The Journal of Health Politics, Policy and Law (JHPPL) just published my first peer-reviewed original research article. I worked with Dr. Paul Shafer on this article as he had requested and received through Freedom of Information Act (FOIA) week by week county level application counts for 2014-2017.
We asked a simple question that Balloon-Juice readers already sort of knew:
What happened to ACA enrollment during the 2017 Open Enrollment Period when the messaging and implementation regime switched from being very supportive to very opposed?
There had been several analysis that suggested a 4% to 5% decline in total open enrollment. The problem with these analysis is that the counterfactual (2017 trend would have looked like 2016 if the Obama HHS had run the entire open enrollment) was lightly tested.
We ran a descriptive analysis that basically was a repeat of the previous analysis. It suggested that something big happened.
We then ran two, more rigorous models. The first was a difference in difference design which found that the last two weeks of the 2017 Open Enrollment Period had a 30% reduction in applications for insurance. An event study design found a 24% reduction in applications in the last two weeks of 2017 compared to the pre-inauguration trend.
So what does this mean?
9/11 Motivated buyers who know that they need health insurance already bought or renewed their coverage earlier in the open enrollment period so the last wave is the healthiest wave and determines if a year is morbid or healthy for insurers.
— David Anderson (@bjdickmayhew) June 18, 2019
Folks who know that they are likely to incur a $100,000 or more in claims are very likely to sign up for health insurance that is community rated, guarantee issued and heavily subsidized no matter what the messaging regime is. It is highly salient and highly important that they get their healthcare covered. They will crawl through glass for coverage. We can also assume (but did not prove as our data was not that granular) that the vast majority of this population had signed up during the time period the Obama administration ran Healthcare.gov as the coverage that could be bought in the last two weeks of the Open Enrollment Period did not start until March 1st. I think it is reasonable to assume that folks who know that they have a $100,000 claim year would want coverage that started on January 1st instead of running naked to save a few thousand dollars in premiums while spending $10,000 or more in medical expenses.
However, folks who think it is likely that they will be in the bottom 50% of the US healthcare spending distribution ($0-$1,000 in total spend) tend to enroll late. They are the ones who need strong encouragement to go onto the Exchanges and buy. This cohort is where the drop-off is likely to have happened. And once they are not in the 2017 pool, they can’t auto-enroll for 2018 or benefit from Silver Loading as it is unlikely in a low outreach/negative messaging environment that they would have the same probability of looking at the Exchanges for 2018 as unenrolled individuals than if they had been enrolled in January 2017. This is a continual drop in enrollment over several years.
Messaging matters, operationalization matters, but we can’t say a ton more. We can’t say if the executive order drove X percentage of the drop, or if the cancellation of advertising was all of the drop and then some more, or if the general elite political messaging that Obamacare sucked and it would be replaced with something better real soon now did anything. We have hints of a story in the data but nothing strong enough to write on. Further research is needed on these factors (and I’m on a great team doing some of that work) but messaging and political support matters for enrollment in public programs.