[Ed. Note: On Friday, May 2 and Saturday, May 3, 2014, the Petrie-Flom Center hosted its 2014 annual conference: "Behavioral Economics, Law, and Health Policy." This is an installment in our series of live blog posts from the event; video will be available later in the summer on our website.]
Our esteemed moderator Gregory Kurfman of the New England Journal of Medicine oversaw a session that dug deep into how defaults work and why. The result was a better understanding of the regulatory tool most associated with soft paternalism, but doubt about whether its operation in healthcare is as libertarian or asymmetrically paternalistic as advertised.
The session started with Anna Sinaiko presenting her paper, “Terminated From Medicare Advantage, Now Choose.” (Coauthors Richard Frank and Richard Zeckhauser.) The paper is focused on patients whose chosen Medicare Part C insurer leaves the program, forcing the patient to re-make a choice she already made.
First, some motivation: Anna points out that plan terminations are increasingly common, creating “terminated choosers,” that is, beneficiaries who had chosen an insurer only to have to choose a different one. Terminated choosers exist not only in Part C (Medicare Advantage), Medicare Part D (Prescription drug), Medicaid (to the extent that states use private insurers), and even private employment. So what should we think about TC’s?
The paper investigates: (1) What do TCs choose, (2) what should they do, and (3) what should choice architects do to align #1 and #2?
Through Medicare advantage, 14.5 million beneficiaries who opt to do so obtain their government-paid coverage through private insurers. The study identifies 234,000 TCs during the study period (2006-2010).
The study finds that the default appears to have power over the TCs even though they had once opted into Medicare. While “[o]ptimal policy is not clear,” Sanaiko proposes that we might make the default a “smart default,” that is, one that aligns closest with the TCs prior choice.
Next, David Tannenbaum presented his paper, “Default Behavior as Social Inference.” The study investigates why defaults for things like retirement savings “stick.” While the conventional wisdom is that inertia/procrastination/bias causes stickiness, Tannenbaum’s paper investigates the extent to which information signaled by the default is what leads employees to follow the default. The mechanism underlying stickiness matters, Tannenbaum notes, both for how status quo bias works and for how we should use it.
Tannenbaum studied these questions by posing vignettes to laboratory participants and through the behavior of students given choices about when to make their papers due, teasing out the extent to which the inferred advice from the default option. He drew out several interest results, but I want to focus here on the key result of his first (laboratory) study. That study found that across contexts an auto-enrollment default was taken to signal more information than a non-enrollment default. In his words, default is especially sticky “when the default is viewed as an intervention.”
Tannenbaum’s finding is especially interesting in light of a key normative argument for libertarian paternalism that has come up throughout the conference, namely, that regulators “have to” pick a default, so might as well choose the default they think best. Tannenbaum’s research suggests that regulators are actually doing more to alter behavior when they set a default as participation than when they set the default as non-participation. (But perhaps the perception that a “default” is an intervention may ultimately reflect the prior governmentally-created default?)
Third up was Elliot Doomes presenting his paper “How Choice Architecture Can Define a New Era of Providing Mental Health Treatment.” (Coauthor Aeva Gaymon Doomes.) Doomes’ coauthor (and partner) is a doctor of psychiatry, and Doomes started the paper by focusing on the magnitude of the mental health problem in the United States. He also noted that the ACA mandates not only that mental health benefits are “essential health benefits,” but that such benefits be put on parity with other benefits. Doomes hopes that this will lead more medical professionals to train in mental health and take insurance.
Doomes then noted two barriers to mental health treatment: (1) access, exacerbated by a lack of mental health professionals, and (2) stigma associated with seeking mental health care. The latter barrier is Doomes’ launching off point for a potential intervention: He proposes screening for mental health illness by default in schools and workplaces.
A default mental health screening system operated through schools and workplaces could undercut the stigma associated with mental health care and thereby expand access to mental health services dramatically, Doomes hopes. On an optimistic note, he suggests that such an approach could not only address the mental health problems of those who currently go without care but also limit the frequency of school and workplace tragedies.
The last presentation of the day was a great way to sign off. David Orentlicher revisited what has been called the most famous default, that is, the default rule for organ donation. And he had some push back for the conventional wisdom in his paper, “Presumed Consent to Organ Donation.” Orentlicher points out that in public surveys some 85% of Americans say would be willing to be organ donors, so it is not surprising that we think that switching the default to opt-out (that is, presuming consent to organ donation), would be a good thing because it would overcome intertia or signaling and thereby increase donation rates.
But this story is not so straightforward, Orentlicher says, for a few reasons:
- Opt out policies can disproportionately affect the less-informed or less-educated.
- We should be skeptical about inferring that public survey data reflects what Americans really want, because people answering surveys are inclined to give the answer the asker wants. For example, about 45% of Americans opt into organ donation, but about 62% of Americans say they have opted in on the phone.
- Many donors who fail to opt out may in fact be unwilling donors. Which, Orentlicher asks, is the greater harm: failing to transplant an organ from a willing donor, or transplanting an organ from an unwilling donor?
- The high participation rates in presumed consent European countries, like Spain, may reflect other features of those countries.
So, now what? A complicated story and questionable empirical data creates, for Orentlicher, skepticism about whether presumed consent is a good example of libertarian paternalism. Rather, because the default pushes people to donate organs to benefit others, not themselves, and the resulting “choices” do not obviously better reflect the individual welfare of the chooser, presumed consent laws may work better as an example of Korobkin’s libertarian welfarism.