Show Summary Details

Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE ( © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

date: 25 March 2019

Adam Smith’s Libertarian Paternalism

Abstract and Keywords

Richard Thaler and Cass Sunstein argue for “libertarian paternalism,” defined as the strategy to devise policy that will “maintain or increase freedom of choice” and at the same time “influence people’s behavior in order to make their lives longer, healthier, and better”. These two goals are often in conflict, and striking the right balance between them has proved difficult in both theory and practice. Where does Adam Smith fall in this debate? This chapter argues that Smith developed his own version of “libertarian paternalism.” It differs in important ways from that of Thaler and Sunstein, but it shares with them an attempt to balance respect for individual autonomy with a desire to help people lead better lives. Smith’s position accommodates the importance of both liberty and paternalism in enabling individuals to construct lives worth living, while avoiding some of the problems that have beset more recent versions of libertarian paternalism.

Keywords: Adam Smith, libertarian paternalism, autonomy, choice, Richard Thaler, Cass Sunstein

1. Introduction

In their influential book Nudge, Richard Thaler and Cass Sunstein argue for what they call “libertarian paternalism,” which they define as the strategy to devise policy that will “maintain or increase freedom of choice” (the libertarian part) and at the same time will “influence people’s behavior in order to make their lives longer, healthier, and better” (the paternalistic part).1 Their goal is to help people make the decisions they would have made “if they had paid full attention and possessed complete information, unlimited cognitive abilities, and complete self-control” (Thaler and Sunstein, 2009: 5). This seems like a high standard, and as a result some critics have charged them with erring rather on the side of paternalism than on that of libertarianism,2 but they insist that their intention is only “to influence choices in a way that will make choosers better off, as judged by themselves” (ibid.; emphasis in original).

The desires to allow people freedom of choice, on the one hand, and to help them make better choices, on the other, are often in conflict, and striking the right balance between policies designed to encourage the former or the latter has proved difficult in both theory and practice. John Stuart Mill, perhaps the leading classical proponent of the libertarian position, argued that “the sole end for which mankind are warranted, individually or collectively, in interfering with the liberty of action of any of their number is self-protection” (Mill, 1978 [1859]: 9). Yet even Mill felt the necessity of limiting that principle: in addition to children, Mill argued that “we may leave out of consideration those backward states of society in which the race itself may be considered as in its nonage” (Mill, 1978 [1859]: 9–10). Some more recent critics of the Millian libertarian position have built on insights from Thaler and Sunstein to argue that policy should abandon the libertarian aspect altogether and instead embrace a fully paternalistic “choice architect” position.3

(p. 240) Where does Adam Smith fall in this debate? Although Smith did not explicitly develop a theory of freedom, nevertheless we can reconstruct a conception of political freedom and a conception of what it means to be a free person by working backward from some of the specific claims he makes about moral agency and from some of his policy recommendations. This chapter will argue that Smith developed his own version of “libertarian paternalism.” It differs in important ways from that of Thaler and Sunstein, but it shares with their version an attempt to balance respect for individual autonomy with a desire to help people lead better lives. The chapter will argue that, contrary to some mischaracterizations of Smith’s argument that he was a pure libertarian, Smith acknowledged the important role of a kind of paternalism in helping people make better choices. On the other hand, his sensitivity to the importance of individual choice for developing good judgment leads him also to recommend a robust role for autonomy. It will be argued, then, that Smith’s position accommodates the importance of both liberty and paternalism in enabling individuals to construct lives worth living, while at the same time Smith’s position avoids some of the problems that have beset more recent versions of libertarian paternalism.

2. Self-Command and the Impartial Spectator

Let us begin by summarizing Smith’s position. Smith claims in his 1759 Theory of Moral Sentiments that one of his four central virtues is “self-command.” In fact, he claims that self-command is what gives the other three central virtues—namely, “prudence, justice, and proper beneficence”—their “principal lustre.”4 But what is required for self-command to be a virtue? It seems to require that one have an ability to separate oneself both from one’s own passions and from the influences of others around one, and then to direct one’s judgments and actions with both deliberation and conscious intentionality. Now, Smith had argued that the desire for mutual sympathy of sentiments is one of our strongest social desires, indeed the desire that ultimately gives rise to our shared moral sentiments.5 It is when we discover that others enter into, or “sympathize” with, our own sentiments that we experience the pleasure of mutual sympathy; because the achievement of this sympathy (or “concord” or “harmony”) of sentiments is pleasurable to us, that gives us incentive to judge or behave in ways that generate this mutual sympathy. By contrast, when we realize that others do not enter into our sentiments, we experience the displeasure of an “antipathy” of sentiments, which encourages us to revise our judgments or behavior. The result is a mutual drive toward shared expectations about one another’s behavior, a process that over time gives rise to a shared set of standards about morality, manners, and etiquette. Self-command is a virtue for Smith because it enables us to constrain our judgments and behaviors so that they accord, not only with what we might like, but also with what others expect of us and with what comes to constitute praiseworthy behavior.

(p. 241) Smith’s full argument is, however, more complicated than this. In particular, it involves the creation of the perspective of an “impartial spectator,” which ultimately becomes for Smith the standard of judgment.6 According to Smith, people are born with no morality whatsoever. A baby knows only its own wants. The baby has no notion of a proper (or improper) thing to ask for, of a proper (or improper) way to ask for it, or of embarrassment for having asked for something it should not have. Hence the baby attempts to have its wants satisfied simply by alarming its caregiver with howls and cries. Yet we do not blame the baby for such self-indulgence: it is not yet capable of considering propriety or others’ interests; besides, Smith says, it is probably encouraged in its self-centeredness by its indulgent parent or nurse.7

According to Smith it is not until the baby has grown to a child and begins playing with his mates that the child has the jolting experience of realizing that he is not the center of everyone’s life, only of his own.8 Smith writes that this is the child’s introduction into the “great school of self-command” (Smith, 1982: III.3.22): it is on being with others and experiencing them judging oneself—even if only implicitly, by, say, not playing with one or simply ignoring one’s demands to have one’s desires satisfied—that one experiences the displeasure associated with an antipathy of sentiments. After the initial jolt, one casts about to find a way to relieve the displeasure, eventually hitting upon using one’s “self-command” to modify one’s sentiments and behavior so that they more closely match those of one’s playmates. At that point an exquisite new pleasure is experienced, that of the mutual sympathy of sentiments, and a new and enduring desire for that pleasure has been aroused. The experience of being judged thus triggers in the child what Smith calls “an original desire to please, and an original aversion to offend his brethren” (Smith, 1982: III.2.6). From that point on, according to Smith, the child regularly engages in trial-and-error investigation into what behaviors will achieve this sympathy and thus satisfy this desire.

These trial-and-error attempts lead the individual to adopt habits and then rules of behavior and judgment that increase the chance of achieving mutual sympathy. By the time the child has become an adult, he has adopted a wide range of principles of behavior and judgment that he can apply in many different situations. Since everyone else is engaging in precisely the same investigation, all our disparate sentiments tend to gravitate toward mutually acceptable means.9 This is the invisible-hand mechanism that Smith thinks generates commonly shared standards of behavior and judgment, indeed a commonly shared system of morality.

Another aspect of human judgment-making, however, is self-judgments: we often approve or disapprove of our own sentiments or actions, we often in retrospect feel pride or shame for what we felt or did, and we often make resolutions to behave in certain ways or to refrain from behaving in certain ways in the future. Smith accounts for this aspect of human morality by reference to the perspective of what he calls the “impartial spectator.”

Smith argues that the process of passing judgment on oneself is, despite what one might initially suspect, quite similar to that of passing judgment on others. In the latter case, “We either approve or disapprove of the conduct of another man according as we feel that, when we bring his case home to ourselves, we either can or cannot entirely (p. 242) sympathize with the sentiments and motives which directed it”; similarly, in the former case, “we either approve or disapprove of our own conduct, according as we feel that, when we place ourselves in the situation of another man, and view it, as it were, with his eyes and from his station, we either can or cannot entirely enter into and sympathize with the sentiments and motives which influenced it” (Smith, 1982: III.1.1). In judging oneself, one must divide oneself “as it were, into two persons”: the “I” as agent or person principally concerned, and the “I” who is judging the person principally concerned (Smith, 1982: III.1.6). Smith argues that the habit of judging one’s own character in this way is what develops into what is often called one’s “conscience.” This may sound like a cumbersome process, but Smith’s argument is that in practice it happens much more readily than it appears. The idea is this: when you are reflecting on your own sentiments, you consider what another person informed of your situation would think. Would an impartial observer approve of your conduct? If so, then you may proceed; if not, then not.

Why should we care what an imaginary observer thinks of our conduct? For two reasons, Smith thinks. First, because your habit of doing so is so deeply ingrained in you already that if you do not heed your conscience you will be unhappy. We can see the importance of this claim by recalling a classic problem in the history of philosophy. In the second book of Plato’s Republic, Socrates is asked to respond to one of the most difficult and enduring problems in moral philosophy, namely: Why would a person not commit injustice if he could be absolutely certain he could get away with it? The question is prompted by the story of a young shepherd who discovers a ring with the magical power of making its wearer invisible. According to the story, the shepherd uses the ring’s power of invisibility to seduce the king’s wife, then to conspire with her to surprise and murder the king, and finally to install himself upon the throne. The lesson of this story, according to Glaucon (who tells it), is that “one is never just willingly but only when compelled to be” (Plato, 1992: II.360c). Socrates thereupon commences a lengthy reply to this challenge, arguing ultimately that the life of injustice is not in fact more desirable than the life of justice: “And haven’t we found that justice itself is the best thing for the soul itself, and that the soul—whether it has the ring of Gyges or even it together with the cap of Hades—should do just things?” (Plato, 1992: X.612b). Socrates further suggests that the just person will fare better in the afterworld than the unjust person—providing another reason to act justly in this life, even if one could act otherwise with impunity.10

The question of why one should be moral has persisted, however, and Smith develops his own answer. According to Smith, man desires not only to be praised, but has a further desire “of being what he himself approves in other men” (Smith, 1982: III.2.7). It appears that for Smith this further desire results from the continued workings of the desire for mutual sympathy of sentiments.11 Smith says that all of us have at some time been frustrated or chagrined when people disapproved of our conduct because, we believe, they did not fully understand the circumstances involved (Smith, 1982: III.2.31–2). An only partial familiarity with another’s situation might well bias or prejudice the judgment one makes, and all of us have experienced the unpleasantness of being on the wrong end of a biased, prejudiced, or partly uninformed (even misinformed) judgment. (p. 243) Smith argues that unpleasant experiences like these encourage us in such circumstances to repair, not to the judgment of actual observers, but instead to the judgment of an imaginary, informed, disinterested observer. Smith calls this standard an “impartial spectator.” Smith believes that this standard comes about naturally, or unintentionally—as a result, on the one hand, of individuals’ wanting mutual sympathy of sentiments and, on the other, of their frequent frustration at others’ inability or unwillingness to expend the effort necessary to understand another’s full situation before passing judgment on his sentiments and actions. When mutual sympathy is not forthcoming from actual, and often partial, spectators, we may find solace in an imagined impartial spectator who, because of his impartiality, would approve of our conduct.12 Though the perspective of this imagined spectator is constructed through our own experiences and is hence itself liable to various partialities and biases, Smith argues that his judgment will more closely approximate an idealized standard because it will be informed by our realizations that more accurate judgments ensue from fully informed but disinterested observers. Our desire for mutual sympathy of sentiments then expands to include, and can be satisfied by, a sympathy achieved with such an imagined idealized observer. We are thus naturally led to consult this imaginary observer frequently. Over time the practice of doing so becomes habitual, and so we will as often consult this imagined impartial spectator’s sentiments as we will the sentiments of actual spectators.

Now we can see Smith’s answer to Plato’s Ring of Gyges problem. Even if we were guaranteed that no actual person would discover our misconduct, we would still be unable to avoid the damning judgment of our conscience, our imagined impartial spectator. The force of habit means that this impartial spectator’s judgment occurs to us even unwittingly, and in such a case it would inform us that our misconduct is such that we ourselves would condemn it in another person. That realization is sufficient to trigger the antipathy, and thus displeasure, of a failure of an imagined mutual sympathy of sentiments. “The man who has broke through all those measures of conduct, which can alone render him agreeable to mankind,” Smith writes, when “he looks back upon it, and views it in the light in which the impartial spectator would view it, he finds that he can enter into none of the motives which influenced it. He is abashed and confounded at the thoughts of it, and necessarily feels a very high degree of that shame which he would be exposed to, if his actions should ever come to be generally known” (Smith, 1982: III.2.9). The worse the misconduct, according to Smith, the greater the shame or guilt the person feels. If the misconduct is of the most grievous kinds, there is no escape from the judgment of the impartial spectator:

These natural pangs of an affrighted conscience are the daemons, the avenging furies, which, in this life, haunt the guilty, which allow them neither quiet nor repose, which often drive them to despair and distraction, from which no assurance of secrecy can protect them, from which no principles of irreligion can entirely deliver them, and from which nothing can free them but the vilest and most abject of all states, a complete insensibility to honour and infamy, to vice and virtue.

(Smith, 1982: III.2.9)

(p. 244) For Smith, then, the reasons to follow the rules of morality include: (1) the anticipated pleasure resulting from a mutual sympathy of sentiments of actual spectators who approve of one’s actions, and (2) the pleasure resulting from a sympathy of sentiments with an imagined impartial spectator; but also (3) the fear of an anticipated displeasure resulting from the judgment of other actual spectators who know of one’s misdeed, and (4) the displeasure resulting from an antipathy of sentiments with an imagined impartial spectator. These conspire to provide a powerful incentive to follow those rules of conduct that one approves in others and not to follow rules of conduct that one disapproves in others.

Thus the first reason Smith believes we should pay attention to what our conscience tells us is to avoid the substantial risk that we otherwise run of feeling unpleasant shame and disgrace at our own misconduct. But the second reason is the other side of this equation: you are, or might be, able to achieve a pleasurable sympathy of sentiments with your imagined spectator. And the more thoroughly this imagined spectator knows the situations of your sentiments and conduct, and still approves, the more exquisite the pleasure you feel from the ensuing mutual sympathy of sentiments. With diligence and application, a person can train himself, Smith thinks, to heed mainly, even perhaps exclusively, the judgment of the impartial spectator, and such a person will as nearly approximate the conduct of a truly virtuous person as a human being can:

To a real wise man the judicious and well-weighed approbation of a single wise man, gives more heartfelt satisfaction than all the noisy applauses of ten thousand ignorant though enthusiastic admirers. He may say with Parmenides, who, upon reading a philosophical discourse before a public assembly at Athens, and observing, that, except Plato, the whole company had left him, continued, notwithstanding, to read on, and said that Plato alone was audience sufficient for him.

(Smith, 1982: VI.iii.31)13

Smith is adamant about this point. Unmerited applause brings no real pleasure, he insists, since we are always “secretly conscious”14 to ourselves that we do not in fact merit the applause. At the same time, the pain we feel from unjust censure can sting more deeply and remain with us far longer than it should: “Unmerited reproach, however, is frequently capable of mortifying very severely even men of more than ordinary constancy. . . . [A]n innocent man, though of more than ordinary constancy, is often, not only shocked, but most severely mortified by the serious, though false, imputation of a crime; especially when that imputation happens unfortunately to be supported by some circumstances which give it an air of probability” (Smith, 1982: III.2.11). Though such a person should repose and take comfort in the knowledge that an impartial spectator would not so condemn him, Smith allows that only people of the firmest “constancy” are capable of that kind of moral discipline.

On Smith’s account, then, a morally mature person is one who judges himself, and others, from the perspective of an imagined impartial spectator who is fully aware of his situation but has no personal stake in the outcome. Self-command is the virtue that is necessary to enable the morally mature person to discipline his actions in accordance (p. 245) with the judgment of the impartial spectator, which at times means acting in opposition to what actual spectators expect or approve.

3. Smith’s Libertarian Paternalism

Though we can, then, on Smith’s account, distance ourselves from others’ expectations of us, one thing from which Smith thinks we cannot distance ourselves is the desire for mutual sympathy of sentiments. This may seem like a liability, but it is actually a crucial part of Smith’s account. One reason Smith can be optimistic about the development of relatively beneficial decentralized social orders—that is, orders not administrated by central authorities—is precisely because he believes we all feel the pull of the desire for mutual sympathy of sentiments. This is Smith’s libertarian paternalism.

The self-command Smith extols comprises two main things. The first is a strong and natural pull toward compliance with others’ expectations. It is a kind of centripetal social force. This is a good thing: Because a chief component of happiness is loving relations with others—Smith claims that “the chief part of human happiness arises from the consciousness of being beloved” (Smith, 1982: I.ii.5.1)—each of us thus has to figure out how to establish and maintain such relationships. That means that we will have to moderate our behavior so that it falls within what others find to be acceptable parameters.15 This gives us reason to seek out companionship with others, but more than that it encourages us to coordinate our behaviors, beliefs, even tastes with those others who matter to us. Their expectations become, then, nudges for us, and as naturally social creatures we thus face a natural hypothetical imperative: If I want to be happy, I need to accommodate myself to others’ expectations. The key is that we apply such nudges to others—rewarding and punishing behaviors, as the case may be—even when others do not understand why we do so, or would prefer to behave differently from the way we are encouraging. When parents give their children guidelines and corrections for behavior, these paternalistic nudges are aimed at helping them comply not only with the parents’ own expectations but with what parents know will be others’ expectations of their children—again, even when children do not understand or do not agree with the nudges. Parents understand, at least at some level, that their children’s relative ability to figure out and even anticipate what others will expect of them will be a skill that is necessary (though, of course, not sufficient) for successful relations with others—and, ultimately, one key ingredient for happiness. So the good parent seeks to give his children a leg up on successful social relations by training them in a particular signature of behavioral habits designed to help them adopt behaviors that are likely to achieve mutual sympathy of sentiments, and avoid behaviors likely to achieve an antipathy of sentiments, with others.16

What constitutes appropriate accommodations that any of us should make can fall, however, only within reasonable parameters. And, crucially for Smith, we always retain the ability—in some cases, perhaps even the duty, as when the impartial spectator (p. 246) requires it—to break from others’ expectations. This is the second part of Smith’s conception of self-command. The desire for mutual sympathy of sentiments means that the default becomes to comport with others’ expectations, and our ability to exercise self-command indicates our relative ability to do so. Yet as adults we can also choose when not to comply with others’ expectations. This is Smith’s psychological, moral, and, ultimately, political libertarianism. Now, knowing when to comply and when to defect is not easy. It is a function of good judgment, which Smith, following Aristotle, believes is a skill that must be practiced to be effective. But practice alone is not enough. We must also have feedback, and this feedback must actually have some purchase on us. That is precisely the role that the desire for mutual sympathy of sentiments plays. When others do not enter into our sentiments, it generates a displeasure in us—we cannot escape that. But what we can do is select those opportune moments when the displeasure is worth it—when, that is, some other good (perhaps compliance instead with our imagined impartial spectator) outweighs the cost of failing to achieve mutual sympathy of sentiments with actual spectators. Every morally mature person of good judgment deliberately distances himself from potential instances of mutual sympathy of sentiments at various times in his life. It is often difficult to know when we should do so, and we often get it wrong. But we possess the ability to choose to depart from others’ expectations, and virtue will require that we do so in many circumstances throughout our lives.

Hence the picture of human freedom that emerges for Smith is one of libertarian paternalism, on the individual level as well as—therefore—on the political level. At the individual level: the behavioral principles framed from others’ expectations of us are their nudges, and the overt recommendations we make to people to follow those principles are part of the paternalistic aspect of human sociality. Our default of following the nudges is thus a proper response to that paternalistic impulse. Yet those occasions when we demur are instances of our freedom to follow different paths.

According to Smith, this conjunction of liberty and paternalism is part of what makes morality beautiful, and what makes people morally beautiful. In The Theory of Moral Sentiments, Part IV, Smith argues that beauty pertains to (1) order and (2) fitness or propriety. So the morally beautiful person is one whose regular course of behavior follows from a relatively small set of foundational principles of character, and yet whose actions in any particular case are the proper ones in those particular circumstances. This incorporates both aspects of libertarian paternalism. The set of default principles are derived from induction on the basis of experience with other people’s expectations; the rules we develop inductively are recommended to us paternalistically, as conducive to happiness (or “utility” or “conveniency,” in Smith’s words17). On the other hand, virtue also requires us to depart from those paternalistic recommendations in the appropriate circumstances. The facts (1) that we can do so, and (2) that we exercise our “self-command” when we do so appropriately, are indicative of our moral freedom—the “libertarian” aspect of Smith’s conception of human moral agency.

But Smith’s libertarian paternalism applies at the political level as well. Smith’s survey of historical and empirical evidence leads him to conclude in The Wealth of Nations that those societies that allow labor to divide and people to enjoy the rewards, or suffer the (p. 247) defeats (as the case may be), of their activities are the ones in which prosperity increases, in which a “universal opulence” and a “general plenty” ensue.18 He concludes that government should do only a few things: enforce justice,19 against both foreign and domestic aggression, and provide those few “publick works” that both provide benefit to the nation as a whole and cannot be provided by private enterprise.20 Smith’s category of “publick works” seems quite small, however: Smith suggests it is only a few things, mainly infrastructure that facilitates commerce, as well as partially subsidized, locally controlled, primary schooling. But what about all the other things that are required for a fully virtuous life? Smith worried in Book V of The Wealth of Nations, for example, about the deleterious effects that extreme division of labor could have on the minds of laborers—it could render them as “stupid and ignorant as it is possible for a human creature to become” (Smith, 1976: V.i.f.50)—and he suggests requiring every citizen to learn to read, write, and do basic arithmetic and geometry as a partial remedy.21 But in The Theory of Moral Sentiments Smith speaks at length about the various aspects of “beneficence”—that is, positive, including paternalistic, actions that one should take on others’ behalf—that are required to be fully virtuous. If the state takes no cognizance of such matters, how can we be sure people will be able to develop into virtuous creatures?

The Smithian answer is the libertarian paternalism that underlies his conception of the moral marketplace. The state is discharged from superintending the beneficent actions in which people should engage not only because it is incompetent to do so—such matters cannot be determined in the abstract or from afar, so dependent is proper beneficence on local details of particular situations—but also because people’s natural desire for mutual sympathy of sentiments will already do the job as well as can be hoped. The patterns of localized moral judgment become like prices: they convey information (about people’s expectations, their tolerances, their willingness to go along with novelty, and so on) that are responsive to people’s particular circumstances and exploit people’s knowledge of those circumstances in a way that no distant third party, however intelligent or well intentioned, could possibly do.

4. Justice and Beneficence

What we are calling Smith’s libertarian paternalism parallels, and relies in part on, the sharp distinction Smith draws between what he calls “justice” and what he calls “beneficence.”22 The next step in capturing Smith’s position, then, requires an understanding of Smith’s conception of those two virtues.

Smith argues that “because the mere want of beneficence tends to do no real positive evil,” it follows that (1) beneficence therefore “cannot be extorted by force” and (2) “the mere want of it exposes to no punishment” (Smith, 1982: II.ii.1.3). According to Smith, acting with justice toward you leaves neither you nor anyone else worse off than you already were, though it may not by itself leave you or anyone else better off. For this reason, Smith calls it “a negative virtue” (Smith, 1982: II.ii.1.9), claiming that “We may often (p. 248) fulfil all the rules of justice by sitting still and doing nothing” (ibid.). The person sitting still and doing nothing is not acting with positive virtue—in other words, is not generating any improvement—“But,” Smith contends, “still he does no positive hurt to anybody. He only does not do that good which in propriety he ought to have done” (Smith, 1982: II.ii.1.3).

So failing in proper beneficence—a category that for Smith includes things like charity, compassion, generosity, and “humanity”—might give us reason to disapprove of and be reasonably disappointed by another’s behavior, but it does not license coercive punishment like jailing or fines. If I do not do you the good office you hoped or expected I would, you may be disappointed, even justifiably so; but you are no worse off than you were before. Because I have done you no “real positive hurt,” meaning I have not worsened your ex ante position, you may not take positive action to punish me. By contrast, if I fail to act with justice toward you, that means I did indeed do “real positive hurt” to you. I left you worse off than you were before, and that gives rise to justified resentment, which licenses punishment.

According to Smith, there are only three rules of justice: “the laws which guard the life and person of our neighbor,” “those which guard his property and possessions; and last of all come those which guard what are called his personal rights, or what is due to him from the promises of others” (Smith, 1982: II.ii.2.2). We act justly, then, according to Smith, when (1) we do not kill or molest others, (2) we do not steal from or defraud others, and (3) we do not break our voluntary contracts or promises. By contrast with justice, beneficence involves making at least one person better off. Assuming, however, that that improvement was not required of me by promise, contract, or other specific obligation, you cannot have had an enforceable expectation of improvement.

This seems a rather thin conception of justice: What about charity, or helping others who need it, especially when we are in a position to help? Smith’s position is that beneficence is indeed frequently morally required, but since it is positive, not negative, it is not part of justice, and therefore not justifiable for state or other third-party coercion.

Smith offers several reasons for supporting his thin conception of “negative” justice. Here are three, each of which links to an important difference between Smithian libertarian paternalism and Thaler and Sunstein’s more recent version.

Smith’s first reason supporting his “thin” conception of justice rests on his empirical claim that no society can subsist unless its members respect these rules of conduct. Even a society of “robbers and murderers,” Smith says, must at least “abstain from robbing and murdering one another” (Smith, 1982: II.ii.3.3). On the other hand, according to Smith, a society can subsist if its members respect these rules of justice but do not act with beneficence toward each other. Because beneficence “is less essential to the existence of society than justice,” Smith concludes that “Society may subsist, though not in the most comfortable state, without beneficence; but the prevalence of injustice must utterly destroy it” (ibid.) Justice, therefore, “is the main pillar that upholds the whole edifice” of society, while beneficence “is the ornament which embellishes, not the foundation which supports the building”; for that reason, Smith argues, it is “sufficient to recommend, but by no means necessary to impose” beneficence (Smith, 1982: II.ii.3.4). (p. 249) So: justice is both necessary and sufficient for the existence of society, but beneficence is neither necessary nor sufficient. That means that they enjoy a lexical priority—justice first, beneficence only thereafter—and, if we accept the Smithian claim that the state is justified in providing only what is necessary for human society, then it follows that it is justified in providing only justice.

A second reason Smith supports this thin conception of justice is its relative ease of administration: it is relatively easy (1) to capture its essence in simple rules, (2) to detect infractions of it, and (3) to remedy infractions. By contrast, beneficence is far more difficult to describe in rules, far more difficult to detect in its absence, and far more difficult to remedy.23 Unlike infractions of justice, improper beneficence can be detected and adjudicated only on the basis of detailed, context-specific knowledge of the situation, persons, and matters involved in particular cases. We might all agree, for example, that we should be generous and that generosity is a virtue; nevertheless, it would be very difficult to generate a set of precise rules that will allow us to determine what generosity requires of me, here, now. In practice, we have to rely on practical judgment, which Smith, like Aristotle, believes does not operate by mechanical execution of general rules.

Smith’s final reason is because it allows for a proper sensitivity to individual circumstances. What counts as being sufficiently generous depends on the particular circumstances of the case in question: the history and situation of the people involved, their available means and tradeoffs and opportunity costs, and even their goals and ambitions are all material considerations. There is also typically a range of behaviors or actions that might qualify as properly beneficent, which means that often no single course of action will be required to satisfy one’s obligations. Therefore, beneficence cannot plausibly be incorporated into the definition of justice, which, because it can license coercion, requires predictable application of clear rules. Smith’s thin concept of justice restricts it, therefore, to those few areas of conduct that it can plausibly and effectively address, and leaves to localized judgment the determination of what the virtues of positive beneficence require in light of particular circumstances.

5. Smithian Libertarian Paternalism

We can now see how Smith’s version of libertarian paternalism contrasts with the “libertarian paternalism” of more recent vintage, namely, that of Thaler and Sunstein. In both cases we have recommendations about what behaviors to engage in or to avoid based on a (kind of) hypothetical imperative: for Smith, it is the hypothetical imperative deriving from the desire of having a happy life, while for Thaler and Sunstein it is more like having a healthy life. And in both cases we see the desire to satisfy the twin goals of enabling individual free choice while also encouraging and enabling people to make better choices.

Yet there are several important differences between the two accounts. First, Smith’s argument that justice (as he understands it) is both necessary and sufficient for society to subsist means that whatever paternalistic nudging we think might help people (p. 250) make good decisions must be (1) pursued entirely within the realm of private—i.e., nongovernmental—social relations, (2) tailored to individuals’ localized circumstances, and (3) pursued only after observance of the rules of justice. Second, Smith’s argument that justice is relatively easy to administrate centrally, while beneficence (here including paternalism) is much more difficult, means that on Smith’s account paternalism must be decentralized and local if it is to be effective. Third and finally, Smith’s argument that effective beneficence requires extensive knowledge of, and must be adapted to, people’s individual circumstances means that beneficence (and paternalism) will fail if it proceeds from centralized and general laws or regulations.

One place where it might seem that there is a difference is in Smith’s relatively stronger emphasis on liberty and relatively weaker emphasis on paternalism than that of Thaler and Sunstein. It might seem as though Smith’s definition of justice is a principled one, debarring—on principle, as it were—the state from engaging in any beneficent activity (which might include paternalistic nudging). But that is not Smith’s position. When he argues for “publick works” that would (1) provide general benefit and (2) be unable to be provided by private enterprise, a latter-day Smithian might conclude that the members of Smith’s putative set of public works are very few indeed—perhaps zero. There are, after all, privately provided roads, bridges, and other infrastructure, private schools and educational programs, and so on. Perhaps even security and dispute resolution can be provided by private enterprise. What, then, is left that markets could not provide? Although this might ultimately be an empirical question, the entire range of nudges Thaler and Sunstein suggest are justified precisely as areas in which private enterprise either cannot or at least does not help society meet optimal results. That means they would qualify for consideration even under Smith’s otherwise stringent requirements. The analogy between moral community and political community would hold: just as one should follow the impartial spectator’s recommendations, even on those occasions when one does not wish to do so, so too there might be times when we should impose nudges on others, even when others, at least at the moment, do not wish to us to nudge them. For Smith, as long as we still allow others the possibility of choosing otherwise, despite our expectations or nudges, we can respect their autonomy and still encourage them to make wise choices even despite themselves.

6. Conclusion: Smithian Libertarianism?

One might wonder whether Smithian libertarian paternalism as it is presented here counts as properly paternalistic at all. Most contemporary accounts of paternalism include some measure of coercion: either directly, by requiring compliance with a set of policies or behaviors regardless of what individuals wish; or indirectly, by imposing benefits on favored, and costs on disfavored, activities or behaviors, again regardless of (p. 251) individual desires. But if Smith has ruled out state action for beneficent purposes, and relegated “nudging” to the realm of private relations and associations among individuals, then to what extent is his position truly paternalistic? Perhaps Smith’s position is like Mill’s, according to which one may remonstrate, reason, persuade, or entreat another, but not compel or visit another with evil to get him to comply.24 So perhaps Smith is just a Millian libertarian after all?

Smith’s position cannot be reconciled with the paternalism of Conly or Ubel, which argues for an abandonment of a concern for autonomy altogether. Importantly, they argue for paternalism not only via government action but also via private action: they call on businesses, schools, and other social enterprises also to implement a “choice architecture” strategy with the aim of helping their constituents make good, or better, choices. That position is much closer to Smith’s. Smith would not object to private persons and private enterprises adopting a deliberate choice-architecture strategy, and indeed he would recommend they do so on the grounds that their localized knowledge, combined with their localized feedback, can be an important part of the process of encouraging their associates to develop both good judgment and good behavior. At the same time, Smith believes that the withholding of mutual sympathy of sentiments is itself the imposition of a cost, since each of us naturally desires this mutual sympathy. If others choose not to approve of my conduct, this counts, then, as a paternalistic punishment because I do in fact desire their approval and experience the lack of it as a displeasure.

Smith would argue, however, that the Thaler and Sunstein program overreaches when it takes the important insight that sometimes others can help us make better decisions for ourselves and transfers that responsibility to governmental agents. For the reasons discussed earlier, Smith thinks doing that would undermine the benefits that paternalism can potentially provide. Because the relative success of nudges is dependent on the details of particular situations, the level at which nudges are applied matters. The Smithian position would thus call for a decentralized federalism: allowing and encouraging the greatest scope of independent paternalism at the most local levels, with less and less paternalism allowed as the level of decision-making ascends from private individuals, groups, or firms, to municipalities, to states, to federal governments.

On Smith’s account, proper paternalistic recommendations arise decentrally from the experiences and expectations of people’s actual lives and interactions. Like the social planner attempting to set prices or allocate resources centrally, however, Smith would argue that legislators or regulators are not in a position that would allow them to exploit people’s localized knowledge of crucial relevant details about their lives—their unique circumstances, associations, relationships, opportunities, schedules of value, hierarchies of purpose—and thus the paternalistic nudges of centralized legislators would, like those of the central economic planner, likely fail to comport with what would actually conduce to people’s well-being. Thus Smithian libertarian paternalism—unlike the Thaler and Sunstein version applied to legislators—would exploit people’s local knowledge and respond to people’s lived experience.

(p. 252) It would also have what Smith believes is the additional considerable benefit of allowing no centralized group of persons to assume the authority of directing the lives of others about whose unique circumstances they know little. In a famous passage, Smith calls a centralized authority undertaking to guide people in proper directions a “man of system,” who, according to Smith, overestimates his knowledge of others’ situations, and thus overestimates his ability to actually benefit others. Smith writes that the “man of system”

seems to imagine that he can arrange the different members of a great society with as much ease as the hand arranges the different pieces upon a chess-board. He does not consider that the pieces upon the chess-board have no other principle of motion besides that which the hand impresses upon them; but that, in the great chess-board of human society, every single piece has a principle of motion of its own, altogether different from that which the legislature might chuse to impress upon it.

(Smith, 1982: VI.ii.2.17)

It is crucial to see, however, that Smith’s criticism here is directed at centralized legislators, not at local fellow citizens—a distinction Smith himself recognizes when he elsewhere allows that a local “civil magistrate” may, after all, “prescribe rules . . . which not only prohibit mutual injuries among fellow-citizens, but command mutual good offices to a certain degree” (Smith, 1982: II.ii.1.8). Smith rails against the “arrogance” of the “most dangerous” prince who believes himself competent to organize all of society according to his “ideal plan of government” (Smith, 1982: VI.ii.2.17–18) and, by extension, competent to excogitate and implement beneficial nudges from his centralized Olympian perch; and yet Smith recognizes the potential for local communities to organize themselves in mutually beneficial ways using even paternalistic means.

Smith’s conception of political freedom and human agency thus incorporates a robust notion of political and economic community by giving people strong incentives to associate beneficially with others. At the same time, it allows the generation and application of robust moral communities, complete with moral agents who feel the pull of the desire for mutual sympathy of sentiments but yet retain the ability to make independent decisions. Smith writes, “A moral being is an accountable being” (Smith, 1982: III.i.3.[3]). That means that Smith’s moral agents assume both a responsibility for themselves and an accountability to others—which just is his libertarian paternalism.


I thank the Center for the Philosophy of Freedom at the University of Arizona for allowing me space and time to draft the initial ideas that led to this chapter. I would also like to thank Carmen Pavel, David Schmidtz, and two anonymous referees for excellent constructive criticism of an earlier draft of this chapter. Remaining errors are mine alone.


Conly, Sarah, 2013. Against autonomy: justifying coercive paternalism. New York: Cambridge University Press.Find this resource:

Griswold, Charles L., Jr., 1999. Adam Smith and the virtues of enlightenment. New York: Cambridge University Press.Find this resource:

Mill, John Stuart, 1978 (1859). On liberty. Indianapolis: Hackett.Find this resource:

Otteson, James R., 2002. Adam Smith’s marketplace of life. New York: Cambridge University Press.Find this resource:

Plato, 1992. Republic. Translated by G. M. A. Grube and C. D. C. Reeve. Indianapolis: Hackett.Find this resource:

Raphael, D. D., 2007. The impartial spectator: Adam Smith’s moral philosophy. New York: Oxford University Press.Find this resource:

Smith, Adam, 1976 (1776). An inquiry into the nature and causes of the wealth of nations. Edited by R.H. Campbell and A.S. Skinner. Indianapolis: Liberty Fund.Find this resource:

Smith, Adam, 1982 (1759). The theory of moral sentiments. Edited by D. D. Raphael and A. L. Macfie. Indianapolis: Liberty Fund.Find this resource:

Thaler, Richard, and Sunstein, Cass, 2009. Nudge: improving decisions about health, wealth, and happiness. Rev. and exp. ed. New York: Penguin.Find this resource:

Ubel, Peter, 2009. Free market madness: why human nature is at odds with economics—and why it matters. Cambridge: Harvard Business Press.Find this resource:

White, Mark D., 2013. The manipulation of choice: ethics and libertarian paternalism. New York: Palgrave Macmillan.Find this resource:


(2.) See, for example, White, 2013.

(3.) See, for example, Conly, 2013 and Ubel, 2009.

(4.) Smith, 1982 (1759): VI.iii.11.

(5.) Smith writes, “But whatever may be the cause of sympathy, or however it may be excited, nothing pleases us more than to observe in other men a fellow-feeling with all the emotions of our own breast” (Smith, 1982: I.i.2.1); similarly, “Nature, when she formed man for society, endowed him with an original desire to please, and an original aversion to offend his brethren” (Smith, 1982: III.2.6). For discussion, see Otteson, 2002: chap. 1.

(7.) See Smith, 1982: III.3.22.

(8.) In order not to beg any questions, I maintain throughout Smith’s exclusive use of the masculine pronoun.

(9.) Smith writes that we “always endeavor to bring down our own passions to that pitch, which the particular company we are in may be expected to go along with” (Smith, 1982: I.i.4.9). Yet he also claims: “though [different people’s] sentiments will never be unisons, they may be concords, and this is all that is wanted or required” (Smith, 1982: I.i.4.7).

(10.) Much of the Republic is an extended response to Glaucon’s challenge. Socrates’s suggestion that the just person will fare better in the afterworld is contained in his myth of Er, which begins at Republic X.614b.

(11.) Other scholars have different views about the nature and origin of this desire. See, for example, Griswold, 1999: chap. 3.

(12.) Smith writes: “But though man has, in this manner, been rendered the immediate judge of mankind, he has been rendered so only in the first instance; and an appeal lies from his sentence to a much higher tribunal, to the tribunal of their own consciences, to that of the supposed impartial and well-informed spectator, to that of the man within the breast, the great judge and arbiter of their conduct” (Smith, 1982: III.2.32).

(13.) As the editors of the Glasgow edition of TMS note, the story Smith has in mind probably refers to Antimachus, not Parmenides. Parmenides probably died before Plato was born. See Smith, 1982: VI.iii.31 n27.

(14.) Smith, 1982: III.4.12 and passim.

(15.) Consider, for example, the small but telling case of joke-telling, one of Smith’s favorite examples: knowing what constitutes a proper joke to tell, and how much—indeed whether—we should laugh at a joke, is determined by judgment that has been honed by experience. See Smith, 1982: I.i.2.1, I.i.3.1–2, and III.i.4.

(16.) Parents no doubt also want to enjoy the pleasure of imagining that others will form positive judgments of them when their children behave well, and avoid the discomfort associated with becoming aware that others might form negative judgments when their children behave badly.

(17.) See Smith, 1982: IV.1.1–2.

(18.) Smith, 1976 (1776): I.i.10.

(19.) More is said about Smith’s conception of justice in the next section.

(20.) See Smith, 1982: II.ii.2.2, Smith, 1976: IV.ix.51, and Smith, 1976: V.i.c.1.

(21.) See Smith, 1976: V.i.f.54–5.

(22.) Smith respects the distinction between “beneficence” and “benevolence”: by “beneficence” Smith means taking positive action to do good for another; by “benevolence” Smith means wishing another well. His discussion here concerns the difference he sees between justice and beneficence.

(23.) Smith offers the following analogy to illustrate the differences: “The rules of justice may be compared to the rules of grammar; the rules of the other virtues, to the rules which critics lay down for the attainment of what is sublime and elegant in composition. The one, are precise, accurate, and indispensable. The other, are loose, vague, and indeterminate” (Smith, 1982: III.6.11). Smith claims that although there are “general rules of almost all the virtues,” nevertheless, reflecting his theory of the dynamic and social way in which rules of conduct develop, he also claims that “to affect, however, a very strict and literal adherence to them would evidently be the most absurd and ridiculous pedantry” (Smith, 1982: III.6.8).

(24.) This is a paraphrase from Mill, 1978 (1859): 9.