Important Correction to Yesterday's Post

I need to issue an apology/mea culpa for a fairly major screwup I made yesterday.

In my post rounding up projections of 2026 year-end ACA exchange enrollment from several major insurance carriers, I correctly tallied up the estimates from four of the five insurers via an article by Amy Lotven of Inside Health Policy:

• Centene projected a loss of 1.5 million people
• Evevance projected a loss of 400,000
• UnitedHealthcare projected a loss of 500,000
• Molina expects to lose 430,000

Combined, these come to 2.83 million across the four, which would be pretty dramatic if it proves accurate.

However, I then fumbled the ball when I looked at the Q4 2025 earnings call for another major ACA player, Oscar Health.

Oscar reported having around 2.0 million ACA exchange enrollees as of the end of 2025, which I got right.

However, I then misread their projections for 2026; I saw CEO Mark Bertolini’s statement that they expect “market contraction may track toward the lower end of our original projection of 20% to 30%” which I took to mean losing around 400,000 (20%) of that 2,000,000 total.

Somehow I completely missed another critical quote from the summary found right in the opening paragraph: Oscar actually ended 2026 Open Enrollment with a whopping 3.4 million enrollees…which they expect to drop to around 3.0 million by the end of the first quarter.

This means that instead of losing 400K, Oscar expects to gain around a million enrollees, single-handedly cancelling out a large chunk of the 2.83 million loss from the other four carriers combined.

While my larger point still stands (a net loss of 1.83 million enrollees from carriers making up less than half the total would still be pretty significant), it was still a pretty big “oops” on my part.

I promise to be more careful in the future. I’ve gone ahead and corrected the original post with a note that I had to completely rework the entire Oscar section.

Having said that, there are actually two other related points which also came out of that correction:

• First, the fact that even with the higher OEP enrollment, Oscar also notes that they expect their effectuated enrollment total to drop nearly 12% from February 1st to April 1st, just 2 months later (from 3.4M to 3.0M).

If that drop proves accurate and is representative of the market as a whole, it would amount to around 2.7 million enrollees dropping coverage nationally by April…and even that assumes that all 23.0 million who selected plans during OEP were actually effectuated as of Feb. 1st, which almost certainly isn’t the case.

February 1st effectuations have never reached higher than 93.6% of the OEP total (and never broke 92% in non-enhanced subsidy years). Assuming it reached 93.6% of OEP this year, only around 21.6 million would actually be effectuated this month…and if that drops by 11.7% between now and April 1st, it’ll be down to around 19.0 million by then…a whopping 4.0 million lower than the 23M OEP plan selection tally.

• Second, the same section also goes on to note that the percent of enrollees who selected Silver plans dropped in half, while the number selecting Bronze plans jumped by 50% year over year, concluding that "people are carrying higher deductible plans.”

This goes to my other major point about enrollees being forced to “buy down” to worse plans.

One more side note on yesterday’s post: UnitedHealthcare said they expect to lose around 500,000 enrollees by the end of 2026, but Lotven’s story doesn’t mention how many that’s a drop from. My own Rate Change project brings up around 1.76 million enrollees via UnitedHealthcare carriers across 34 states in 2025.

(This wasn’t easy to determine given that UnitedHealthcare has a stunning 2,400+ subsidiaries, many of which don’t include UnitedHealthcare in their names, but I was able to track down the ACA market carriers which don’t thanks to this handy analysis.)

This puts total December 2025 enrollment across all 5 of these carriers at around 10.7 million out of roughly 22.4 million total, or roughly 48% give or take…so my estimate of the 5 comprising “nearly 50% of total enrollment” was accurate, anyway.