Paternalism in Economics1 Daniel M. Haybron and Anna Alexandrova Draft: December 26, 2011

1  Introduction   Economists are notoriously averse to paternalism. Yet the reigning methods of policy analysis in normative economics frequently counsel profoundly paternalistic policies. Or so we shall argue. We take our cue from the current debate over the use of happiness and other psychological measures of well-being in economics. The debate concerns those who take happiness to be a relevant policy consideration, employing the methods of psychology to study the impact of economic policies on well-being, and those who don’t. We will call the former approach happinessdriven economics (HDE) and the most important variety of the latter minimalism.2 Minimalists argue that economics should minimize its normative commitments by adopting a preference satisfaction theory of well-being and a decision procedure based solely on optimizing preference satisfaction; and, crucially, minimizing its use of psychological notions, relying instead on an austere methodology of revealed preference—“choice,” in a loose manner of speaking.3 Happiness-driven economics need not take happiness to be the sole, or even a central, focus of normative economics; it may see it only as one significant concern among others. But it does trade freely in psychological notions like happiness, asserting that economic policy analysis must go beyond a narrow focus on choice behavior. For convenience, our discussion centers on “happiness,” using that term loosely to refer to mental states like subjective well-being, life satisfaction, or emotional well-being, which have dominated recent discussion of well-being policy. However, the argument more broadly concerns the use of well-being indicators for policy, and the points made about happiness will generally apply to well-being, on any of the major theories of well-being. Happiness-driven economics has been widely accused of paternalism, particularly by friends of minimalism. Such critiques, we will argue, are well-founded in relation to some forms of HDE; but the objections fail regarding other varieties of HDE: there is nothing inherently paternalistic about happiness-driven economics. The question of paternalism in HDE raises broader questions, however, about the potential for paternalism in economic policy analysis. We do not deny that paternalism is sometimes justified to help people achieve what they value. And avoiding paternalism, even where it isn’t warranted, can be surprisingly difficult. But one approach that especially lends itself to deeply objectionable forms of paternalism, ironically, is minimalism. This claim will occupy the lion’s share of the paper. Doubling the irony is the fact that HDE can be an important tool for avoiding paternalism. In short, the minimalist critics of HDE have gotten things almost exactly backwards 1

For their helpful discussion and feedback on earlier versions of this paper, we wish to thank audiences at the “Freedom, Paternalism and Morality” workshop at Bowling Green State University and the “Law and Happiness” workshop at Saint Louis University, as well as John Helliwell, Neil Thin, Richard Arneson, Douglas Husak, Bonnie Wilson, and Matthew Cashen. 2 One need not be a minimalist to oppose HDE, but minimalism exemplifies the main strand of opposition within economics. As well, our chief target in this paper is the minimalist exclusion of psychological phenomena generally from economics, not just happiness. 3 Loose, because revealed preference does not distinguish between choices and other forms of behavior, including involuntary behavior; indeed, thermostats make “choices” in this sense.

1

when it comes to the avoidance of paternalism. Even where HDE genuinely offends antipaternalist sensibilities, few of the measures actually proposed for HDE even approach the extent to which minimalist policies can disregard the highest priorities of their supposed beneficiaries, indeed amounting to hard paternalism. We want to stress at the outset that our target is by no means the whole of mainstream economics. To a great degree our criticisms aim not at economists but at policymakers and agencies who have taken up principles inspired by minimalism in setting policies. Nor are we calling for a revolution in economics. For all we shall argue, introductory economics classes might proceed largely as before. Indeed, minimalist principles might remain appropriate for substantial parts of economics; we do not claim that the entire field needs to engage with psychological research. Note that plenty of economists have already abandoned minimalism, and might regard our discussion as congenial to their efforts. Yet our target, minimalism, nonetheless constitutes a huge strain of economic thought, influencing even many of those who do not fully accept it. We begin with a characterization of minimalism. Then, using a broad definition of paternalism, we examine the antipaternalist credentials of minimalism and find them wanting. Finally, we consider how policymakers might avoid, or at least minimize, paternalism, arguing that HDE should be part of a less paternalistic approach to policy analysis.4

2.  What  is  minimalism?   In standard economics minimalism takes the form of two commitments: epistemological and normative. In this paper we focus on the latter, but it will help to briefly characterize epistemological minimalism. It stresses observability, precision and parsimony.5 Observability is a requirement that applies to economists’ characterization of preferences. According to the Revealed Preference Methodology (hereafter RPM), unobservable psychological quantities, such as happiness, commitments or other mental states, should not be part of economic theory or economic explanations. Here’s a recent endorsement of RPM: Economic phenomena consist of individual choices and their aggregates and do not include hedonic values of utilities or feelings. Therefore, it is not relevant for an economic model to explore the feelings associated with economic choices. The point of revealed preference theory is to separate the theory of decision making from the analysis of emotional consequences of decisions. (Gul and Pesendorfer 2008, p. 40) And more strikingly: Populating economic models with “flesh-and-blood human beings,” was never the objective of economists. (Ibid 43) Gul and Pesendorfer provocatively embrace the term “mindless economics” for this minimalist view of their field: minimalist economics is, literally, mindless economics.

4

A couple of caveats: for convenience, we will sometimes speak loosely of policy analyses as being paternalistic, even if paternalism only emerges in their application. Second, our discussion of HDE sets aside concerns unrelated to paternalism such as the reliability of happiness measures (see Haybron 2008), and does not try to mount a general defense of HDE. 5 See Alexandrova and Haybron forthcoming for more on epistemological minimalism.

2

Normative minimalism is a set of implicit principles of welfare economics. It purports to keep value commitments to a minimum, if not to avoid them altogether, notably by orienting normative economics solely toward the satisfaction of preferences, and thus (ostensibly) deferring to individuals’ own value judgments. Normative minimalists usually adopt an actual preference satisfaction theory of well-being and welfarism, the view that only well-being is morally significant. These assumptions are not an explicit part of economic theory and there are certainly economists who do not endorse them. But they are so common as to be definitive of standard welfare economics; or, as Edward Glaeser recently put it, they constitute “the moral heart of economics”: Improvements in welfare occur when there are improvements in utility, and those occur only when an individual gets an option that wasn’t previously available. We typically prove that someone’s welfare has increased when the person has an increased set of choices. When we make that assumption (which is hotly contested by some people, especially psychologists), we essentially assume that the fundamental objective of public policy is to increase freedom of choice.6 Of course, providing policy recommendations requires more than the preference theory of wellbeing and welfarism. We also need principles for aggregating individual preferences at a societal level. Here, normative minimalists have two procedures to offer: the Pareto Principle and CostBenefit Analysis (hereafter CBA). Since the Pareto Principle has a very narrow range of application (that is, only to cases where there are no losers from a proposed policy), we shall concentrate on CBA. In its classic form, CBA is a minimalist procedure par excellence. It requires only one extra step beyond Pareto efficiency: an outcome favored by CBA is a potential Pareto improvement if the losers under the proposed policy could be compensated by the winners, a condition known as Kaldor-Hicks efficiency. CBA attempts to provide a decision criterion without relying on any interpersonal comparisons of psychological states, without incorporating moral and political values, without engaging in any deliberation about the relative merits of the policy proposals in question and often without even actually talking to or otherwise communicating with the affected parties. Instead, CBA proposes that we canvass actual raw preferences for a policy, often using nothing more than past choices in similar situations as the indicators of preferences. Once the preferences of all affected parties are represented by the present monetary value and aggregated (without consideration of their comparative wealth or any other morally relevant factors), the benefits of adopting a project must be compared to its costs. This standard form of CBA has elicited much criticism and many proposed fixes for its shortcomings.7 Nevertheless, in its classic formulation CBA retains a powerful hold on public policy discussions and it is important to understand exactly where it goes wrong. So here our target is the sort of CBA that eschews all values except as they appear in preferences and takes preferences to be fully revealed by choices. Let us call this Minimalist CBA or MCBA. Here we 6

New York Times, January 25th, 2011. Online at: http://economix.blogs.nytimes.com/2011/01/25/the-moral-heart-ofeconomics/?scp=3&sq=glaeser&st=cse 7 See Hausman and McPherson 2006 for a summary of the criticisms, and papers from a 2000 issue of the Journal of Legal Studies, collected in Adler and Posner 2001, for proposed fixes to classic CBA (Sunstein 2000) or the impossibility of such fixes (Nussbaum 2000, Richardson 2000). Adler and Posner 2006, Orr 2007 and Schmidtz 2001 provide qualified defenses of CBA.

3

are concerned with one problem that, as far as we can tell, has been overlooked by the critics of MCBA. It can override values and commitments that people hold dear, purportedly with the aim of furthering their interests or other goals – a patently paternalistic implication. But before we justify this claim, we will take a brief foray into definitions of paternalism.

3.  What  is  paternalism?   When faced with charges of paternalism, a common strategy is to adopt a narrow definition of paternalism on which the charges won’t stick.8 We take the opposite tack, adopting a broad motive-based conception of paternalism—broader even than the definitions used by critics of happiness-driven economics (Hausman and Welch 2009). Our conception of paternalism is a modified version of Seana Shiffrin’s influential analysis (Shiffrin 2000, p. 218). Intuitively, the paternalist asserts some degree of control over an agent’s own affairs. But she does so, not merely for her own purposes or out of sadism, which is why a bully or an oppressor need not be paternalistic. Rather, a paternalist is primarily concerned with managing your affairs with the aim of getting them right. And the wrong of paternalism, when wrong, is that it involves usurping an agent’s authority to manage her own affairs. Let’s make this more precise. According to Shiffrin, paternalism by A toward B may be characterized as behavior (whether through action or through omission) (a) aimed to have (or to avoid) an effect on B or her sphere of legitimate agency (b) that involves the substitution of A’s judgment or agency for B’s (c) directed at B’s own interests or matters that legitimately lie within B’s control (d) undertaken on the grounds that compared to B’s judgment or agency with respect to those interests or other matters, A regards her judgment or agency to be (or as likely to be), in some respect, superior to B’s. This account of paternalism has some interesting features. First, it does not require restricting or otherwise actively interfering with B’s choices, still less coercion. For example, people sometimes prefer having fewer choices to more, so policies that disregard such preferences to foist more options on those individuals, for their own good, would be paternalistic. Second, paternalism need not be contrary to B’s judgment, as it may instead circumvent B’s agency in carrying out her judgment. You might want to quit smoking but lack the willpower; my hiding your cigarettes doesn’t override your judgment yet is paternalist, because I’m preventing you from managing your own affairs by circumventing your agency. Thirdly, Shiffrin’s account does not require paternalism to be aimed at the agent’s well-being. Indeed, it needn’t be concerned with anyone’s welfare: forcing me to be more virtuous, talented, skilled, well-rounded, beautiful, or otherwise excellent can quite obviously be paternalistic, even if there’s no thought of me or anyone else benefiting. We agree with Shiffrin on these points, but propose three modifications to her account. First, it is not clearly necessary, pace (b), that A’s judgment or agency be substituted for B’s: perhaps A designates a mechanical process, bureaucracy or panel of experts to decide for B. She might even regard her own judgment as worse than B’s. We should say instead that A acts paternalistically by substituting some other entity’s judgment or agency for B’s (including, perhaps, her own). 8

E.g., arguably, Trout 2005.

4

The second modification concerns condition (d), which seems too narrow: A need not regard her (or any other) judgment or agency as superior to B’s. Perhaps A holds a high opinion of B’s judgment, but doesn’t bother to find out what B really wants before acting on B’s behalf (indeed, perhaps she does so precisely because her high opinion of B causes her to presume agreement between them): she just assumes that B wants something, and secures it for her. A pushy caregiver for a wheelchair-bound sibling, say, might regularly if inadvertently override his sister’s preferences, for instance making travel plans for her without getting her opinion, which he simply assumes will mirror his. (“Of course she wants to go to Orlando! Who wouldn’t?”) In such a case A acts paternalistically toward B, not out of any presumed superiority, but simply out of insensitivity or inattentiveness to B’s attitudes. Call this inattentive paternalism. (Or, alternatively, implicit, versus overt, paternalism.) What such cases have in common with Shiffrin’s condition (d), we think, is a lack of deference to B’s judgment or agency. Meddling (say), out of either presumed superiority or insensitivity, is paternalistic because in both cases it reflects a failure or unwillingness to defer to B’s judgment or agency in matters over which B has legitimate authority. Accordingly, we propose to replace condition (d) with: (d’) manifests a non-deferential attitude to B’s judgment or agency with respect to those interests or matters. A “non-deferential attitude” in turn involves an unwillingness to defer to B’s judgment or agency, or forgoing reasonable opportunities to defer to B’s judgment or agency. An “unwillingness to defer,” as we understand it, can include refusing to acknowledge B’s authority to make the ultimate decision, even while granting B’s wish. Third, it is not clear that the present account, with b’ and d’, adequately distinguishes between paternalism and ordinary self-serving bullying or oppression. Condition (c) specifies that the behavior must be “directed” at B’s interests or affairs, but that is consistent with aiming to harm B. We can amend this to say that the behavior is done for the sake of B’s interests or other affairs: there is an intention to get them right. So we arrive at our final characterization of paternalism (with aforementioned changes to condition (b) as well): Paternalism by A toward B is behavior (whether through action or through omission) (a) aimed to have (or to avoid) an effect on B or her sphere of legitimate agency (b’) that involves the substitution of some other entity’s judgment or agency for B’s (c’) done (or omitted) for the sake of B’s own interests or matters that legitimately lie within B’s control (d’) manifests a non-deferential attitude to B’s judgment or agency with respect to those interests or matters. We think this account captures the intuitive notion of paternalism reasonably well, and highlights what is morally significant about paternalism. Further tweaking may be needed to handle other problem cases, so we consider this definition provisional. But we think it is close enough for the purposes of our argument.

4.  Is  minimalism  paternalistic?   We take no stand on agricultural policy. But farmers make a useful case for illustrating our points, both because they and other rural people are unusually at risk of suffering the defi-

5

ciencies of minimalism that we note here, and because the problem has been eloquently voiced by farmers themselves, for instance Wendell Berry. Let’s start with an example. In Costa Rica some years ago, a hotel developer offered Miguel Sanchez, a subsistence farmer, $600,000 for his land. He refused, later explaining, “I have lived in this forest since I was born. I have no desire to live anywhere else. Money can be evil. People will ask me for money, and then they might not wish to pay me back. Where I once had friends, I would have enemies.” And the forest, he added, matters more than the money.9 And thus a subsistence farmer declines the opportunity to become rich. Consider a different sort of case. People in a community may want a convenient new shopping center nearby. But they may also be strongly committed to individual property rights; they would be appalled to see the government invoke eminent domain to forcibly remove people from their homes, thus clearing the way for a shopping center. They would be more horrified still to learn that this action was taken on their behalf, to satisfy their preferences. Their moral commitments not only cause them to oppose eminent domain for trivial purposes such as this, but to regard their convenience preferences as having no weight at all in deciding whether to seize their neighbors’ homes. From their perspective, their wish for a new Starbucks offers not the slightest reason even to contemplate tossing people from their homes. In cases like these, values, moral or prudential, silence other preferences, neutralizing them for deliberative purposes.10 Silencing can come in various forms and degrees, but for now it suffices to note that such values—call them value commitments—are not merely strong preferences; they are strong preferences that serve as constraints on the satisfaction of other preferences. Note that whether people are reasonable in having such values is beside the point: what matters for our purposes is that people do, or could, have preferences like this. And that much should be plain enough. Indeed, the whole point of positing moral rights, as virtually everyone does, is to place constraints on the satisfaction of people’s preferences. Constraints are central to Kantian and other deontological theories, of which MacIntyre once wrote, “For many people who have never heard of philosophy, let alone Kant, morality is roughly what Kant said it was,” and there’s good empirical data to back this up.11 And while rights tend not to play any fundamental role in consequentialist ethics, we cannot think of any living consequentialist philosopher who denies a role for rights, or related constraints, in actual deliberative practice. (No one thinks judges ought to decide cases solely in terms of utility maximization.) Commonsense morality, and philosophical ethics as well, are steeped in constraints. Constraints also arise in people’s conceptions of well-being. Those holding Aristotelian or Stoic views of well-being, for instance, may regard virtue or excellence as incomparably greater in importance for their lives than other goods, such as comfort. They may be unwilling to trade any amount of excellence for any amount of gain in comfort, or be unwilling to do so below some threshold of excellence, or above some threshold of comfort. Indeed, probably most people have prudential values like this, holding commitments they see as central to their identities, integrity, and well-being. To compromise those values would be to compromise their own 9

From Andre Carothers, “Letter from Costa Rica,” E Magazine, September 1993. The notion of silencing derives from John McDowell’s writings, e.g. McDowell 1979. We do not mean to endorse his views about silencing, and understand the notion loosely here. It may ultimately be more useful to frame the point in other terms, e.g. “exclusionary reasons” (Raz 1999). We use ‘prudential’ to denote the sort of value involved in well-being. 11 MacIntyre 1998, p. 190. For empirical discussion, see, e.g., Baron and Spranca 1997, Tetlock 2003, Cushman, Young, et al. 2010. 10

6

welfare. Many farmers for instance are, like Miguel Sanchez, committed to various values such as not “robbing the land,” taking good care of their animals, and for that matter leading the life of a farmer.12 They may also cherish the freedom of self-employment, having to answer to no one. These are, for them, a matter of integrity and essential for their well-being. They may be unwilling even to contemplate trading off these values against mere preferences, say for extra money or consumer goods, or would do so only in limited circumstances. Their value commitments serve as constraints on the fulfillment of other, ordinary preferences in the pursuit of personal wellbeing. Some such values (“treat my livestock well”) may also be moral values, but others (“be a farmer”) may be wholly prudential: they might matter to the person simply as aspects of personal well-being, or as elements of a meaningful life. In itself this preference structure might pose no problems for CBA, since (setting aside mistakes) it will presumably inform people’s preference orderings. Deep trouble arises, however, when we try to evaluate actual policy options, for the hierarchy of preferences can make it very hard to impute preferences to people. MCBA in particular has to make do with an austere diet of observed choices, considering only the overt behavior involved and not the reasoning behind it. Values collapse into mere preferences, and constraints vanish from the scene altogether. Returning to the farmer case, MCBA treats farmers’ values as mere preferences, imputing monetary equivalents for each of them, say by looking at how much money is needed to get farmers to quit the land and give up farming. Suppose there is some amount of money for which farmers in a region have been observed to be willing to sell off their land. It may seem we can infer that they value their property to that extent, but the appearance may be misleading. Perhaps—indeed, this may be the standard case—they accept the trade purely to protect other cherished values; for example, you sell your land during hard times to save your family from ruin. They may not generally be willing to make such tradeoffs at all; a secure farmer, like Sanchez, might not accept any amount of money to sell his land, because mere commodities (mere wants) aren’t comparable at all to his core values, which serve to constrain the pursuit of commodities. And so, when a Sanchez does sell his land, it could be for a song—whatever it takes to protect his other cherished values. Similarly, that some impoverished Indian parents are willing to sell their children into slavery for $12 doesn’t mean they don’t view their offspring as priceless; perhaps the amount is just enough to keep their other children from starving. Market prices and WTP/WTA assessments in these sorts of “Sophie’s Choice” cases yield approximately zero information about people’s preferences. This is probably not fantastical case-mongering: governments have not infrequently undertaken policies that drive farmers from their land—fishermen from their boats, etc.—to join the wage economy. Mexico, for instance, allegedly set out to modernize its economy in recent decades, adopting policies (such as NAFTA) meant to move small farmers into manufacturing jobs. A great many have indeed left the farm as a result.13 The subsequent earnings of those who actually manage to get such jobs might often exceed the amounts for which they were—often under duress—willing to sell their property, creating the illusion of a benefit where, quite possibly, many see the move as a life-shattering disaster that no amount of money can compensate. In such cases, policies are imposed on people, for their sake, that they would have judged abhorrent had anyone bothered to ask. Even when officials correctly discern that those individuals are los12

On “robbing the land” and kindred values, see Berry 1990. See, e.g., Mohanty 2008, and a brief but illustrative discussion by development researcher Timothy A. Wise, “Small-Scale Farmers and Development: Assume a different economic model” at http://triplecrisis.com/small-scalefarmers-and-development/. 13

7

ing out, the policy may still fail grossly to take seriously the depth of their opposition. “We regret your loss, but assure you that we gave your concerns full consideration”—when in fact the only concerns they registered were the prices some people had been observed to accept in extremis, which are indicative of basically nothing except that there may be even worse fates than accepting such a bargain. To treat people in this manner is, we contend, paternalistic. The same problem arises in a different way in the eminent domain case, where purely moral values are at stake. Here, MCBA gives the preferences for the shopping center full weight. Yet most of the policy’s supposed beneficiaries might vehemently object to having their convenience preferences counted as a reason for evicting their neighbors from their homes. As a result, it is entirely possible that most citizens, including those who supposedly benefit from the resulting policy, would reject the measure. This would not be a particularly democratic procedure. In any event, people’s preferences are being used, on their behalf, in ways they strenuously oppose. And so their judgments about how to promote their own interests end up being overridden by policymakers. It is rather like a gangster who fulfills his wife’s wish for a diamond ring by killing a bystander, over her protests, and handing the victim’s ring to his disgusted companion, congratulating himself on his generosity. This sort of treatment is, we submit, paternalism. It may be objected that these cases, however morally problematic, do not amount to paternalism, since policymakers have no intention of overriding people’s preferences. They might firmly believe they’re carrying out people’s wishes. Given their notorious aversion to paternalism, it is safe to say that minimalist economists would generally fit this description. Consumer sovereignty is the name of the game, and calling such economists paternalists can seem like calling the Pope an atheist. We agree that minimalists’ explicit intentions are non-paternalistic. The problem is what they leave out: having accepted an oversimplified model of human preference, minimalists fail seriously to consider how people actually think about their lives. Put more concretely, a policymaker who settles on policies with deeply disruptive consequences for people’s lives, or which conflict with people’s deeply held values, and does so while seeking no more information about people’s attitudes than MCBA can provide, is almost certainly forgoing reasonable opportunities to see what the interested parties think about the matter. This is what we called inattentive paternalism: paternalism through being insufficiently attentive to people’s attitudes concerning their own affairs. And the structure of the mistake is like the well-meaning but insensitive sibling: “Of course they’re better off! They’re making more money than ever before.” In fact our cases involve paternalism in multiple forms: take the farmer case again, but imagine as well that the policy is meant to benefit, not just the farmers, but also consumers who will enjoy lower prices. In this situation we may have threefold paternalism: first, the committed farmers forced off their land, whose prudential values are discounted; second, the farmers who gladly quit their land, but who benefit only coincidentally from the policy—their values are still ignored or discounted; third, the consumers whose moral values would decisively rule out treating farmers that way—they would much rather deal with higher prices than have farmers forced off their ancestral lands. In each case, the policy aims to account for people’s interests without taking their view of the matter seriously. Conceivably, policymakers relying solely on choice information might settle on a policy that every last one of the stakeholders vigorously opposes: neither the “winners” nor the losers want it. Presumably, at least some of that information is available to them, if only they are willing to look beyond choice data. When a tin-eared government imposes uniformly despised policies on their citizens, for their own good, because they refused seriously to consider what the concerned parties actually want, that sounds like paternalism to us. Note, pace the second case, that even giving people what they want can be paternalistic, if

8

it is done without serious regard to what they think. A popular and benevolent monarch, who takes himself not to be bound by the people’s will, treats his subjects paternalistically even if they get everything they want. MCBA-based policy threatens paternalism not just for farmers and fishers, but—at least—for those subject to any policy with highly disruptive effects on people’s lives: unemployment policy, climate change policy, major public works like dams, etc. Climate change policy poses exceptionally grave risks of paternalism, since the value commitments of essentially all human beings, for centuries to come, may be at stake. Simply to weigh GDP impacts of policy options likely means ignoring most of the costs that people really care about. Intelligent minimalists, of course, will want to count nonmonetary costs too, but their instruments aren’t remotely up to the job. Mere observation of choice behavior in past circumstances, particularly the circumstances of a tragic commons, is unlikely to tell us very much about how people’s values bear on the prospect of an unprecedented, centuries-long global catastrophe. The average person has probably done very little to help preserve coral reefs, but from that fact we can hardly infer that they won’t much mind leaving a reef-free planet for their children. It is difficult to imagine a more urgent demand on policymakers deciding on public works than that the magnitude of such costs to the affected individuals be responsibly accounted for. Because such policies have such profound impacts on people’s lives, they will invariably infringe on individuals’ value commitments. Even where the policies would gain people’s assent, MCBA’s deafness to their views means that MCBA-driven policies of this sort are always, as a matter of fact, paternalistic. We have here not the mere possibility of paternalism, but, plausibly, a pervasive actuality. Moreover, the paternalism involved is hard paternalism: though no overt coercion or even limitation of freedom need be involved, this sort of paternalism does not, like soft paternalism, help people achieve what they value.14 Rather, it substitutes different ends for them altogether. While the permissibility of soft paternalism is a matter of dispute, hard paternalism is deeply unpopular among liberals, and most certainly economists. In fact the rejection of hard paternalism may be a defining feature of liberalism. And hard paternalism in the service of values that no one cares very much about must have a very small constituency indeed. That the unwitting endorsement of hard paternalism, possibly on a massive scale, should be a prominent feature of mainstream economic thought is no small irony. After all, a commitment to promoting freedom, including sharp restraints on paternalism, arguably forms the moral core of mainstream economics. A large swath of contemporary economics thus is profoundly morally incoherent.

5.  On  the  distinction  between  values  and  preferences   We are by no means the first to observe that the difference between values and mere preferences makes trouble for economics. A number of philosophers have drawn a distinction between values and preferences.15 This distinction has been used to criticize CBA for its inability to recognize and respect two very different forms of valuation. It is this inability that motivates Anderson’s claim that the “norms of consumers’ sovereignty amount to a tyranny over citizens

14

Save per accidens, as for the farmers who happen to get what they want, even though their values were not taken seriously. 15 E.g., Sagoff 1986; Anderson 1995; Sen 1977.

9

when applied to the domain of public policy” (210). Though Anderson does not frame her objection in terms of paternalism, the worry resembles those raised here. Some have attacked the distinction between values and preferences as insufficiently sharp and unmotivated (Orr 2007). Our argument need not assume a hard or deep distinction. We might, for instance, simply distinguish different levels of preferences, for example higher- and lower-order (e.g., Sen 1977). Importantly, the basic problem of paternalism arises even if we treat all preferences as brute preferences, for the well-known reason that, on any remotely plausible notion of preference, choices radically underdetermine preferences: only a small fraction of our preferences are actually revealed through choice, so inferring preferences from choice behavior will always be risky. (As Richard Arneson helpfully put it to us, revealed preference information is highly “gappy.”) When oft-traded market goods are at stake, the risks are typically minimal: your shopping behavior may well give pretty reliable evidence about your preferences for apples. But when talking about Very Strong Preferences, to say nothing of values, things are different, particularly if those preferences are rarely tested in the crucible of the market, and even then only under duress. Perhaps Sanchez’s neighbors sold their lands for a paltry sum; but from that fact we can infer almost nothing about their preference orderings, beyond the bare fact that something made the deal worthwhile to them at that time. Maybe they would normally refuse any amount of money, like Sanchez, but the baby needed an operation. To assign valuations based solely on observed choices in cases like these would be to fail to take seriously those individuals’ preferences. It would be paternalistic, at least if there is some better way to take account of their preferences. We don’t need to posit values at all to make that point. That said, we think it important to acknowledge the existence of values beyond brute preferences. Even if our best normative theory did not ultimately require any preference/value distinction, and even if any such distinction were obscure, the fact remains that people do make distinctions of that sort. You might think value commitments, “protected values” or “sacred values” obscure and unfounded, and for that reason leave them out of your normative theory. But mountains of evidence attest that people do have such values, for better or worse, and are very strongly committed to honoring them. However benighted such folk might be, it is paternalistic to disregard the structure of their preferences—viz., their values—when setting policies on their behalf. The problems raised here do not indicate that MCBA is incomplete, needing supplements from other kinds of information; they tell us that MCBA gets the wrong answers about what people want and hence about the welfare or utility impacts of policy options. It gets wrong precisely the sort of information welfare economics is supposed, by its own lights, to be about. Would that the deliverances of MCBA were merely incomplete; on the contrary, it’s bad information. In cases of paternalism, it tells us people 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 2000, policymakers 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.”

6.  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.

10

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 policymakers 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 differences 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 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 policymakers to disregard obvious facts about human life. Even where direct indicators of preference are lacking, policymakers 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 policymakers 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

11

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.16 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 well-being 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 Layard (Layard 2005) and less visibly by Daniel Kahneman (Kahneman et al. 2004). Insofar as HDE is Benthamite, then paternalism is indeed a serious problem: Benthamite utilitarianism notoriously offers 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.17 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 policymaking, 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.18 HAT illustrates 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 inoffensive goals as the reduction of suffering. 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 16

E.g., Mitchell 2004, Glaeser 2006, White 2010, Barrotta 2008, Hausman and Welch 2009. This 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. 18 For a few likely examples, see Diener, Lucas, et al. 2009, Bok 2010, Frey 2008, Stiglitz, Sen, et al. 2009, Dolan and White 2007, Loewenstein and Ubel 2008, Forgeard, Jayawickreme, et al. 2011, Fleurbaey 2011. 17

12

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 significant consideration when evaluating policies, which is all that weak forms of HDE require, need not be paternalistic, and may be an indispensible 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, policymakers should sometimes attempt to influence, by nonrational 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.19 Where we are liable to fall victim to our own cognitive biases, policy-makers may sometimes intervene and “nudge” us toward 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 off 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 toward behaving more prudently still leave us plenty of room to make irrational choices should we insist on them. Choice architecture is plainly paternalistic, though in these cases soft paternalism (that is, roughly, it plausibly helps people to get what they want).20 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 policymaking, 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. 19

E.g., Thaler and Sunstein 2008, Camerer, Issacharoff, et al. 2003, Loewenstein and Haisley 2008, Trout 2005, Trout 2009. 20 Light paternalism can be hard paternalism if the goals advanced aren’t taken to be the agent’s own.

13

Avoiding, or at least minimizing, paternalism will require admitting a diverse set of indicators of what people care about into the policymaker’s toolbox. This sort of pluralism requires abandoning minimalism in favor of what we have called high fidelity economics, so called because it is offers a fuller and more accurate picture of economic phenomena, even as it trades some degree of exactness and formal elegance.21 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.

7.  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 efforts 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 efforts to avoid paternalism having the opposite effect, ensnaring it in a profoundly paternalistic approach to policy that 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 effective 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, fishers, 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.22 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 be21 22

Alexandrova and Haybron forthcoming. On overconfidence in this regard, specifically focusing on the IMF, see Angner 2006.

14

ing 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 affections, 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.23 When the stakes are high, mindless economics is bully economics.

  References   Adler, M. and E. Posner, eds. (2001). Cost-Benefit Analysis: Economic, Philosophical, and Legal Perspectives. Chicago: University of Chicago Press. Adler, M. D. and E. A. Posner (2006). New foundations of cost-benefit analysis. Cambridge, Mass.: Harvard Univ Pr. Alexandrova, A. and D. M. Haybron (forthcoming). “High Fidelity Economics.” In The Elgar Companion to Recent Economic Methodology, eds. W. Hands and J. Davis. Northampton, Mass.: Edward Elgar. Anderson, E. (1995). Value in ethics and economics. Cambridge, Mass.: Harvard Univ Pr. Angner, E. (2006). “Economists as Experts: Overconfidence in theory and practice.” Journal of Economic Methodology 13(1), pp. 1-24. Baron, J. and M. Spranca (1997). “Protected values.” Organizational Behavior and Human Decision Processes 70(15520873570931760914related:Et8il0c4ZdcJ), pp. 1-16. Barrotta, P. (2008). “Why economists should be unhappy with the economics of happiness.” Economics and Philosophy 24(02), pp. 145-165. Berry, W. (1990). “An Argument for Diversity.” In What Are People For?, ed. W. Berry. New York: North Point Press. Bok, D. (2010). The Politics of Happiness: What Government Can Learn from the New Research on Well-Being. Princeton, NJ: Princeton. Camerer, C., S. Issacharoff, et al. (2003). “Regulation for Conservatives: Behavioral Economics and the Case for 'Asymmetric Paternalism'.” University of Pennsylvania Law Review 151, pp. 1211-1254. Cushman, F., L. Young, et al. (2010). “Multi-System Moral Psychology.” In The Moral Psychology Handbook, ed. J. Doris. New York: Oxford. Diener, E., R. E. Lucas, et al. (2009). Well-Being for Public Policy. New York: Oxford University Press. Dolan, P. and M. P. White (2007). “How can measures of subjective well-being be used to inform public policy?” Perspectives on Psychological Science 2(1), pp. 71-85. 23

The fate of many idled fishermen from Atlantic Canada, who now find employment in the tar sands of Alberta. See, e.g., Stuart Neatby, “Hard Times Sold in Vending Machines: Worker migration from Atlantic Canada to the tar sands,” The Dominion, Issue 48, Autumn 2007. Thanks to Corey Katz for the example. For an illustrative complaint by a fisherman about a policy initiative in the United States, see Ernie Foster, “Here come catch shares: How NOAA and the Environmental Defense Fund plan to destroy North Carolina’s working watermen,” Island Free Press, November 10, 2010.

15

Fleurbaey, M. (2011). “The importance of what people care about.” Politics, Philosophy and Economics: Happiness. (New Orleans. Forgeard, M., E. Jayawickreme, et al. (2011). “Doing the Right Thing: Measuring Well-Being for Public Policy.” International Journal of Wellbeing 1(1), pp. 1-28. Frey, B. S. (2008). Happiness: A Revolution in Economics. Cambridge: MIT Press. Glaeser, E. L. (2006). “Paternalism and Psychology.” University of Chicago Law Review 73(1), pp. 133-156. Gul, F. and W. Pesendorfer (2008). “The case for mindless economics.” The foundations of Positive and normative Economics: A handbook, pp. 3–42. Hausman, D. M. and M. S. McPherson (2006). Economic analysis, moral philosophy, and public policy. New York: Cambridge University Press. Hausman, D. M. and B. Welch (2009). “Debate: To Nudge or Not to Nudge.” Journal of Political Philosophy 18(1), pp. 123-136. Haybron, D. M. (2008). The Pursuit of Unhappiness: The Elusive Psychology of Well-Being. New York: Oxford University Press. Kahneman, D, A. Krueger, D. Schkade, N. Schwartz & A. Stone. 2004. "Toward National WellBeing Accounts". American Economic Review (May) 429-434. Layard, R. (2005). Happiness: Lessons from a new science. New York: Penguin. Loewenstein, G. and E. Haisley (2008). “The Economist as Therapist: Methodological Ramifications of "Light" Paternalism.” In The Foundations of Positive and Normative Economics, eds. A. Caplin and A. Schotter. New York: Oxford University Press, pp. 210248. Loewenstein, G. and P. A. Ubel (2008). “Hedonic adaptation and the role of decision and experience utility in public policy ” Journal of Public Economics 92(8-9), pp. 1795-1810. MacIntyre, A. (1998). A short history of ethics: a history of moral philosophy from the Homeric age to the twentieth century. New York: Routledge. McDowell, J. (1979). “Virtue and Reason.” The Monist 62, pp. 331-50. Mitchell, G. (2004). “Libertarian paternalism is an oxymoron.” Nw. UL Rev. 99, pp. 1245. Mohanty, M. (2008). “Small farmers and the Doha Round: lessons from Mexico's NAFTA experience.” La Chronique des Amériques (Juillet 2008 No 12). Nussbaum, M. C. (2000). “The Costs of Tragedy: Some Moral Limits of Cost-Benefit Analysis.” The Journal of Legal Studies 29(s2), pp. 1005-1036. Orr, S. W. (2007). “Values, preferences, and the citizen-consumer distinction in cost-benefit analysis.” Politics, Philosophy & Economics 6(1), pp. 107. Raz, J. (1999). Practical reason and norms. New York: Oxford University Press. Richardson, H. S. (2000). “The Stupidity of the Cost-Benefit Standard.” The Journal of Legal Studies 29(s2), pp. 971-1003. Sagoff, M. (1986). “Values and preferences.” Ethics 96(2), pp. 301-316. Schmidtz, D. (2001). “A place for cost-benefit analysis.” Philosophical Issues 11, pp. 148. Sen, A. K. (1977). “Rational fools: A critique of the behavioral foundations of economic theory.” Philosophy and Public Affairs 6(4), pp. 317-344. Shiffrin, S. (2000). “Paternalism, unconscionability doctrine, and accommodation.” Philosophy & Public Affairs 29(16275849302651892920related:uPRaJrVu3-EJ), pp. 205-250. Stiglitz, J. E., A. Sen, et al. (2009). “Report by the Commission on the Measurement of Economic Performance and Social Progress.

16

Sunstein, C. R. (2000). “Cognition and Cost-Benefit Analysis.” The Journal of Legal Studies 29(s2), pp. 1059-1103. Tetlock, P. (2003). “Thinking the unthinkable: Sacred values and taboo cognitions.” Trends in Cognitive Sciences 7(10065375210518430634related:qkP54Wxlr4sJ), pp. 320-324. Thaler, R. H. and C. R. Sunstein (2008). Nudge : improving decisions about health, wealth, and happiness. New Haven: Yale University Press. Trout, J. D. (2005). “Paternalism and Cognitive Bias.” Law and Philosophy 24, pp. 393-434. Trout, J. D. (2009). The empathy gap: Building bridges to the good life and the good society. New York: Viking Press. White, M. D. (2010). “Behavioral Law and Economics: The Assault on Consent, Will, and Dignity.” In Essays on Philosophy, Politics & Economics: Integration & Common Research Projects, eds. G. Gaus, C. Favor and J. Lamont. Stanford, CA: Stanford University Press.

17