[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.]
Today opened with a plenary talk by Russell Korobkin, who is the Richard C. Maxwell Professor of Law at UCLA School of Law. His talk was titled “The Choice Architecture Problem and Health Care Decisions.”
He began by suggesting that the fundamental problem that unites the conference is that efficient health care decisions are often too difficult for boundedly rational individual to make optimally. Classical economics suggests that revealed preferences match predicted hedonic experience. Behavioral economics shows us that this is not true, and provides insights into why. The million dollar question is what should policy makers do with these insights.
Yesterday, most of discussion oriented around the solution of libertarian paternalism. But as Korobkin notes, that is just one approach and it has several drawbacks that keep it from providing a complete solution. First, policy makers don’t always know what is best, so we don’t know which way to nudge. Second, the best choice is often heterogeneous. Third, the best choice for individuals is not necessarily socially optimal. With these limits in mind, Korokin’s central claim is that the menu of non-coercive responses to bounded rationality should be expanded. He proposes two additional choice architecture options: (1) simplification, and (2) libertarian welfarism.
His first proposal is to simply decisions. To set up and illustrate this approach, he shows how it can be applied to the problem that health care costs in the US are fast approaching crisis point. Moral hazard is a significant source of this inefficiency: when people do not need to pay, they over consume. Technological innovation exacerbates the problem: new drugs are vastly more expensive than older drugs, and sometime with only a small difference in health outcomes. Further, the law prevents insurance companies from mitigating these problems: “medical necessity” laws mean that it is difficult for insurance companies to deny payment.
There are a few ways that we try to solve the moral hazard problem. The first is to create “more skin in the game.” But point of service medical care decision making it too difficult for people to make optimal decisions. People will reduce consumption, but not in rational or efficient ways. Another way to solve the problem is “consumer directed health care.” But this too is flawed: health care decisions involve complexity, novelty, inconsistent comparative information, and emotions, all of which invoke heuristics that lead to bad decisions. A third option is substituted decision-making: e.g., allow insurance companies decide. But this just trades one moral hazard problem for another, as the company has incentives to provide little care. Finally, there is “pay for performance”. But this creates a conflict of interest between the patient and doctor, who is trying to save costs rather than provide the full amount of authorized care.
Korobkin’s proposal is to simplify patient choices with “relative value health insurance.” Under this approach, a government body would rate treatments based on their relative value: for example, dollars per QALY gained. Private insurance companies would then sell insurance covering treatment to any level. This would simplify choice because consumers wouldn’t make “point of service” decisions, but rather decisions at point of choosing a health care plan. They would merely be trading off 2 attributions: cost vs. depth of service.
His second proposed addition to the “choice architecture menu” is what he calls “libertarian welfarism,” or nudging individuals towards the social optimum. To illustrate this approach, he applies it to the problem of organ donor shortages. Evidence suggests that a presumed consent default (i.e., requiring people to opt out of donation) would mitigate the problem. But, donation is probably not privately optimal for most individuals: the donor doesn’t benefit, and there could be some risk. This suggests that a libertarian paternalist would stay with opt in, rather than moving to opt out.
His proposal is that we should not only use nudges for paternalist reasons, but also for welfarist reasons. The missing category is “libertarian welfarism.” We can use techniques of nudging to do what is in interests of society. On this approach, we should have presumed consent for organ donors. Unlike libertarian paternalism (where externalities are ignored and the best choice for the actor is sometime unclear), libertarian welfarism provides clear normative justification for taking externalities into account and the social optimum is sometime clear.
He closed with some thoughts on the relationship between libertarian vs. coercive welfarism and the question of whether why, if our goal is social welfare and the optimal outcome is clear, we should not use coercive policies. He offers a few reasons. The first is that freedom of choice enhances social welfare, all else being equal. The second is that libertarian welfarism can allow for cheaper provision of social welfare: when we don’t need 100% compliance, libertarian welfarism provides social optimum at the lowest cost. For example, for people who have strong opposition to organ donation, they are not required to do it; but for people who have no strong preference, they will become organ donors. So the rule separates high cost from low cost providers of organs.
The Q&A began with a comment from David Orentlicher, who noted that there are some benefits to coercion: e.g., we are more willing to be socially minded when everyone else is doing it; if everyone can opt-out of being an organ donor, there is less reason to do it. Korobkin responded that he isn’t proposing that we rule out coercion, just that there are situations in which coercion isn’t needed/best.
The next question was from Chris Robertson, who suggested that considerations of justice might suggest that we should offer incentives to those who donate organs (otherwise, there are free-riders.). If we were to create such incentives, then there would be a justification for presumed consent on paternalist grounds, as it would be in the person’s interests. Korobkin agrees that there are various ways to change the incentives, and that he is merely trying to provide two more options.
Glenn Cohen asked about the relationship between libertarian paternalism vs. welfarism: whether paternalist policies can be “stronger” than welfarist. Korobkin replied that it depends on your views about paternalism vs. welfarism.
Matt Lawrence asked about the level of generality at which treatments would be rated (e.g., MRI vs “MRI with concussion”). Korobkin said that the ratings should be as specific as possible. The ratings would be fixed. They would not be up for debate between insurer or doctor.
Abby Moncrieff asked whether we will ever be able to create a relative value scale in many areas. There is so much uncertainty about diagnosis in first instance, in which case it seems like it will be impossible to determine relative value. Shouldn’t we be trying to push back on the idea that “more care is better”? Korobkin agreed.