Anderson On Health Insurance

Senators Cassidy (R-LA) and Crapo (R-ID) released their health plan proposal for the ACA.

It does not extend premium subsidies but it does provide limited cost-sharing subsidies that are likely to be valuable for healthy and higher income enrollees.

That is the basis of their plan:

 

There is a lot going on here.

1.  Direct payment of HSA funds is great for pretty healthy folks who can afford full priced premiums
2. The decision to fund Cost Sharing Reduction subsidies in 2027 would end the practice of silverloading.
   1. Silverloading inflates Silver premiums
   2. Silverloading is great for enrollees with incomes between 200% to 400% Federal Poverty Level (FPL)  
   3. Silverloading has next to no impact on 100-175% FPLers nor over 400% FPLers
   4. Silverloading and Premium Alignment have juiced the Texas and Florida ACA markets
3. Catastrophic plans have a price advantage on average over Bronze plans for unsubsidized buyers only because risk adjustment does not move Catastrophic funds to cover Silver, Gold and Platinum plan enrollee expenses while Bronze premiums partially cover other enrollees’ expenses
4. Catastrophic plans for everyone worsens both the Metal Level and Catastrophic risk pools
   1. Great for remaining subsidized metal level buyers as higher morbidity is great in generating bigger premium spreads
   2. Really bad for non-subsidized Catastrophic and metal level buyers
5. There is a world where the ACA individual market acts as a well subsidized high cost risk pool — this is not that world.

 

The short version is that this plan does nothing for premium affordablity, and for those who actually need care a $1000 HSA contribution instead of extending the enhanced premium tax credits is still a money loser.  Getting rid of Silverloading increases premiums paid by middle class Americans for Bronze and Gold plans.

There are CHOICES BEING MADE HERE.

Cassidy and Crapo — great for healthy over 400% FPLersPost + Comments (21)

A friend of mine asked me to check out Jackson County Illinois for its ACA pricing.  There is a single insurer in the state.  Illinois requires the Silverload for Cost Sharing Reduction subsidies to be heavily loaded onto Silver plans. Silver plans are important because the second least expensive Silver plan determines the amount of subsidies an individual is eligible for.  Non-subsidized buyers care about premium levels while subsidized buyers care about premium spreads from the benchmark.  The bigger the spread, the more affordable the plan is.

And this is fascinating as I pull up the the pricing for a non-subsidized 45 year old.  I am pulling up the cheapest Bronze plan, and then the two cheapest Silver plans.

 

There are a couple of things to notice.

First, the premium spread between cheapest Bronze and the benchmark silver at this age is $595.  If Illinois did not have a mandatory non-Hyde abortion benefit requirement that can not be paid for by federal subsidies, then anyone with incomes under 399% Federal Poverty Level at this age or older will qualify for a zero premium Bronze plan.

The raw Silver Spread between the benchmark Silver plan and the cheapest Silver plan is $49.  Under current law, no one who wants to buy a Cost Sharing Reduction Silver plan could be exposed to a zero dollar Silver plan if Illinois allowed for those plans to be offered.

This is weird especially as Blue Cross and Blue Shield of Illinois, the only insurer in the county, offers a more expensive Silver plan as the 3rd Silver.  They could have dropped the current benchmark from the offerings and made a bigger spread without any work OR they could have engaged in substantial strategic pricing decisions to make sure that the largest population of ACA enrollees that they are guaranteed to get will be exposed to a very low but non-zero plan.

If we are to assume that sicker low income enrollees will pay premiums and healthier ones won’t and we assume that BCBS-IL gets to determine the extensive margin by changing the least expensive plan’s net of subsidy costs, BCBS-IL is leaving money and enrollment on the table.

This is just WEIRD.

Pricing strategies in the ACAPost + Comments (4)

The GOP — just in time again — is trying to come up with a health policy.  Right now there are a lot of flavors of direct to consumer subsidies instead of intermediating through insurance companies.  There are a lot of flavors as Dr. Adrianna McIntyre identifies in this skeet:

 

We also need to know *which* subsidies are in play. As I understand it:

Paragon has proposed converting CSRs to HSA contributions

Cassidy has proposed converting ePTCs to HSA contributions

Trump has proposed (I think? this is least clear) converting *all* subsidies to HSA contributions.

[image or embed]

— Adrianna McIntyre (@adrianna.bsky.social) November 19, 2025 at 10:50 AM

The two things that I’m pinging on is distribution  and incentives.

Lump sum distributions are great for pretty healthy folks.  Most years I am a light user of healthcare services with a one or two PCP appointments, an urgent care visit because of a bum ankle or a bad sinus infection that won’t clear on its own and zero to two generic antibiotic prescriptions at $5 a piece from Target or Walmart.  If I was to get a $1000 Health Savings Account distribution I’m golden in my normal years as my out of pocket spend would be net zero.  This year I’m a low to middle spender as I’m getting a mental health tune-up in therapy.  A $1,000 lump sum distribution would reduce my out of pocket spend in half.

There are people in my life who their good, low spend year is a $20,000 year and their high spend year is a $50,000 to $100,000 year.  A $1000 contribution does nothing for them. Their insurance limits their catastrophic cost exposure.

First dollar aid at the trade-off of lifting last dollar caps is great for 50% to 70% of the low using population.  It is hideous for the portion of the population that actually incurs substantial medical costs.

Secondly, incentives matter.  Right now, I have a deductible.  I am currently paying fully out of pocket for therapy.  I am happy to do so as it is helping me out. I am unlikely to hit my deductible cap in either 2025 or 2026 at the current utilization pattern.   But cutting the cost of therapy substantially by effectively reducing my deductible means I am likely to max out my deductible in 2026 which all of a sudden substantially reduces the cost of a few likely deferrable things to near zero.  I might consumer more fairly low value medical care as a function of timing.

Is that what we really want?

 

 

 

FSA and distributional consequencesPost + Comments (41)

On September 4th, the Centers for Medicare & Medicaid Services (CMS) released a new policy statement for the ACA individual marketplaces.

The U.S. Department of Health and Human Services (HHS) announced today it is implementing important measures to expand access to more affordable catastrophic health coverage through HHS’ new hardship exemption guidance. This guidance streamlines access to more affordable catastrophic coverage for consumers who are ineligible for advance payments of the premium tax credit (APTC) or cost-sharing reductions (CSRs).

Through these efforts, more Americans will be able to qualify for catastrophic health coverage based on need, beginning November 1^st with the start of open enrollment. Catastrophic plans generally have lower monthly premiums, are designed to protect consumers from very high medical costs in the event of serious illness or injury, and are required to cover three primary care visits pre-deductible. Consumers under the age of 30 have always been eligible for catastrophic plans through HealthCare.gov.

Coleman Drake, Dylan Nagy and I were curious.  Would this policy have bite?  Would it impact many people and would its impact vary by income and subsidy system in place?

We wrote a rapid research letter published at Health Affairs Scholar attempting to answer these questions.  But first let’s lay out some background.

Catastrophic plans have a simple benefit design — they have the legally allowable maximum out of pocket limit and the deductible is equal to the maximum out of pocket limit.  Preventive care services and three primary care visits are provided without cost-sharing.  Structurally they are equal to some of the lowest actuarial value Bronze plans in benefits and networks.

There are two massive differences between Catastrophic and Bronze plans.  First, Catastrophic plans are NOT  eligible for premium tax credits.  Buyers pay full price no matter their income.  Bronze plan buyers can receive premium tax credits that may reduce premiums down to zero.

Secondly, Catastrophic plans are risk adjusted only against themselves.  Catastrophic plans are very high cost sharing plans bought by highly likely to be healthy and young enrollees with decent to high incomes right now.  Money flows between Catastrophic plans to equalize risk but no money goes from Catastrophic to the Metal Level Pool.  Bronze plans have a substantial portion of the total premium going out the door to cover some claims in Silver, Gold and Platinum plans.  This risk adjustment quirk means that the Catastrophic plan offered by the same insurer on the same network as their lowest cost Bronze plan is typically cheaper in gross premium than a similar benefit structured Bronze plan.

So does this matter?

Well — it depends.  We estimated net premium paid by enrollees in 2025 for the lowest cost Bronze and lowest cost Catastrophic plan in a given county-year on Healthcare.gov in the approximately 1500 counties where both are offered.  We simulated 2026 subsidies under both a current law (2026 subsidies are back to the standard schedule) and current policy (enhanced subsidies continue) scenario.

What did we find?

 

Under an enhanced subsidy scenario, this policy does not offer any enhanced affordability through at least 500% Federal Poverty Level (FPL) for a single 45 year old.  Bronze plans, due to premium tax credits cost less for a similar benefit.

Under the current law, return to standard subsidies scenario, Catastrophic plans are modestly cheaper by about $100/month in premium than Bronze plans at and above 400% FPL.

However this is only a modest amount of help as someone at 401% FPL who had a low cost Bronze plan could plausibly experience an $800/month premium shock while the Catastrophic plan replacement might “only” be a $700/month premium shock.  Sure it helps on the margin, but it is marginal.

The current CMS leadership is prioritizing affordability for the non-subsidized portion of the market.  This is a legitimate policy choice.  Expanding Catastrophic eligibility will modestly benefit the 400% or higher group.

Now pulling this back out to the intersection of actuarial projections, policy and federal budgets.  Actuaries were terrified of this proposal as it came in late and it came in unexpected.  There is likely some movement between Bronze to Catastrophic next year, but no one knows who, how many and how sick.  We can think that the most likely movers are likely healthier than average Bronze enrollees, and we can also sort of expect that these movers are more expensive than the marginal current Catastrophic buyer so the actuaries don’t know how to predict the risk pools of both the Catastrophic pool and the Metal Level pools.  If there is a lot of movement, both pools are likely underpriced.  If there is no movement, both pools are appropriately priced given current projections.  There is a lot of uncertainty.  Some insurers have pulled their catastrophic plans from the market as they don’t want to eat surprise risk.

From a policy perspective, if there is a lot of movement from Bronze to Catastrophic, the Metal Level market is highly destabilized as risk adjustment is going to be, using a technical term, bonkers.  There won’t be enough low risk folks kicking into the pool to cover higher risk folks.  Assuming a decent amount of movement, insurers in 2027 will jack up their rates, which means any federal cost savings in 2026 likely will be consumed by increased subsidies in 2027.

Catastrophic Plan Affordability in the ACAPost + Comments (1)

In Health Affairs Scholar, my co-authors and I analyzed the Herfindal Hirschman Index (HHI), a measure of concentration,  of enrollment in the on-Exchange ACA marketplaces from 2014-2023.

Our big contribution was breaking down HHI by the number of insurers in a given county-year.  HHI is calculated by taking the market share of each insurer, squaring it, and then summing these products together.  We then multiply by 10,000 for ease of interpretation.  The Federal Trade Commission (FTC) gets worried when a market HHI is above 1800, and stressed out when HHI is above 2500.  They sort of freak out when it is above 5,000.  A single insurer county by definition has an HHI of 10,000.

 

So what does this mean?

Of course, by definition, markets are very highly concentrated when there are only one or two insurers in a given county-year.  But there is a lot of concentration when measured by on-Exchange enrollment in county-years where the theoretical minimum HHI is fairly low.  There are highly concentrated county-years when there are seven, eight or nine insurers in a given county-year.

Increasing the number of insurers in a market may or may not do anything to market concentration.  We see some county-years where there are 4 insurers with higher market concentration as measured by enrollment marketshare than counties with a pair of insurers.

Why does this matter?

Well, we should expect market exit.  This provides support some there are different types of exit.  An insurer with 1% market share leaving the market won’t have substantial consequences if the market goes from 3 to 2 insurers.  However we could expect much strong consequences of an insurers with 40% market share leaving a market and the market goes from 8 to 7.  There is a lot of nuance here that we need to think about.  Adding a completely non-competitive insurer to a county likely won’t improve quality or lower gross premiums because they are not a competitive threat to anyone.  We really need to wrap our heads around the impacts of what types of insurers enter and leave markets to better project the consequences.

 

Market concentration in the ACAPost + Comments (12)

The Affordable Care Act’s enhanced Premium Tax Credits (PTC) are due to expire on 12/31/25.  The standard PTC are currently law for 2026.  Individuals earning over 400% Federal Poverty Level are ineligible for standard PTC.

Insurers are sending out the annual renewal notices.  Insurers attempt to calculate the gross and net after PTC premiums individuals will pay if they passively and automatically re-enroll.

We are going to see millions of stories like the one below:

Me and my partner’s “Bronze” health insurance plan with Florida Blue is increasing from $511.22 a month to $1,711.75 a month next year. Not quite sure how to fit “$20,541 a year in health insurance premiums” into the budget.

[image or embed]

— Jason Steele (@filmcow.bsky.social) October 20, 2025 at 10:29 PM

I know nothing about this family. My guess due to their 2025 net premium is that they make someplace between 400% and 600% FPL.  And they are getting hit with a $1,200 premium shock for a low cost plan.

And yes, they are eligible for a catastrophic plan (but given recently accepted research) but that might only lower the premium shock to $900 or $1,100 a month. They can’t really downgrade the quality of their insurance much more.

They have three options.  Find other, likely underwritten coverage, go uninsured or find a massive percentage of their family budget for this bill shock.

Millions of households  are having that conversation this week.

Premium Shock Notification WeekPost + Comments (49)

I think too much about risk adjustment. I think too much about insurance pricing with the basic formula that premiums are roughly equal to claims costs plus overhead plus variance costs plus profit. And then I saw this Skeet yesterday. And my head hurts.

I am a well educated, pretty media literate person whose job it is to be professionally annoying. I cannot figure out if my cancer medication, which costs $17k a month (I do not pay anything thanks to a combo of insurance/access programs), is going to be tariffed 100%.

— Lizzie O’Leary (@lizzieohreally.bsky.social) September 29, 2025 at 9:51 AM

I immediately (as I do) thought about risk adjustment in the ACA. ACA risk adjustment is intended to make insurers compete on actually being good instead of being good at dodging bad risk. It transfers money through a complex formula from insurers in a given state-year that have a population that has low predicted predictable costs to insurers that cover a population that codes as having high predicted predictable costs. Yeah, say that three times fast. It is imperfect, but it is pretty decent.

Pricing for ACA risk adjustment is, in normal policy years, tough.  A cluster of diagnosis codes are assigned a risk adjustment factor by the federal government.  This risk adjustment factor is the average incremental cost that cluster of codes generates in prior year data with potential modest adjustments (hi Hep-C drugs).  The actuaries then have to use these factors to price plans.

Plan pricing is hard.  Actuaries have to guess first who their client/employer will be covering and what do these folks look like. This projection is based on both aspects the insurer controls like their own premiums and their network and benefits but also other insurers’ premiums and benefits. Then the actuaries have to guess how many other people will enroll in a rating area or the state and what does their health status look like.  From these sets of compounding guesses, the actuaries need to project how much sicker or healthier their clients’ membership looks relative to the state average and then multiply that difference by the state average premium.  This will produce an approximate risk adjustment flow either into the plan or out of the plan.  This flow is critical to setting premiums.

Yeah, there is a lot of guessing here.  This is TOUGH!

And now tariffs are a huge new factor.  Actuaries now have to guess what percentage of their risk-adjustable conditions are exposed to what degree of tariffs.  Conditions which have no tariff exposure will be accurately priced.  Conditions which are 100% exposed to 100% tariffs are grossly mispriced.  Lots of conditions are likely in-between.  And there is likely within condition variance as Physician Group A likes to prescribe fully tariffed drugs while Physician Group B typically prescribes lightly tariffed drugs.  This is an omni-shambles.

There are plenty of things to be concerned about.  This is not in the top 38 but ACA plan pricing is likely systemically wrong if these tariffs go through.

 

 

 

Potential Tariffs, Risk Adjustment and Insurance PricingPost + Comments (11)