Richard Thaler, co-author of Nudge, wrote an op-ed in The New York Times this weekend arguing that we should change the default option for organ donation. Reading the article helped crystallize for me a vague concern I’ve had with all this behavioral economics-inspired, benevolent-paternalistic behavior modification that has gotten so much attention lately among the smart set. But I’m getting ahead of myself.
The poster child for “changing the default option” (I’ll explain that in a minute) is 401(k) enrollment. We all know that, in general, people don’t save enough for retirement. It has something to do with overly high discounting, or fallacies in how we perceive present and future utility, or something like that, but that’s not relevant here. One problem is that not many people enroll in their 401(k) plans, even though a 401(k) plan is free money from the government, because of the tax deduction. The solution is to change the default option so that people are automatically enrolled in the 401(k) when they join the company, and some modest amount of their salary (like 2%) is automatically directed to an appropriate investment, like a stock index fund (not a money market fund, which is another bad default option). The arguments for doing this are: (a) you’re not forcing anyone into anything, since it’s easy and free to switch out of the 401(k) if you want to; and (b) it works – by changing the default option, you can get higher 401(k) participation, which we’ve posited is a good thing.
Thaler extends this argument to organ donation. 12,000-15,000 people each year end up brain dead but temporarily on life support. It’s good for society if more of them agree to become organ donors in advance. However, in most (or all?) states, you have to explicitly sign up to be an organ donor, which most people don’t get around to doing, even if they would like to. The alternative to this “opt in” system is an “opt out” system, where everyone is automatically a organ donor, but it is free and easy to opt out.
The results? “Consider the difference in consent rates between two similar countries, Austria and Germany. In Germany, which uses an opt-in system, only 12 percent give their consent; in Austria, which uses opt-out, nearly everyone (99 percent) does.”
Thaler recognizes that making everyone organ donors by default is politically impossible, so he ends up recommending a “mandated choice” system, where you have to choose one way or another. Before I move on, let me say that I agree with this. But let’s go back to the opt out system (default option = organ donor) and ask, leaving aside political issues, whether this is a good thing.
The thing that started me thinking was this paragraph:
“In the world of traditional economics, it shouldn’t matter whether you use an opt-in or opt-out system. So long as the costs of registering as a donor or a nondonor are low, the results should be similar. But many findings of behavioral economics show that tiny disparities in such rules can make a big difference.”
The general argument for changing the default option is twofold. First, you are preserving free choice; second, you are getting better outcomes. The thing that bothers me is that you are getting outcomes that people would not have chosen. The default option is getting some people to do what they don’t want to do. Let’s put this in the 401(k) context. I’m going to make up the numbers, but the principle is sound.
Let’s say that, if people thought about it carefully, 75% would enroll in their 401(k) plans. The other 25% absolutely need the money now (college debts, sick children, sick parents), or they are rationally income-smoothing over their lifetimes (taking income now because they expect their incomes to go up in the future). If the default option is not to enroll, you get only 35% participation, because people are busy and don’t think about it carefully. But if the default option is to enroll, you get 90% participation, for the exact same reason.
Now how does this look? The arguments are still pretty strong. The participation rate is closer to what it would be if people thought about the question carefully, because the default option is what a majority would choose. But now you’ve swung too far the other way. You have fewer people doing something they wouldn’t have chosen, but you’ve still tricked 15% of the people into doing something they wouldn’t have chosen.
However, you can partially justify this: some people who, on reflection, would not enroll, really should enroll. That is, part of the error in low enrollment rates is due to the default option; but part of it is due to people making mistakes about whether they should save or spend, and those mistakes are due to cognitive fallacies we all suffer from. But this is much shakier ground to be on. Protecting people from mistakes due to inattention is a form of paternalism I am perfectly comfortable with; using their inattention to protect them from logical mistakes they would have made I am less comfortable with.
This becomes more visible when we move from 401(k) plans to organ donation. There are well-meaning people who think that everyone who is eligible should be participating in a 401(k) plan – it is free money, after all. So if the default option gets more people to sign up than would otherwise, you can rationalize it by saying (a) it’s good for them anyway and (b) it’s good for society (people not saving for retirement creates an externality). With organ donation, though, (a) falls away completely; while many people (including me) don’t care what happens to their organs, some people care a lot; if you make them donors by default, not all of them will make the effort to opt out, and now you’ve tricked them into something that is bad for them. You can balance that against the social welfare created by having more organ donors, and I think that’s a legitimate position to take, but it’s not one that flies very well in our society.
To be clear, again, Thaler is arguing for mandated choice, not making everyone organ donors by default. But I think the reasons that I (and probably others) are uncomfortable with the latter apply, perhaps to a varying extent, to all of these attempts to use the default option to change people’s behavior – even when changing their behavior is in the their own interests.
By James Kwak