Wednesday, May 18, 2016

How many Americans who buy their own insurance are unsubsidized?

One further footnote on the McKinsey & Co. report on the state of the post-ACA individual market for health insurance, discussed in my prior post.

McKinsey opines that the market is unlikely to enter a death spiral because ACA premium subsidies insulate most participants from premium hikes. The report estimates that 69% of current enrollees are subsidized. That strikes me as at least slightly exaggerated.

Kaiser, in its 2015 survey of non-group enrollees on- and of-exchange, found that just under half were subsidized -- or about 59% if you exclude those in grandfathered and grandmothered pre-ACA plans, which are a different risk pool (McKinsey's "methods" note indicates that they include enrollees in non-compliant plans in their estimate of off-exchange enrollees, though they may also assume that some of those in noncompliant plans are in subsidy range)*. UPDATE 5/20: Kaiser's 2016 individual market survey results, reported today, find that 64% of individual market enrollees are now in marketplace plans. 83% of marketplace enrollees are subsidized, suggesting that 53% of the individual market is subsidized.

The off-exchange market is something of a black box, so it's not surprising that estimates would differ. Also, Kaiser's estimate is from 2015, and the ACA marketplace grew modestly this year. Still, on McKinsey's own terms, I think their estimate is a bit inflated.  Here's its basis:

Not everyone in that 300-400% FPL category is subsidized. A subsidy kicks in if the premium for the benchmark silver plan costs more than 9.7% of income. For solo young adults, subsidies can fade out below 300% FPL, though for households with two or more enrollees, they generally (always?) extend to 400% FPL. According to the Kaiser subsidy calculator, the average age nationally at which subsidies zero out for a solo enrollee at 300% FPL is 34 (the exchanges would probably have a lot more young enrollees if premiums were capped at 8.5% of income, as Hillary Clinton is proposing).

HHS enrollment stats indicate a fair number of unsubsidized buyers under 400% FPL. In the 38 states using the platform, accounting for 74% of enrollees, just 2% reported income over 400% FPL, while another 7.6% did not supply income data (and so probably mostly earned too much to qualify for subsidies). But only 85% of all enrollees obtained subsidies -- so a good number of those who did not must have had incomes under 400% FPL.  8% of those who reported income, or about 711,000, had incomes in the 300-400% FPL range. In total 1.44 million enrollees were unsubsidized.

Not all unsubsidized buyers with incomes under 400% FPL were in the upper ranges of subsidy eligibility. In 2015, California reported** a number of unsubsidized buyers in every income category, and the lowest percentage was 1.3% (in the 138-150% FPL band). Some buyers in every category apparently were disqualified, either because they indicated they would not file a tax return or because their employer offered insurance that they found unaffordable or otherwise opted not to buy. It would be useful to know those percentages nationally.

Update: It occurs to me that the McKinsey estimate of the percent subsidized should not really depend on income breakouts. Since we know, as the report notes, that 83% of marketplace customers were subsidized, the question is how many enrollees there are outside the marketplace [note update above: Kaiser's 2016 survey suggests that 64% of individual market enrollees are in marketplace plans]. To address that question, McKinsey pretty much offers a black box, referring  to their own "MPACT coverage model":
MPACT is a micro-simulation model that uses a comprehensive set of inputs and a distinctive approach to modeling consumer and employer behavior to project how health insurance coverage may be changing post-reform. MPACT contains 300 million “agents” representing all residents of the United States. Each agent is characterized by his or her county of residence, type of insurance coverage, and eight demographic variables. Over the course of the microsimulation, agents in each geo-demographic segment make health insurance purchasing decisions depending on their eligibility, prior purchasing behavior, demographics (including health risk status), subsidy eligibility, and penalty impact, among other factors.

* Corrected -- I originally assumed that McKinsey's estimate did not include those in ACA-noncompliant plans.

**See the June 2015 profile, available here.

No comments:

Post a Comment