coons_c_weber_m_eds_paternalism_theory_and_practice
.pdfPaternalism in economics |
171 |
are getting what they want, or are bearing only modest costs, when in fact the policy might be an unmitigated catastrophe from their perspective. In such cases MCBA is far worse than useless, and, contrary to Sunstein,23 policy-makers may well be better advised to rely on gut instincts and hunches – say, a nagging sense that there might be something wrong with ruining people’s lives for the uncertain prospect of a marginal public gain. If nothing else, their decisions would not then bear the intimidating imprimatur of “science.”
avoiding paternalism
“The economist to whom it is no concern whether or not a family loves its farm will almost inevitably aid and abet the destruction of family farming.” (– Wendell Berry, “An Argument for Diversity”)
The question is how we can do better. What would non-paternalistic policymaking look like? It depends on what policy-makers have “reasonable opportunities” to learn concerning what people want, or would want, them to do on their behalf. This will naturally be a matter for some debate. An obvious first step, just noted, would be to abandon MCBA for more mindful forms of CBA that go beyond choice behavior in teasing out people’s preferences, for instance asking them about their wishes, as in contingent valuation surveys. As well, we can take seriously the di erences among brute preferences and values.
For example: When the local city council contemplates using eminent domain to turn people’s homes into a shopping center, CBA will have a place, for instance telling them whether the action has more costs than benefits even from a narrow market perspective. But citizens will also want to know why some people won’t sell: Is the price just too low? Are they holding out for a bigger settlement? Or are these cherished homesteads whose elderly residents can imagine no other satisfactory place to live? If the latter, then the costs of the policy to those individuals may be far higher than any market indicators could tell us. Those costs need to be noted, even if we cannot quantify them in any precise way. And how would the alleged beneficiaries of the proposal feel about having some lives ruined so they and some others can enjoy a few new retail outlets? The obvious solution, where possible, is to ask them. We might find that their convenience preferences vanish without a trace when elicited in this context.
Sometimes it is not feasible to ask, and at any rate such methods have their own troubles: Talk is cheap – a major motivator of mindless economics – and
23 Sunstein, “Cognition and Cost–Benefit Analysis.”
172 daniel m. haybron and anna alexandrova
the preferences people express in surveys are not always reliable. But one doesn’t always need to run a formal study to know something about people’s values; anyone with a modicum of sense knows that people often have profound attachments to their homes and communities, which is frequently why the question of eminent domain arises in the first place. Measures or no, there is no excuse for policy-makers to disregard obvious facts about human life.
Even where direct indicators of preference are lacking, policy-makers can employ measures of other sorts, for instance seeking information about outcomes that people are independently known to value very highly, such as happiness. Here we think happiness research can be among the means for policy-makers to avoid paternalism. You don’t need to ask your constituents, every time you decide a policy question, whether they want to be happy. You can usually take that one pretty much for granted, and indeed put a pretty heavy weight on it: When your decisions undermine the happiness of your constituents, you have a pretty good start on knowing that you did something wrong. Similarly, we don’t fret about governments fighting the spread of malaria, since that’s something people obviously want to avoid. If it is true, for instance, that economic growth often yields poor happiness dividends, and that economic growth is valued substantially as a means to happiness, this could be evidence that some growth-oriented policies stand in conflict with people’s priorities: People might prefer policies that advance happiness (such as unemployment reduction) to those that promote growth but not happiness. Knowingly to promote pecuniary goals for the public over goods that citizens are known to care more about is, surely, paternalistic. In short, when asking people directly for their opinions on the measure you’re proposing isn’t practical – as is the case for most policies – then considering information about its likely impact on things they care very much about, like happiness, may be essential to taking their priorities seriously and treating them with respect. In the case of happiness, some sources – e.g., shorter commutes – often require collective action to promote, and people’s choices often fail quite obviously to match their priorities – e.g., obesity. So consideration of the happiness impacts of policy options may be necessary for taking individuals’ values seriously. Not to consider happiness impacts in policy, where reliable enough information exists, can be paternalistic.
This may seem a peculiar thing to say, since policy applications of happiness research are very often the target of complaints about paternalism.24
24E.g., Mitchell, “Libertarian Paternalism Is an Oxymoron”; Glaeser, “Paternalism and Psychology”; White, “Behavioral Law and Economics”; Barrotta, “Why Economists Should Be Unhappy with the Economics of Happiness”; Hausman and Welch, “Debate.”
Paternalism in economics |
173 |
In some cases, the critics have a point. In thinking about paternalism, we can usefully distinguish three basic varieties of HDE. First, Benthamites argue that the promotion of mental states such as happiness or subjective wellbeing should be the sole aim and normative criterion of policy. Normative economics in turn should center on such mental states rather than preference satisfaction. This classical utilitarian position is advocated most visibly by Richard Layard25 and less visibly by Daniel Kahneman.26
Insofar as HDE is Benthamite, then paternalism is indeed a serious problem: Benthamite utilitarianism notoriously o ers no fundamental protections against paternalistic interference, restraining the state only insofar as restraint is thought to maximize happiness. If you are allowed to do what you want, get what you want, or have any say at all in how you are to lead your life, this is mere coincidence, an indulgence granted by the utilitarian despot on the grounds that you happen to live in a beneficent world where the sum of utilities is best served by permitting you to choose. You are, in this scheme, merely the steward of your life, not the sovereign master.27 That you are given a long leash does not make you any less bound.
We don’t expect many utilitarians to be impressed by this argument. And of course there are important ways in which many Benthamites reject overt forms of paternalism; some indeed deem themselves libertarians. Layard’s own proposals do not, by and large, seem overtly paternalistic, the paternalism in question merely being implicit or inattentive. The main point is to concede that at least one form of HDE may indeed be deeply paternalistic.
At the opposite end of the spectrum stand what we’ll call the happiness-aware traditionalists (HATs), who don’t think that happiness or well-being should be the sole or perhaps even a central concern of policy-making, but nevertheless treat it as an important concern. Normative economics likewise needs to trade in such notions, but without major revision to the standard framework, which might still be seen as fundamentally concerned with preference satisfaction. Arguably, most advocates of HDE and happiness policy are closer to this end of the spectrum than to the Benthamite view, and a perusal of their policy recommendations raises few red flags about paternalism – with one major family of exceptions to be noted momentarily.28
25 Layard, Happiness. |
26 Kahneman et al., “Toward National Well-Being Accounts.” |
27This is a slight exaggeration: A steward would have a special responsibility for her life, whereas the utilitarian scheme has no fundamental place for such responsibilities.
28For a few likely examples, see Diener et al., Well-Being for Public Policy; Bok, The Politics of Happiness; Frey, Happiness; Stiglitz, Sen, and Fitoussi, Report by the Commission on the Measurement of Economic Performance and Social Progress; Dolan and White, “How Can Measures
174 daniel m. haybron and anna alexandrova
HATs illustrate just how modest HDE can be. A particularly weak form would claim only that information about happiness or well-being sometimes merits consideration when evaluating policies, allowing that stringent moral principles protecting liberty might sharply limit what governments can do to promote such values. No controversial stand need be taken on the right theory of well-being, and indeed we might restrict our concern entirely to such ino ensive goals as the reduction of su ering. In fact we need not make the promotion of happiness, or even the mitigation of unhappiness, a goal of policy that new policies might be developed to pursue, taking happiness merely to be a consideration to be weighed when assessing policies initiated for other reasons. By this point, the happiness police are nowhere in sight, and it is hard to see what must be paternalistic about this sort of policy.
In short, simply counting information about happiness as a signifi- cant consideration when evaluating policies, which is all that weak forms of HDE require, need not be paternalistic, and may be an indispensable means to avoiding it. But a third variety of HDE is, for the most part, self-consciously paternalistic; we will call its proponents choice architects. According to choice architects, policy-makers should sometimes attempt to influence, by non-rational means, people’s choices to correct for common mistakes. Thus we have proposals such as “libertarian paternalism,” “light paternalism,” and “asymmetric paternalism,” which urge policy-makers to act as choice architects aiming at promoting well-being in those domains where the minimalist focus on revealed preference leads to undesirable results.29 Where we are liable to fall victim to our own cognitive biases, policy-makers may sometimes intervene and “nudge” us towards making choices that better serve our well-being, or at least our own ends. (Choice architects can of course be more aggressive than this, but most advocates press this relatively conservative line.) Everybody agrees that we would be better o if, at very little cost to our present well-being, we saved more for retirement or ate more healthily or drove a little slower. The “nudges” that push us towards behaving more prudently still leave us plenty of room to make irrational choices should we insist on them.
of Subjective Well-Being Be Used to Inform Public Policy?”; Loewenstein and Ubel, “Hedonic Adaptation and the Role of Decision and Experience Utility in Public Policy”; Forgeard et al., “Doing the Right Thing”; Fleurbaey, “The Importance of What People Care About.”
29E.g., Thaler and Sunstein, Nudge; Camerer et al., “Regulation for Conservatives”; Loewenstein and Haisley, “The Economist as Therapist”; Trout, “Paternalism and Cognitive Bias” and The Empathy Gap.
Paternalism in economics |
175 |
Choice architecture is plainly paternalistic, though in these cases soft paternalism (that is, roughly, it plausibly helps people to get what they want).30 Is this a problem? It is certainly cause for concern, and reason for caution. As with soft paternalism generally, it is not easy to formulate rules for the acceptable use of nudges, and we will not try to do so here. To some extent it is a question of costs and benefits; most people endorse even strongly paternalistic programs like Social Security because the stakes are so high, and the benefits so far in excess of the costs. Once we’ve accepted that kind of policy, it is hard to see the objection to nudging people into far more adequate retirement savings by changing the default options on their employment forms. At some point, however, agential prosthetics that benignly help people to make the choices they really want to make can lapse into rank manipulation, ceasing to treat people with respect. Whatever that point is, we can at least conclude that choice architecture need not be objectionably paternalistic, and can be less problematically paternalistic than MCBA.
While happiness will be among the considerations noted in nonpaternalistic policy-making, it will not of course be the only one. For various reasons, it is hard to determine what people really want us to do on their behalf, and there is likely no canonical means for representing that information in policy deliberations. Take the case of obesity: Few people have values that are genuinely advanced by a grossly unhealthy diet; yet people sometimes endorse such diets out of ignorance or rationalization. Do we attend to their present expressed views about their diet; their present expressed views about the importance of a long, healthy life; their views about the importance of happiness along with data about the impact of obesity on happiness; the considered judgments such people are found to give when informed and reflective; the ex post judgments people tend to make later in life; the actual choices they make (including membership in Jenny Craig . . . ); or . . . ? Arguably, all of these perspectives are relevant to assessing what manner of living best reflects their priorities. Such complexities suggest it will be hard to avoid concerns about paternalism, just as it is hard to avoid them when making decisions for a sick parent.
Avoiding, or at least minimizing, paternalism will require admitting a diverse set of indicators of what people care about into the policy-maker’s toolbox. This sort of pluralism requires abandoning minimalism in favor of what we have called high fidelity economics, so called because it o ers a fuller and more accurate picture of economic phenomena, even as it trades
30 Light paternalism can be hard paternalism if the goals advanced aren’t taken to be the agent’s own.
176 daniel m. haybron and anna alexandrova
some degree of exactness and formal elegance.31 Hi-fi economics will sometimes be less precise, and have more extensive empirical and normative commitments than minimalist economics. But it more faithfully represents the economic landscape, and better reflects the values of economic agents.
conclusion: burying the minimalist nanny state
Summing up: The central charge against minimalist economics is moral incoherence, specifically where that framework extends to the policy realm for weighing the costs and benefits of policy options. In their e orts to avoid the specter of philosopher-kings imposing alien values on the populace, minimalists have inadvertently spun a set of policy tools fit for a philosopher-king, minus the philosophy. (In a more literal sense than one might realize; the mistake here is essentially refusing to think in a serious way about the consequences of one’s rulings.) To avoid or at least reduce paternalism requires a pluralistic approach, considering a wide range of information about how policy options bear on people’s values, including indicators of happiness and other goods known to be valued. Contrary to popular complaints, happiness policy need not be paternalistic at all (though of course it can be).
A secondary claim is that minimalist paternalism is profoundly objectionable, steamrolling individuals in ways that should make even hard paternalists blanch. (Again, we do not oppose all paternalism. But this variety is beyond the pale, overriding people’s values in order to promote things they care less about, or not at all.) Notice that the incoherence charge holds even if the paternalism charge doesn’t: Paternalistic or not, refusing to take account of crucial information about people’s preferences when deciding policies on their behalf plainly contradicts any conceivable rationale for economists’ concern for optimizing preference satisfaction, ideals of “consumer” sovereignty, etc.
Again, our target is not economics, period, nor even “mainstream” economics. Yet our target, minimalism, nonetheless constitutes a large chunk of economic thought – to the extent that Princeton’s Gul and Pesendorfer define the entire discipline in minimalist terms, and Harvard’s Glaeser deems minimalism part of the “moral heart of economics.” If he’s right – which we doubt – then the moral heart of economics is deeply confused, its misguided e orts to avoid paternalism having the opposite e ect, ensnaring it in a profoundly paternalistic approach to policy that
31 Alexandrova and Haybron, “High-Fidelity Economics.”
Paternalism in economics |
177 |
systematically discounts people’s strongest preferences about their lives. So proceeding might make the equations easier to solve, but it is no way to let people’s priorities be e ective in deciding the course of their lives.
Economics has long been the object of considerable resentment in many corners of the world, particularly among factory workers, farmers, fishermen, small business owners, and others whose cherished commitments are especially vulnerable to the vagaries of policy. For decades, such people have often felt pushed around by governments and hubristic technocrats aiming to bring about a more prosperous future.32 We suspect that much of the resentment has less to do with sore losers than with a distinct sense that no one is listening; that their concerns aren’t being taken seriously, or that they are being railroaded by policies supposedly designed with their interests in mind, not that anyone bothered to ask. It is one thing to lose your livelihood, trade, land, community, and culture to the inevitable forces of change. But it is quite another to lose them to the fiat of tone-deaf administrators whose calculations could assign no price to such a ections, leaving only the market data which suggested you were actually getting a pretty good deal and would surely be happy about your lucrative new job toiling over the tar sands of some distant land.33
When the stakes are high, mindless economics is bully economics.
32On overconfidence in this regard, specifically focusing on the IMF, see Angner, “Economists as Experts.”
33The fate of many idled fishermen from Atlantic Canada, who now find employment in the tar sands of Alberta. See, e.g., Neatby, “Hard Times Sold in Vending Machines.” Thanks to Corey Katz for the example. For an illustrative complaint by a fisherman about a policy initiative in the United States, see Foster, “Here Come Catch Shares.”
chapter 9
Choice Architecture: A mechanism for improving decisions while preserving liberty?
J. S. Blumenthal-Barby
introduction
Extensive research in the social sciences has shed light on two important phenomena. One is that human decision-making is full of predictable errors and biases that often lead individuals to make choices that defeat their own ends. For example, individuals incorrectly predict what will make them happy in the future and are biased towards short-term consequences over long-term ones despite attaching more importance to their long-term goals. Let us call this the “bad choice phenomenon.” The second phenomenon is that individuals’ decisions and behaviors are powerfully shaped by their environment in logically irrelevant ways; they are impacted by the order in which options are presented, or tone in which they are presented, and even the smell in the room. Let us call this the “influence phenomenon.”
Some have argued that it is ethically defensible that the influence phenomenon be utilized to address the bad choice phenomenon.1 They propose that “choice architects” learn about the various ways to influence choices and then work to design environments, broadly construed, that promote choices that make people better off. Specifically, choice architects
1Thaler and Sunstein, Nudge. This argument, spearheaded by Thaler and Sunstein, has become extremely popular in the US and the UK. Often labeled “libertarian paternalism,” the program is being integrated into policy-making and millions of dollars are being dedicated to investigations of how these findings can improve people’s health, financial, and environmental choices. Despite its integration into policy, a fullthroated defense of the program and analysis of its limitations and boundaries has not been given, which is the aim of this paper. I would also like to comment on the terminology used in this paper. I avoid using the term “libertarian paternalism” in this paper for two reasons. First, I think the term was a bad choice because “paternalism” is by definition interfering with someone’s liberty [to make them better off ], and Thaler and Sunstein want to claim that their efforts to make people better off are not interferences with liberty. Second, avoiding provocative or misleading terminology such as “libertarian paternalism” or “nudging” (another term that has been used in this context) allows me to analyze the issues at stake directly without engaging in semantic debates. I do use “choice architecture” as a descriptive term, which I define in the Introduction section of this paper.
178
Choice Architecture: Preserving liberty? |
179 |
can influence individuals by arranging the order or tone of the presented options, offering incentives, utilizing social norms, offering vivid examples or representations, creating an anchor for thought processes, having them write down an intention or action plan, engaging them in commitment strategies, creating a default, and mapping choice consequences to something that they care about such as money.
Ethical arguments for creating choice environments that lead people to make better choices revolve around two claims: (1) It makes people better off, and (2) it does so in a way that is entirely compatible with individual liberty. The goal of my paper is to examine these two claims and to see if they hold water. I will first turn to the conceptual and normative concerns with the claim that choice architecture makes people “better off,” and then in the second half of the paper I will turn to the soundness of the claim that choice architecture is compatible with liberty.
making people better off
The claim that choice architecture can serve as a mechanism for making people better off carries several ambiguities and objections, which I will explore. I will address the following five issues: (a) What is meant by “better off ” and whether choice architects rely on a perfectionist account of well-being; (b) who exactly is the intended recipient(s) of the motivated choice architecture; (c) how to handle the concern that choice architects are subject to biases and errors that affect their ability to influence people beneficently; (d) how to handle the concern that choice architects are subject to motivational problems such as malevolence or indifference that affect their ability to influence people beneficently; and (e) whether choice architects are subject to epistemological deficiencies that affect their ability to influence people beneficently.
What does “better off” mean?
The claim that choice architecture can be used to make people better off raises the question of better off by whose or what standards? Proponents of using choice architecture often claim that the answer is: By an individual’s own standards. In other words, they reject the promotion of any objective or perfectionist standard of well-being, and instead rely on a subjective, informed desire account of well-being.2 The issue is complicated, however,
2 Ibid., 5.
180 |
j. s. blumenthal-barby |
when we (1) consider not just what proponents claim to be doing, but what they actually are doing based on examples they give, and (2) realize the difficulties in determining what will make people better off “from their own point of view.”
Turning to the first point, Gregory Mitchell notes that even though proponents of choice architecture “do not choose one particular conception of the good that the planner should try to achieve, most of their illustrative examples involve making people healthier or wealthier.”3 This is in fact true when one canvasses the various areas in which choice architecture strategies have been proposed: organ donation, saving money for retirement, promoting healthy foods, environmental conservation, increasing exercise, increasing medication compliance, smoking cessation, and increasing productivity.4 Indeed, there is a sense of an underlying perfectionist standard of good, namely health and wealth, to which the various nudges direct the masses. A perfectionist standard of the good holds that what is good for a person is fixed independently of her attitudes towards it; it promotes an ideal way of living that people can achieve to a greater or lesser extent, and allows for comparisons regarding the quantity of the good that people have.5 But it is important to realize that this sort of perfectionism is not what proponents of choice architecture have in mind. Their concerns lie not with choices in favor of goods other than health and wealth per se, but instead with choices that people would not have made had they paid full attention, possessed, understood, and unbiasedly weighed all of the relevant information, and exhibited self-control.6 Designing choice environments that promote health and wealth is justified not because health and wealth are goods independently of anyone’s attitude towards them, but because (1) the vast majority of people do prioritize them,7 and (2) health and wealth are instrumental for the achievement of most other goods, regardless of what goods people aim to achieve.
This brings us to the second complication to the choice architect’s claim to promote people’s own conceptions of “better off.” It is difficult to conclude what will make people better off, even from their own point of view. As it turns out, individuals are horribly poor assessors and predictors of what will make them better off.8 Moreover, there is often a tension
3Mitchell, “Libertarian Paternalism Is an Oxymoron,” 1261.
4For a thorough litany of examples see Thaler and Sunstein, Nudge and the blog, nudges.org.
5 Arneson, “Perfectionism and Politics,” 38. |
6 Thaler and Sunstein, Nudge, 5. |
7Note the difference between this premise and the premise that health and wealth are things that any rational person prefers more rather than less of (a Rawlsian primary goods premise).
8Gilbert and Wilson, “Miswanting.”