4.29% enrollment loss is the cost of Trump’s First Day EO

I’ve been playing a bit more with the 2016 and 2017 QHP data in an attempt to figure out the incremental cost of the Trump Executive order.  I think 4.25% is a good lower estimate.

My data is still here:

Data and Methods

I again excluded Kentucky and Louisiana.  Kentucky was switching from Kynect to Healthcare.gov while Louisiana had a mid-year Medicaid expansion.  I wanted to isolate the effect of the executive order from whatever the general trend in enrollment was.  I used the CMS enrollment snapshot for 2016 and 2017 that contained January 14th.  2016 was goes through January 16 while 2017 only goes through January 14th.  The 2016 report contains two extra days worth of data and more importantly, 2016 contains a deadline day as people who buy coverage by the 15th would see their policy start on February 1st.  We know deadlines spur enrollment.

CMS recognized this problem:

More than 8.8 million Americans were signed up for 2017 coverage through HealthCare.gov as of January 14, 2017. This compares to about 8.7 million sign-ups as of January 14 last year, as Americans continue to demonstrate strong demand for 2017 Marketplace coverage.

So on the 14th of each year, 2017 was running slightly ahead of 2016.  My data due to timing constraints will show 2016 running slightly ahead of 2017.   This is fine as the known flaw in the data favors the argument that the executive order had no impact.

So the question is what was the deviation from 1/15 to 1/31? If the Executive Order and the dropping of advertising and potentially elite knowledge networks disseminating anti-enrollment messaging or more likely fear, uncertainty and doubt about PPACA being a good play?

Analysis and Conclusion

2017 using my known flawed data was running .96% behind 2016 on the January 14th inclusive update.  2017 ended up running 5.25% behind 2016 on Healthcare.gov states.  The increment (using favorable to the null hypothesis data) slowdown in pace that can be attributed to Trump Administration actions is 5.25-.96 or 4.29% of enrollment was lost due to the executive order and other Trump administration actions such as shutting down some outreach and advertising in the last eleven days of enrollment.

4.29% is a minimal level of enrollment loss.  Using the January 14th pace, 2017 was running 1.1% ahead of 2016.  Charles Gaba is collecting data from the state based exchanges.  The state based exchanges ran their own marketing campaigns that did not get shut off on 1/20/17.  He is showing at least a 1.5% enrollment increase.  So more aggressive baselines can credibly argue that the Trump Administration actively discouraged 6% of the market from signing up.

Finally, here are some charts that I had fun creating as I worked through this problem.

Partisanship and enrollment changes on Healthcare.gov

Charles Gaba outlines what happened on Healthcare.gov for final enrollment. The numbers came in lighter than expected.

Then, Donald Trump actually took office…and in a one-two punch, not only signed an executive order which specifically instructed the HHS Dept. to do everything in their power to sabotage the implementation of the ACA, but also attempted to pull the plug on the critical last-minute advertising blitz which is so vital to reminding procrastinators (particularly young ones…you know, the ones who help the risk pool?) to get off their duffs and actually sign up.

The CMS numbers:

On January 31, 2017, Open Enrollment for 2017 coverage ended with more than 9.2 million plan selections in states that use the HealthCare.gov eligibility and enrollment platform.

I was curious about something over the weekend but I was with my kids. I had to chase them and learn about rocket ships.

At the state level, does partisanship have a relationship to enrollment levels. Would states with a higher Trump vote percentage have greater declines in enrollment?

My data is here for any and all to view:


I excluded Kentucky from the analysis as they transitioned from Kynect to Healthcare.gov.  I also suppressed Louisiana as they had a mid-year Medicaid expansion so we should expect a massive drop in Louisiana enrollment no matter what (and that is what happened).

UPDATE 1: Removing Hawaii as an outlier transforms the relationship from statistically significant to not significant

The short answer is yes, and significantly so.It is not a perfect relationship as South Dakota and Utah each saw significant gains.  But each point of Trump voters led to 4.5 basis points less enrollment.  A state that perfectly mirrored Trump’s actual popular vote total would have had effectively not lost enrollment.  A 50/50 state would have lost 2.5% of the QHP’s.

What is the mechanism?  I don’t know.  I would bet that elite cuing and partisan information channels would be a major part of this explanation.  Trusted sources in red states were telling people that the markets were blowing up and outreach efforts were either hobbled or sidelined so the marginal people who were making the last buy/no buy decisions were making incremental no buy decisions.  I think if we see an older than expected risk pool, that will be a chunk of evidence in support of that idea.

The timeline to run

The Republican Party does not have the luxury of time to get their act together on health policy. They have maybe fifty days to get a coherent plan with the possibility of passage before external actors move to foreclose on Republican policy option space:

Below is the Center for Medicare and Medicaid Services plan filing timeline.

Read more

Keep on calling

Let’s keep on working:

Right now the best targets are probably Flake (R-AZ), Heller(R-NV) and Toomey(R-PA). The vote count assuming no other surprises is 50-50 so we don’t need a no vote, we can work with an urgent need to stay in their state to listen to constiuents, a strong desire to wash their hair, unexpected HVAC issues or a sudden sphincter spasms that keep them on the Senate can.

We won’t get many wins, but let’s see if we can get a win here.

Call Congress.

Good news through the lack of bad news

One of the first potential blow-up points for the ACA in the Trump Administration was today. Insurers are required to offer lower deductibles and cost-sharing to people who buy on-Exchange policies and who make less than 250% of the Federal Poverty Level. This is the Cost Sharing Reduction (CSR) subsidy. Insurers offer the better product and the Federal government pays the insurers on the back end for the increased value of the Silver plan.

There is a lawsuit (House v. Burwell which will be renamed) that the District court judge ruled that the House’s argument that the CSR subsidies were discretionary spending that had not been appropriated was correct instead of accepting the Obama administration’s argument that the spending was mandatory and thus automatically appropriated. It has been appealed by the Obama Administration and once the administrations changed, the Appeals Court granted a delay in the appeal as the Trump lawyers figure out what they want to do.

However, the current CSR system where money goes out the door to pay insurers is based on administrative interpretation of the law. At any point in time, the Administration could decide to re-interpret the law and agree with the House that the funding is discretionary and that there is no money available to legally pay the insurers.

As I outlined in early January, this would blow up the Exchanges:

CMS in their 2017 QHP contracts allowed carriers to pull products from the market if the CSR subsidies disappear and it looks like that would be the plan of CHC to pull their Silvers….Carriers have to offer Silver plans to participate on Exchange. If they yank all of their Silvers, they have to yank everything on Exchange.

And carriers will flee if CSR disappears as they will not eat a 30% revenue loss for a high cost population in a market that they don’t know if it will be around long enough to actually make money on.

So far there is no executive order or administrative re-interpretation of the CSR funds.

This is good news. The markets and the plans survive for another month.

Call Congress (and see results)

If you’re happy and you know it —- Call Congress
If you’re scared and you know it — Call Congress
If you’re mad and you know it — Call Congress

Keep up the calls. Call about what you are passionate about, call about what you care about. Call about what you can.

And see results:

As you’re calling Democrats who are in tough Districts (Senate Heitkamp is in a Trump +30 state) tell their staffers thank you. Tell them you appreciate that their boss is going out on the limb for our vision and values of America. The staffer answering the phone has no power and is getting deluged right now. Be nice and then register your opinion.

We’re getting results, so keep calling

Trump and the global creative class

Richard Florida was one of my professors in grad school.  He was an adviser to a couple of projects that I loved working on and he served as one of my early job references.  I think his Creative Class work is interesting despite significant causality concerns of the chicken and egg and its inability to really speak to distributional issues within regions.

The core insight that powers his work is that it is wise to pay a lot of attention to where people with lots of choices end up and concentrate.  He focuses a lot on younger people who are heavily educated and who have weak ties to their home geographic area.  This is because these individuals can move fairly cheaply and fairly broadly.  A twenty seven year old with no kids and no mortgage is far more mobile than a forty three year old with two kids and seventeen years left on their mortgage.  This is a really smart insight.  Where it goes from there may be another story (bike paths might not save us all).

Why is this important?


We’re much less attractive as a country to very smart, mobile international students and researchers this week than we were last week. And I bet that we’ll be less attractive next week than this week. Think about the incentive for a twenty something looking to do a post-doc? They could come here and work with an awesome group but when they need to travel to a conference overseas they might not be allowed back into the country? Or they could go to Canada or Japan or Australia or the EU where they don’t have that new worry as a cost to their calculus of choice.

Will this stop all international brain drain that the US massively benefits from? No, but it will impede it.