We all saw this New York Times story over the weekend about how the Trump campaign used extraordinarily aggressive and sketchy fundraising tactics last fall:
What the Blatts soon discovered was $3,000 in withdrawals by the Trump campaign in less than 30 days. They called their bank and said they thought they were victims of fraud.
“It felt,” Russell said, “like it was a scam.”
But what the Blatts believed was duplicity was actually an intentional scheme to boost revenues by the Trump campaign and the for-profit company that processed its online donations, WinRed. Facing a cash crunch and getting badly outspent by the Democrats, the campaign had begun last September to set up recurring donations by default for online donors, for every week until the election. [ my emphasis added ]
There is an amazing amount of details in the story that are well worth reading but the word that is sticking out to me is default.
Defaults matter a lot because there is a simple fact of both physics and human behavior — inertia matters a lot. It is easier to get a person to do nothing than to get them to do something. Therefore, defaults matter a lot. This is one of the key insights behind the entire “nudge” movement. Setting defaults matter because most of the time, people will not readily or easily move out of the default.
In the NY Times story, the Trump campaign used an “opt-out” default for people to make a recurring donation. The donors had the ability to move but that ability was getting buried further and further underneath an avalanche of words that were designed to impose cognitive costs and short-circuit critical thinking. Opt-outs mean a lot of people will stick around while “opt-in” solutions that require checking a box will get a lot fewer people making an active choice to make a recurring donation. Opt-outs can be beneficial; 401Ks are currently opt-out enrollment as a means of a socially beneficial nudge.
Why does this matter for health insurance?
Picking insurance is hard. Picking insurance is painful. Figuring out optimal or at least satisficing choices is demanding. People have a lot of choice inertia.
A new working paper by Brot-Goldberg and his merry gang of collaborators found that the downsides of bad defaults don’t motivate people to move insurance plans in the Medicare Part D market.
We leverage two unique natural experiments to show that, in public drug insurance for the low-income elderly in the U.S., defaults have large and persistent effects on plan enrollment and beneficiary drug utilization. We estimate that when a beneficiary’s default is exogenously changed from one year to the next, 96% of beneficiaries follow that default. We then develop a general framework for choice under costly cognition that allows for the possibility that either paternalistic defaults that steer consumers to plans that suit them (Thaler and Sunstein 2008) or ‘shocking’ defaults that trigger consumers to make active choices (Carroll et al. 2009) could be optimal. We show that optimal default design depends on a previously-overlooked parameter: The elasticity of active choice propensity with respect to the value of the default. Leveraging variation in the match value of randomly-assigned default plans, we estimate an elasticity close to zero: There is little difference in the probability of active choice between beneficiaries assigned a well-matched default versus beneficiaries assigned a poorly-matched default.
Translating this out, they took advantage of the fact that a lot of people qualify for zero premium Medicare Part D plans and when a plan left a market, these folks were re-allocated to other zero premium plans by default with an opt-out to choose something else. Almost no one chose something else even if the new default did not have the drugs that the person needed or had way more cost-sharing. There are two lines of thought on how to get people into good plans. The first is to use smarter defaults with behind the scenes technocratic tinkering that apply a set of rules to place people in plans that are not bad but may not be optimal. The second is to default people into something horrendous and use the hideousness of the choice to overcome inertia. That second option does not work in this context. In Medicare Part D, since very few people will opt-out of a bad choice, smart(er) defaults are the best option to avoid bad choices and improve aggregate choice quality.
My frequent co-author and friend Coleman Drake and two other co-authors have analyzed the ACA markets in California for inertia in another new working paper. They find big inertia effects independent of network preferences that are expensive and tough to overcome:
We find that nearly all inertial plan choice results from inattention and hassle costs, the former more so than the latter. As a result of these two sources of inertia, consumers lost over a billion dollars in forgone surplus in 2018, or $1,440 to $1,584 per household per year
Finally, Adrianna MacIntyre and team at Harvard, in a just released working paper, looked at the movement of folks who failed to pay premiums towards zero premium plans in Massachusetts. This smart default matters a lot:
Using data from Massachusetts’ health insurance exchange, we study an “automatic retention” policy intended to prevent coverage interruptions among low-income enrollees. Rather than disenroll people who lapse in paying premiums, the policy automatically switches them to an available free plan until they actively cancel or lose eligibility. We find that automatic retention has a sizable impact, switching 14% of consumers annually and differentially retaining healthy, low-cost individuals.
Any time that we expect people to actively opt-out by checking a box or setting up a payment scheme, we can expect a lot of failures for people who really would like to frictionlessly opt-out. These costs are individually small and could seem to be easily surmountable, but over time they add up to become significant hurdles that are more often taxing on cognitive load and management budgets rather than cash budgets. Active opt-outs reduce movement out of a default. Active opt-ins reduce movement into a choice that is not a default. Setting up a default and then imposing significant costs to move from that default creates inertia. The Trump campaign used this insight to very aggressively fundraise. Health insurance marketplaces can use these insights to place people into non-dominated plans and hopefully exceed that minimal goal of avoiding domination by defaulting people into plans that are pretty decent even if not optimal.
marklar
Really nice post, David. I’ve sent it to a former student of mine who is doing graduate work in medical decision making out in Oregon.
My favorite example of nudges in the medical system is organ donation. In Europe, many countries require you to opt out. In the U.S., we have to opt in. Not surprisingly, the majority of Europeans in those countries are organ donors, while the majority of Americans are not.
This would be such an easy fix for us, but, you know, freedom.
Omnes Omnibus
Sounds a lot like libertarian paternalism. Not really a bad idea. Also, I didn’t come up with the name so don’t bitch at me about it – you know who you are.
OzarkHillbilly
I have one thought on this: Make it easy for people know what plans offer.
The Medicare sight is a nightmare. The A, B, C, D, E, F, G, H, I, JKLMNOP plans all have very similar descriptions with very minor (or major if one thing or another is actually important to a person) differences. I got dizzy just reading them all and trying to keep them straight. Finally played pin the tail on the donkey and opened up a type so I could compare the plans offered under that type, and even then it was ridiculously difficult to figure out so again I played PtTotD.
What I’ve got is pretty good, it’s not carpenter’s but it is better than my wife’s employer’s policy. There might very well be something better but Just the thought of wading back into that snakepit sends me into an inferno of anxiety.
Cheryl Rofer
So much this.
Raise your hand if your fave thing is to wade through deductibles, doctor and prescription availability, and various metal names.
Yeah, I thought so.
We have so many demands on our attention. I particularly do not like dealing with finances – I recognize that’s important, but adding columns of numbers and keeping track of expenditures and income is much less pleasing to me than practicing piano. Or cleaning the house.
I don’t get the correlation between the insurance variables – deductables, co-pays, and whatever else they load in. If there were a formula that developed monthly payments from all those things, it would be easier to compare plans. I suspect the insurance companies have such a thing, but no way will they tell us what it is.
I suspect this is the reason many people say they prefer single-payer. Just one organization to deal with.
Barbara
@marklar: I don’t see organ donation and participating in your company’s 401k plan to be the same thing. Among other things, even if you check the box as being willing to donate an organ, at the point of donation, at least in the U.S., your family’s wishes on the matter will be given greater weight than the checked box (though that might be used to persuade them to agree to donation). Efforts to bypass family consent have been met with a lot of controversy.
Regarding the post, I would add that a significant percentage of beneficiaries are auto-enrolled into Part D plans, so the concept of auto-enrollment based solely on cost is already embedded into the way the program works. Adding plus factors based on comparability/suitability of plans to the auto-enrollment process would be beneficial and, if not reduce, at least not provide undue rewards for the gamesmanship that goes into making sure a plan meets the threshold eligibility requirements for auto-enrollment.
zhena gogolia
@Cheryl Rofer:
I’m with you!
Barbara
@OzarkHillbilly: The Medicare.gov website practically begs you to provide information on the drugs you take so that it can help choose a plan that gives you lower out of pocket costs. It’s not hard, but it’s hardest for those beneficiaries who face the greatest obstacles for signing up for a vaccine appointment. There is not doubt in my mind that even the most thorough or simplified explanations would generate the kind of changes that would be optimal for beneficiaries.
Among other things:
1. People routinely discount the amount of care they will need and the expenses associated with that care.
2. In light of the above, it doesn’t matter how much you try to educate beneficiaries, for many if not most, a zero dollar premium feels like thrift and a monthly premium feels like a waste.
Chyron HR
“But”?
Booger
Maybe we should lean into that, and go with options like “HËLLTH-KÄRE PLÄN BLACK“
EmbraceYourInnerCrone
@Chyron HR: Yep, an intentional scam, from the Trump organization. What a surprise. I feel sorry for them, but they followed a man who has made it his life’s work to scam, cheat and rob people while using the legal system to avoid paying for his crimes. Like Maya Angelou said, When someone shows you who they are, believe them the first time.
Doug R
@marklar: I was just going to comment on organ donation and how it should be opt-out.
I think motor voter laws at least require asking the question, does that make it opt in or out?
germy
Matt McIrvin
In software development I call this the “discipline and punish” approach: if you want somebody to do something, check in a change that breaks the build until they do it. It doesn’t actually work much better there.
Brachiator
@Barbara:
I never knew that, or maybe skipped over it when my sister and I were trying to help our mother with this.
I also have observed that navigating some web sites is difficult for some seniors, and it is not a matter of not being “computer-savvy.” It’s more that reading and navigating web sites is not always easy for older eyes and fingers.
dnfree
@OzarkHillbilly: there are two good sources of help to sort this out. Your local senior center, if you have one, is one source. In the small city where we formerly lived, we went there every year to have them help us sort through Part D plans based on total cost for our medications plus the monthly premium. The people at these centers are usually trained volunteers.
In the bigger city we now live in, those appointments were filled, so we worked with a professional at an agency called LifeSmart to do the same thing. There are other similar agencies, no cost to you and there are rules in place that are supposed to keep them from going beyond the question you asked.
For Part D they are mostly using software provided by Medicare. To David’s point, a couple of years ago that software was set to optimize the premium being the lowest, not your personal out-of-pocket cost. The adviser we worked with knew to change that default to optimize what we would actually be paying. Since I have a couple of expensive drugs that you see on TV, that’s a big difference. If I’d done it on my own I might have missed the new DEFAULT.
dnfree
@Doug R: in Illinois it is opt out at the DMV for organ donation.
Matt
@EmbraceYourInnerCrone:
I saw the bit where the guy in hospice for cancer ended up making a $500 / week donation until he ran out of cash.
My only question for him is, “DOES CANCER HURT, YOU DEPLORABLE SHITHEAD?” The man’s dying and all he can think to do is hurt other people, I hope he suffers long and hard.
taumaturgo
Picking unsuspecting marks is easy for Trump and the insurance cartel. Figuring out how to scam people with a myriad of choices is not demanding at all. People have a lot of inertia as to how to avoid scams.
marklar
@Doug R: The fact that the question is asked is great, otherwise it would virtually never cross people’s mind.
I still see it as ‘opt in’ because the default is that the box isn’t checked. There is a lot of research under the term “status quo bias” that explains how this all works.
BTW, this is also why there was such resistance to the implementation of the ACA when it started, and why Obama knew that once the program was established people would want it to remain. One of the consultants tapped by the Obama administration was Cass Sunstein, who along with Richard Thaler wrote the book “Nudge”, which is an easy to digest trade book focused on the choice architecture that David is writing about here.
VOR
The term of art in the software world for the Trump campaign’s user interface design is “Dark Patterns”. From an article at The Verge:
Starfish
@Matt: My sister had to go on disability leave very suddenly a week or two ago from her job. New York gives everyone disability pay capped at like $170 a week.
Buy food? Pay the pile of medical bills coming her way? Donate to a campaign? It is not.
evodevo
@marklar:
Yep…here in KY it requires you sign on yur drivers license with TWO witnesses…where am I gonna find two witnesses? I would have preferred the European way…
Jesse
In my experience with Act Blue, most donations have set up their donation forms to be default one-time, opt-in recurring. A minority are set up to be default recurring (opt-in one-time donation). Is that the top and bottom of the Trump thing?
Barbara
@Brachiator: I think most people skip over it because while they are clearly asking for the information that doesn’t make the information easy to input or to remember. If the beneficiary is looking for a Part D only plan (not integrated with Medicare Advantage), entering drugs is definitely something you should do because it will really clarify the choices, especially if the person is taking an expensive brand name medication. However, if they are looking for an integrated plan then they will be put in a position of potentially trading off current prescription needs with favorite doctors or other providers. Although, integrated plans tend to have better drug benefits because the integration gives them better management tools and they typically have more money to play with.
Yeah, it’s way too complicated.
Martin
Because I can’t link to it often enough.
Are we in control of our own decisions?
I watch that about every 6 months or so (it helps that it’s entertaining).
Also, CA has newly passed legislation banning dark patterns for privacy decisions, but I don’t think it would have applied here.
Me in Seattle
I read this and shake my head. I have original Medicare and Tricare-for-Life to cover part D and everything else (which includes covering the Medicare deductible). The only issue I have is getting my Doctor’s office to find the right code for my “annual physical” which Tricare Prime covered just find less the $15 co-pay. Apparently Tricare-for-Life doesn’t use those codes.
Anonymous At Work
Sunstein takes a lot of shit, and more than a little earned, for the academic-contrarianism. But the Nudge approach did a lot in the regulations with which I work, *and which were written under his auspices during his time at the White House*, where “opt-in” was changed to “opt-out” for a lot of things and you had to work extra hard to add justifications and policies for the extra work. It became easier and expected to be flexible and do less paper-grinding.
So, shout out to Cass Sunstein and Nudge.