Tuesday, April 03, 2018

"Strong" CSR takeup dropped only slightly on HealthCare.gov in 2018

Update/correction: This year, for the first time, CMS broke out enrollment at each level of CSR (at actuarial values of 73%, 87%, and 94%) -- and separately, metal level enrollment at each income level (indirectly - see note below). It seems that somewhere between 2% and 5% of silver enrollees from 100-200% FPL did not obtain CSR, probably mainly because they were ineligible for subsidies because of an offer of insurance from an employer. There's no read on how many enrollees had incomes below 100% FPL, and how many of those obtained CSR (as legally present noncitizens time-barred from Medicaid can do).  I have extrapolated from 2016, when HHS did report the percentage of enrollees under 100% FPL, edited the post accordingly. Originally, I estimated a 7% drop in CSR takeup among those eligible for "strong" CSR, i.e. with incomes up to 200% FPL. In fact it's more like 2%.

I noted recently with respect to Maryland ACA marketplace enrollment, the downside of "silver loading" the cost of CSR is a drop in takeup of the strong CSR available to enrollees with incomes below 200% FPL. CSR is available only with silver plans (please see the prior post to un-abbreviate all this).

That downside (for the subsidized) is outweighed by the upside. In Maryland, about 30,000 enrollees with incomes over 200% FPL obtained steep discounts in bronze and gold plans, whereas about 6,000 fewer enrollees under 200% FPL obtained strong CSR than would have had Trump not disrupted the market by cutting off federal funding for the benefit (see appendices to post linked to above for a quick explanation).

Now we have enrollment numbers for the all states -- with, as usual, more precise breakouts for the the 39 states using the federal exchange, HealthCare.gov than for the whole country.  In HealthCare.gov states, CSR takeup among those with incomes up to 200% FPL downticked only slightly.  -- to  about 80%, from (probably) about 82% last year. Silver selection among those with incomes ranging from 100-200% FPL dropped from 86.9% in 2017 to 85.3% in 2018. For 2017, we don't know how many silver enrollees in the income bracket did not obtain CSR.

Nationally, in 2018, 53% of enrollees obtained CSR, versus 58% at the end of open enrollment 2017. On HealthCare.gov, the 59.9% obtained CSR in 2017 vs. 54.4%. But the difference was largely concentrated in the 200-250% FPL income bracket, where CSR is negligible (boosting AV from 70% to 73%) and enrollees had ample cause to take gold and bronze discounts. In 2017, 67.6% of enrollees in the 200-250% FPL income category selected silver plans. In 2018, just 53.4% did.

There was also a significant drop in the 150-200% FPL bracket, where CSR boosts the AV of a silver plan to 87%. At 100-150% FPL, however, where CSR-boosted silver is 94% AV, and the percentage of income required to obtain benchmark silver is 2-4%, silver selection rose slightly. Here is the breakout - keeping in mind that 2-4% of those who selected silver did not obtain CSR.

Metal Level Selections at Different CSR-eligible Income Levels (% FPL)
HealthCare.gov states


                              Strong CSR          Weak CSR                
Metal level
 <  1%


                              Strong CSR          Weak CSR                
Metal level
  0.09 %

Viewed another way, the drop in CSR enrollment on hc.gov in 2018 was about 300,000 larger than the total drop in enrollment (458,163 fewer enrollees; 754,207 fewer with CSR). Though the 200-250% FPL income bracket accounts for only about 20% of CSR-eligibles, it accounted for about two thirds of that drop. The drop was proportionately steepest in the 200-250% FPL band, as the chart above indicates. There were 204,954 fewer silver enrollees in the 200-250% FPL group in 2018 than in 2017, a 23% drop, whereas total enrollment in the income band dropped less than 3%.

That drop-off is not surprising, as many enrollees had access to substantial discounts in gold and bronze, thanks to the loading of the cost of CSR onto silver plans only in most states. About two thirds of the drop-off in silver at 200-250% FPL went to bronze, and one third to gold.
CSR procurement lower than silver selection among CSR eligibles

According to the public use files just released by CMS, 4,067, 973 enrollees chose silver plans with strong CSR (2.6 million at the highest level, with an actuarial value of 94%, and 1.4 million at 87% AV).  4,865,013 enrollees had incomes between 100% and 200% FPL.  Another 348,268 are in the "other FPL" category -- i.e., they supplied income data, but they're either under 100% FPL or over 400% FPL. In 2016, HHS reported that 3% of hc.gov enrollees for whom income was known, or about 267,000, had incomes under 100% FPL. That suggests about 240,000 this year. I am estimating, based on past years, that 175,000 of them have incomes under 100%  FPL, most of them legally present noncitizens ineligible for Medicaid but eligible. (Most of the unsubsidized over 400% FPL don't supply income information.)

That would makes 5,105,014 with incomes in range for strong CSR, with roughly 4,068,000 obtaining it, or 80%. As noted above, however, it appears that 85% of enrollees with incomes in the 100-200% FPL range selected silver plans. The unknowns: how many enrollees had incomes under 100% FPL (it was 3% in 2016, when HHS provided the info), and what percentage of those low income enrollees obtained CSR (as legally present noncitizens barred from Medicaid can do).

Note on metal level selection totals 

For each metal level, CMS breaks out what percentage of total enrollees were at each income level. The number that I'm concerned with -- the percentages at each income level that select each metal level -- has to be extrapolated from that, and it's rounded as a whole percentage. For example, in 2017 there were 6,827,122 silver enrollees on hc.gov, and 42% of them had incomes in the 100-150% FPL range. That comes to 2,865,291, a number that really probably should be rounded in some way, given the rounded percentage.  But I take it as 89.3% of the 3,208,242 enrollees in the income bracket.

No comments:

Post a Comment