by Cass R. Sunstein, JD
Default rules, taken as such, do not intrude on autonomy, even if they influence people without persuading them. If default rules give people certain rights automatically (such as the right to free speech), they promote autonomy for just that reason. And to the extent that default rules give people the freedom to focus on their most pressing concerns, and thus eliminate a kind of “bandwidth tax,” they increase autonomy as well. When default rules compromise autonomy, it is not because they are default rules; it is because they invade, or take, certain rights or interests without people’s explicit consent.
This is, in short, my reaction to the interesting discussion of organ donation, default rules, and autonomy in MacKay and Robinson. They may be right to say that for organ donation, presumed consent is an intrusion on autonomy and a pro tanto wrong. But if so, the reason is not that default rules bypass people’s rational capacities. It is that unless they have given their explicit consent, people should not lose parts of their own body (even at the time of death). Whether or not this view is correct, and whether or not we embrace the underlying (Kantian) moral framework, it captures deeply held moral convictions, which help account for widespread public opposition to presumed consent to organ donation.
To orient these conclusions, let us begin with an issue that seems far afield: voter registration. In the United States, most states require people to register to vote; to be a voter, you must effectively opt in. In recent years, Oregon, California, and other states have switched the default: If the state knows that you are a resident and of voting age, you are automatically registered. If you do not want to be a voter, you may opt out.
A state could, of course, force people to choose whether to be registered (“mandatory active choosing”). It could say, for example, that unless people make that choice, they cannot obtain a driver’s license. Or a state could adopt either an opt-in or opt-out default, but also prompt people to choose, for example, with a prominent question—do you want to register to vote?—when they apply for their driver’s license (“prompted choice”).
In my view, the best approach is automatic registration. People have the rights to freedom of speech and religion by default; to obtain those rights, they do not have to register with anyone. The same should be true for voting, at least if fraud can be prevented and if the administrative burdens are not serious. Similar conclusions hold for many other rights and entitlements. For example, federal law in the United States entitles poor children to free school meals. An opt-in system, requiring their parents to sign their children up, inevitably means that numerous poor children will not receive the benefit the law gives them. If states and localities can automatically register eligible children, so that they do not even have to apply, nothing will be lost, and much will gained.
These examples show that opt-out defaults are often best; they simplify life and compromise no one’s autonomy. Human beings have limited bandwidth, and an opt-in design, mandatory active choosing, and prompted choice can tax some of that bandwidth. In its own way, that bandwidth tax is an intrusion on autonomy. For voter registration and school meals, an opt-out system is not a moral wrong in any way, and it seems morally superior to the alternatives.
Actually default rules are pervasive, and they need not threaten autonomy. On the contrary, they might promote it. Your printer has default settings. Your cell phone has default settings, and so does your computer. For some of those settings, you might have to choose (such as language and time zone), but for many others, default settings are in place (such as ring tone, do not disturb, display and brightness, and wallpaper). Your relationship with your employer or your university almost certainly has default terms of various kinds, and some of them are massively important. You might be automatically enrolled in a savings plan, and have to opt out if you do not wish to be—or you might be required to opt into a savings plan if you prefer to be enrolled. You might have to choose a health insurance plan (actively), but once you do, it will have a range of default provisions—for example, no insurance in case of long-term disability, unless you opt in.
In fact, the law of contract consists in large part of default rules; they establish what happens if the contracting parties are silent. In the United States, for example, the default rule is that employment is “at will”; employees can be fired, or leave, for any reason or for no reason at all. Other legal systems have different default rules—for example, a right to be fired only “for cause” unless the contract specifies otherwise.
In all of these cases, default rules can be characterized as nudges, or as “reason-bypassing nonargumentive influences”, but they are not ethically problematic for that reason. No plausible conception of autonomy, Kantian or otherwise, requires employers to ask you to choose whether your printer will be single-sided or double-sided, or forbids employers from automatically enrolling employees in pension plans, subject to opt-out.
But some default rules are morally problematic. To see why, suppose that the Internal Revenue Service issues a new regulation, stating that by default, taxpayers will give 1 percent of their after-tax income to the United Way, unless they designate a different charity, or opt out entirely. Or suppose that state or private employers do the same thing. In the United States and several nations in Europe, strong majorities disapprove of charitable donations by default. One reason may be a belief that it is in the nature of charity to make such donations actively rather than passively; that is what it means to be charitable. But people in diverse nations also seem drawn to a clear moral principle to this effect: If people are going to lose money, it must be as a result of their explicit consent.
Of course, this principle is not unqualified. Thieves must return what they have taken; wrongdoers must compensate those they have wronged. Most people agree that governments are allowed to raise revenue. (Perhaps the tax system can be understood as a product of collective consent, at least in its broad outlines.) But as a general rule, individual consent is necessary if the government, or a private institution, is to take resources from some people for the benefit of others.
Now turn to MacKay and Robinson’s provocative discussion of organ donation in this light. Many people favor a system in which people are presumed to be organ donors, subject to opt-out. Other people prefer a system of mandated choosing. MacKay and Robinson share the latter preference on moral grounds, at least in the sense that they believe that presumed consent, as a form of reason-bypassing, nonargumentive influence, intrudes on people’s autonomy, and therefore counts as a moral wrong, at least pro tanto.
If presumed consent is objectionable, I think it is for a more specific reason, analogous to the principle that leads people to oppose charitable contributions by default: If people are to donate their organs to others, it must be because they have actively consented to that donation. A default rule in favor of organ donation intrudes on autonomy not because of a general concern about default rules, or about reason-bypassing, nonargumentative influences, but because certain kinds of losses must occur only after an explicit expression of a person’s will.
Importantly, the requirement of explicit consent applies in numerous contexts. In a free society, people cannot be classified as Christians by default (say, for purposes of the census) or as Democrats by default (say, for purposes of voter registration). Active choosing is required in many domains. But again, the reason is not that default rules, taken as such, intrude on autonomy. It is that in particular contexts, what is necessary is an explicit indication of people’s values, wishes, and tastes. It is in those contexts—which remain of course to be specified—that active choosing and personal agency are essential.
The ultimate reason might have to do with autonomy; it might have to do with welfare. But return to my general claim: Default rules are not problematic because they are default rules. Reason-bypassing, nonargumentative influences are all around us, and they make us better off—and more autonomous to boot.