if there are massive differences in the gross premium of the benchmark plan.

For instance, if a household has an applicable percentage of 5% of income, they are expected to pay 5% of their income for the benchmark plan.  This is true if the benchmark plan has a premium of 5.1% of their income (.1% household income PTC) and it is true if the benchmark plan is 15% of income (10% household income PTC).  It is not true if the benchmark plan is 4.9% of the household income as there are no negative PTCs.

Furthermore, the ACA subsidy system is set up so that the net premium that a household pays is flexible.  If the household chooses a plan that is priced below benchmark, the household pockets the difference between the benchmark premium and their chosen plan’s premium in the form of a reduced out of pocket net premium.  Households are frequently exposed to zero premium plans.  Conversely, a household that chooses a plan priced above the benchmark pays the entire extra incremental premium for the more expensive plan.

Finally, under current law, not everyone gets a PTC subsidy.  Only households who earn between 100% to 400% FPL and who meet several other criteria are eligible for PTC subsidies.  Most other people are able to buy plans but not get any financial assistance from the PTC.

So what does this mean?

It means that there are two general classes of buyers in the ACA individual market as it is currently constituted.  The first class is the subsidized buyers.  They are fundamentally indifferent to premium levels.  The federal government eats the premium level risk.   Instead buyers who earn between 100% and 400% FPL are sensitive to premium spreads.  How far away is their chosen plan from the benchmark is what determines what they pay.  A subsidized buyer whose preferred plan costs $1400 a month could be seeing a great deal if the local benchmark is $1900 while they could be seeing an extraordinarily expensive deal if the local benchmark is $900.  As a further subnote, individuals who earn under 200% FPL (for a single individual ~$25,500) these folks are hyper sensitive to the spread between the cheapest silver plan and the benchmark (what I’ve called the Silver Gap before).  Spread sensitive folks get no direct benefit from premium reductions as those benefits are collected by the Federal Treasury in the form of lower PTCs.

However, folks who are not subsidy eligible don’t care at all about premium spreads.  They don’t care about the difference and distance in pricing between the plan that they want and any other plan.

Folks who earn over 400% FPL (~$51,000+ year for a single individual) only care about the premium level.  They take on the full burden of premium increases.  They get the full benefit of premium decreases.

Why does this all matter?

There are going to be a lot of proposals designed or at least proclaimed to be targeting affordability.  As long as we have price linked subsidies that apply to a large proportion of the ACA’s covered populations, we will need to analyze almost all policies through the lens of Spread buyers and Level buyers.  Using this lens, the introduction of a premium reducing public option is mostly beneficial to buyers who only care about premium levels  and could be slightly to significantly detrimental to price-sensitive spread buyers.  Program design would matter a lot.

The ACA is complicated and incomplete in its promise. There are ways to get closer to its core promise that everyone can get affordable health insurance (with significant fights over the meaning of “affordable” and “health” and “insurance” and “health insurance”) but in this legislative window of potential policy change, we will need to grapple with the complexities and oddities that the price linked subsidy system that is at the base of the current affordability mechanism imposes on current program design.  That could mean trimming sails, it could mean making explicit trade-offs, it could mean systemic reformation of the affordability mechanism.  But whatever path that is chosen, it must be grappled with.

Price linked subsidies and who cares about spreads or levelsPost + Comments (15)

West Virginia is a state whose population is disproportionally likely to benefit from the Affordable Care Act’s subsidized health insurance markets.  It has a large population of people who need health insurance and who earn between 100% and 400% Federal Poverty Level (FPL).  It has light to non-existent competition on Healthcare.gov so the stable equilibrium may not be converged pricing in every county which compresses the premiums spread.

However since 2017, the total share of Healthcare.gov enrollment that West Virginia supplies has gone down.

For the 2021 OEP, national enrollment increased.  It increased on Healthcare.gov once we adjusted for two more states with 7% of the 2020 enrollment leaving the platform and going totheir own exchanges.  However, West Virginia lost enrollment again.

There are two things going on, I think.

First, we have good evidence that partisanship matters.  Republican leaning counties (and states) tend to have less enrollment at any given price point than a comparable Democratic leaning area.  The elite messaging that conservative leaning partisans have heard is that the ACA is horrendous and should be avoided. That (probably) matters.

This obstacle can be overcome through superior pricing.

Here, West Virginia has made a policy choice.  The state government has decided that they will not structure their market to provide increased affordability for subsidized buyers.  Since 2018, Cost Sharing Reduction (CSR) subsidies are not paid directly by the federal government.  Instead, the benefit must be provided and insurers can seek compensation through indirect means via the premium channel. The overwhelming majority (48) states currently have insurers “Silver load” CSR benefits onto the silver premiums.  This increases the silver premiums relative to gold and bronze plans and produces larger than 2017 status quo counterfactual subsidies.  That means a lot of people are either exposed to zero premium bronze plans OR see gold plans with significantly less cost sharing priced near silver plans.  The value proposition in silver load states for people who don’t want to buy silver plans just got a whole lot better and people who bought silver plans in previous years sometimes move to non-silver plans.

Three states, West Virginia, Indiana and Mississippi, mandate a different CSR compensation scheme.  These states require insurers to spread the CSR costs to all plans in a system known as Broad loading.  Broad loading results in minimal changes in relative premium spreads compared to 2017 policy.  Few people see a better deal now than they did under the Obama administration in these states.

The combination of partisan response and the deliberate embrace of broad loading’s inability to provide a different set of pricing facts and reality to counter-act that partisan response in West Virginia has resulted in a shrinking ACA market in the Mountaineer State.

Policy choices and ACA enrollment in West VirginiaPost + Comments (9)

Yesterday, it leaked that the Biden Administration is inclined to re-open Healthcare.gov. In some ways, Healthcare.gov has been partially open since last April with a broad Special Enrollment Period (SEP), but there is a good argument that a full Open Enrollment Period(OEP) which does not require administrative burden to prove worthiness combined with elite messaging and validation as well as significant advertising will insure more people.

However, one of the challenges of an SEP and a partial year OEP is that the value proposition of the insurance gets worse as we get deeper into the year. ACA insurance contracts are designed on a twelve month cycle. The actuarial value of a plan is for the entire year and the entire pool. Shortening the length of the contract means that the actuarial value (the share that the insurer pays out of general premiums) goes down.

I want to take four utilization patterns and illustrate what is happening when a contract starts in January compared to when it starts in March.  For this purpose, I am assuming a single policy with a combined $3000 deductible/Maximum out of Pocket with no other cost-sharing for simplicity sakes.

• Low Use — barely touches the system, a PCP visit, an urgent care visit and a prescription
• Medium Chronic Condition — routine expenses of prescription drugs and specialist visits but no acute emergencies
• Acute Emergency — BABY!!!  or APPENDICITIS and otherwise looks a lot like a Low Use individual
• Expensive Chronic Condition — My niece Claire with her leukemia treatment is going to be consistently running up high monthly costs so her family is shopping on the combination of network and out of pocket maximums as they hit the MOOP in the first month no matter what.

The low use individual is indifferent to when their contract starts. They are unlikely to hit MOOP at any point in any scenario.  The Medium Chronic Condition person hits their deductible in July under as 12 month contract and in September under a 10 month contract. More importantly, the 12 month contract has the insurer paying for 47% of their claims out of premium revenue while the 10 month contract has the insurer paying only 31% of their annual claims from premium revenue.  The effective AV dropped significantly.  The individual with an acute emergency with a 12 month contract hits their deductible in the first month of a 12 month contract but never hits it under a 10 month contract.  The person with the expensive chronic condition hits the deductible as soon as the policy goes live in either scenario.

These are stylized examples (for example, an acute emergency that happens in July would be fully MOOP-ed in either scenario).

However, the intuition that for a given benefit configuration (deductible, co-insurance, co-pay, maximum out of pocket), the shorter the contract length, the lower the effective actuarial value being offered stands. As plans start later in the year, the effective actuarial value goes down.  A Gold plan transitions into a Silver plan in the middle of the year, and then a nominal Gold plan becomes, in reality, Bronze as hot chocolate weather arrives.  Bronze becomes Tin and Copper as well.  The speed of metal decay will vary depending on cost sharing designs (deductible heavy designs decay faster than co-pay heavy designs) but it is a simple unidirectional reality.

This matters.  The monthly premium for an ACA plan is the same if the plan starts in January or if it starts in March or if it starts in November.  The value proposition for an ACA plan gets worse as the start month gets deeper into the year. The actuarial value declines while the monthly premium stays constant.  This is a problem for future nerds to figure out a solution and then advocates to get that solution implemented via either legislation or regulation, but this is something that we need to be aware of as a limitation on the size of impact an unexpected mid-year OEP could do.

Partial year contracts and declining effective actuarial valuePost + Comments (7)