Gilles Saint-Paul's Tyranny of Utility offers a criticism of the current trend to treat behavioural economics as justifying ever-expanding government intrusion in the lives of private citizens. Although Saint-Paul tends to be uncharitable to his paternalist opponents and to confuse important ethical distinctions between types of policy justifications, the overall call to evaluate more carefully the findings of behaviour economics, as well as to examine more closely what inferences such findings justify regarding the role of the government, is certainly welcome at this point in the philosophical debate.

Saint-Paul begins with an historical examination of political institutions and the conceptions of man on which such institutions relied. Specifically, Saint-Paul considers the conception of man as a rational, unitary individual. Indeed, the historical focus on freedom and personal responsibility requires such a conception. However, as Saint-Paul notes, both the conception of man as rational and the conception of man as unitary have come under fire from philosophers and social scientists. Moreover, behavioural economists now reject the traditional economic view that whatever decision an individual makes reveals his/her actual preferences, and they have shifted their attention to the real rather than the ideal man.

Unfortunately, as Saint-Paul notes, many have used the findings of behavioural economics as justification for expanding the role of the government. The classic example in the literature is that of mandatory pension savings. On a traditional economics view, ‘under-saving’ is not irrational; instead, it indicates that the individual prefers consumption now to consumption later. If, however, we abandon the view that the self is unitary and consistent over time and instead embrace the conception of the individual as multiple incarnations (or selves) within the same body, we can reinterpret the phenomenon of under-saving as one incarnation consuming now and thereby reducing another incarnation's consumption in the future. Thus, the argument on behalf of government intervention goes, we now have a reason for requiring that the current incarnation save now, as it is the most efficient means of protecting the other incarnation.

Saint-Paul rightly notes two troubling outcomes of such argumentative moves: first, the policies indicate a move away from individual freedom and toward some as-yet-undefined conception of social welfare. Indeed, we are likely to promote individual freedom only insofar as it advances social welfare, whatever that may be. Thus, either we are no longer concerned with individual liberty, or we are concerned only insofar as individual liberty is instrumental in advancing other goals. Either of these outcomes, he argues, should be seriously troubling.

Second, the policies also indicate an abandonment of any sense of personal responsibility. Worse still, abandoning our conception of the self as a unitary, rational individual suggests that we ought to abandon any sense of personal responsibility, at least for those individuals who exhibit a tendency toward the kinds of errors in reasoning that suggest a non-unitary self. Thus, even in the paradigm case of acceptable intervention – setting defaults such that individuals are automatically enrolled in savings systems – we see signs of the troubling trend away from an emphasis on responsibility, and toward an emphasis on social welfare. As Saint-Paul suggests, the dismissal of freedom and responsibility as core, intrinsic values ‘paves the road to tyranny’ (122).

Furthermore, if we abandon the unitary, rational individual, we also ought to abandon the view that choices reveal preferences. As a consequence, we can no longer measure individual welfare by looking only at an individual's decisions. According to Saint-Paul, a natural – though troubling – alternative is to measure welfare directly; rather than assuming that a particular decision maximizes that individual's utility, we instead ask whether that decision makes the individual happy. Aside from the problems with self-reporting that will plague policy debates that rely on such measures, our decision to focus on happiness rather than utility generates an additional problem: if Policy A does not make us happy, but it preserves our freedom, and Policy B makes us happy, but does not preserve our freedom, then we now seem to have a reason to prefer Policy B. Of course, perhaps the loss of freedom is itself detrimental to happiness. Indeed, Saint-Paul cites evidence that suggests that this is so. However, such a calculation suggests only that we need to believe that we are free. Thus, if Policy C makes us happy and does not preserve our freedom but maintains the impression of individual freedom, we ought to prefer Policy C to Policy A. Given a focus on what Saint-Paul refers to as ‘feeling-ethics’, he laments, ‘what really matters is feeling free rather than actually being free’ (73). Again, then, we are on our way to tyranny.

Saint-Paul sounds a much-needed alarm in the current debate about state paternalism. In addition to the principled concern about our shifting social values, he notes that as with any empirical findings, we must be careful to determine both what the findings of behavioural economics actually tell us about individual behaviour, as well as how much weight to give those findings in policy debates. Moreover, most of the policy recommendations behavioural economists have offered are, as Saint-Paul says, ‘piecemeal’; that is to say, no single overriding principle supporting a type of policy has emerged. A related worry is that in the absence of a general theory, behavioural economics seems to support the expansion of the role of the government without any limiting principles. Specifically, we are no longer concerned about historically recognized limits on the expansion of the government, such as constitutionally protected rights. Instead, we worry about those rights only insofar as they contribute to our overall social welfare.

Although Saint-Paul is right to express alarm at the rush to implement new paternalistic policies, both his argument against such policies and his suggestion for correcting our social values are problematic. First, Saint-Paul commits ‘paternalists’ to a host of policies that they need not embrace. Furthermore, much of what Saint-Paul rejects as liberty-threatening paternalism is indeed liberty-threatening, but it is not paternalism (at least on any reasonable conception of the term). Thus, although we certainly should have a debate about whether such restrictions of individual liberty are justified, they are not prima facie troubling in the way that paternalistic policies are. Finally, his own suggestion for reforming society (which he admits is merely a sketch) includes neither a principled argument nor a defence of those empirical claims that are at the heart of his position.

Consider the following three cases: public education, sin taxes and parental authority. In each case, Saint-Paul attributes to the paternalist an obviously troubling view, but each attribution seems remarkably uncharitable. First, consider the case of public education. Here, the state has an opportunity to behave paternalistically by teaching a particular set of values, namely those that will increase happiness. According to Saint-Paul, a

Similarly, elsewhere he suggests that if we discovered that a particular type of education reduced the incidence of depression, that discovery would be sufficient for a paternalist to embrace the education, ‘regardless of why it works’ (125). Of course, that paternalists offer a reason for a particular policy does not mean that the paternalist must endorse the policy. Paternalists might have other values, goals, etc. with which the policy would interfere. Moreover, that the paternalist's prima facie reason for a particular policy does not itself limit the scope of the intervention is not obviously problematic. It seems an unreasonable standard to require that considerations in favour of a particular policy themselves limit the scope of that policy.

A similar problem arises with sin taxes. Here, Saint-Paul says, ‘We already know that paternalists think that if [some] product may be misused, this should be corrected by some sin tax that would artificially increase its price’ (143). Why must the paternalist think this? At best, it seems that the prospect of misuse provides a prima facie reason in favour of such a tax. However, the paternalist need not conclude that that reason is indefeasible. If, for instance, the paternalist is also a utilitarian, she may not be concerned with individual liberty per se, but she may still recognize compelling reasons for not implementing a sin tax, despite the paternalist reason for doing so.

In discussing parental authority, Saint-Paul claims that if happiness researchers determined that something like a massive redistribution of children would make the children happy, ‘it is again quite defensible to remove children from their families’ (91). Again, the paternalist need not concede this, particularly if she is a utilitarian. Even a happiness-driven utilitarian will recognize that concerns beyond merely maximizing the children's happiness are at stake. Of course, this would not – and should not – satisfy the libertarian, whose concern with parental authority does not derive from a belief that such authority maximizes overall happiness. Nonetheless, the policy recommendations attributed to paternalists are uncharitable: paternalists need not believe that the fact that Policy X makes Y happier is a sufficient condition for implementing Policy X.

In addition to an unfair description of the paternalist's commitments, Saint-Paul's examples of ‘paternalistic interferences’ are troubling, in large part because many of them involve no paternalism at all. Returning to the view of the self as non-unitary but instead consisting of multiple incarnations, Saint-Paul says, ‘paternalism imposes costs on one individual's incarnation to prevent another incarnation from inflicting harm on third parties’ (100). Of course, this does not seem to fit with most common definitions of paternalism. Restricting A's freedom for A's good is paternalistic; restricting A's freedom for B's good is not. In his description of the modern paternalistic state, Saint-Paul offers the following examples of paternalistic policies: road safety regulations, such as cell-phone restrictions and maximum blood alcohol content for drivers, gun-control laws limiting legal access to weapons, and Sexual Offenses Prevention Orders (SOPOs). In each of these cases, the restriction of one person's freedom is arguably intended to prevent harm to others; at least, such an intention would justify such policies, even if the actual intention were in fact paternalistic.

Consider the specific example of SOPOs. Saint-Paul offers an example of a SOPO that prevented one individual from using public trains except in cases of an emergency. This SOPO was the result of the man's fifth sexual assault of a woman on a public train. The justification for such a restriction on the individual's freedom, Saint-Paul notes, was the individual's ‘alleged lack of self control’ (105). An order correcting for an individual's lack of self-control is not necessarily paternalistic. In this case, the individual was not being restrained for his own good but rather the good of fellow travellers. He was not being protected from his lack of self-control. Others were.

Of course, that the man threatened harm to others – leaving aside Saint-Paul's questionable assertion that ‘the physical damages are very small and no violence was exerted’ – is not sufficient to limit his freedom (105). However, the justification for the limit is fundamentally different than the justification for paternalistic policies. Paternalistic policies are controversial not only because they limit an individual's freedom, but also because they claim to do so for the individual's own good. That the man in this case might be better off if he is not arrested for a sixth sexual assault is not the reason for interfering with his freedom. Thus, although we should certainly consider whether such a restriction of his freedom was appropriate, the restriction itself does not obviously inherit the moral problems associated with paternalistic interference.

The source of the confusion may be seen in Saint-Paul's description of the ‘benevolent paternalist’. According to Saint-Paul, the benevolent paternalist would ‘want to allow for the government to discretionarily intervene in individual choices if it is proven that such an intervention would have positive consequences for total happiness’ (128). This description fails to recognize a relevant consideration: for a paternalist, it matters very much whose happiness is at stake. Intervening to promote general happiness is not itself paternalistic. Intervening to promote the happiness of the individual with whom the state is intervening is paternalistic.

The problem of referring to these policies as paternalistic gets worse when we remember that we are conceiving of (at least some) individuals as non-unitary but instead as multiple incarnations. Returning to the problem of under-saving, it looks as though mandatory pension savings – the paradigm case of paternalistic interference – is not, in fact, paternalistic if we take seriously the claim that the individual is not unitary. Instead, the state is requiring that one incarnation refrain from harming another. Again, whether this is justified remains a legitimate question. After all, that some action would harm someone else is not a sufficient condition for state interference. Nonetheless, it cannot be objected to on paternalistic grounds. Of course it involves a restriction of freedom, but surely we are occasionally permitted to restrict freedom in order to prevent harm to some third party (or, in this case, some third incarnation).

At times, Saint-Paul seems to suggest that such interferences are not permissible merely because they restrict freedom. In considering the role of scientific evidence in policy debates about violence on television and in video games, Saint-Paul says:

Again, we should have a serious debate about what constitutes optimal design; however, that a right is constitutionally protected and fundamental does not entail that it is also unlimited. Just as our commitment to social welfare ought not to be absolute, neither should our commitment to unlimited individual freedom.

Saint-Paul's criticism of the rush to embrace behavioural economics as justification for rampant state paternalism is a welcome warning bell in the current debate. However, his strategy for shifting our focus back to freedom and responsibility is problematic. First, as Saint-Paul notes, we can reasonably ask why someone ought to consider freedom per se to be more valuable than happiness. As far as I can tell, he offers no response to this question, despite his earlier charge that the state ought not to assume a particular ranking of values on behalf of its citizens. Thus, Saint-Paul's argument will likely convince only those with a prior commitment to liberty as both an intrinsic and a primary value.

Second, Saint-Paul suggests that society adopt practices that treat the individual as unitary and responsible: ‘Being treated by society as responsible and unitary goes a long way toward eliciting responsible and unitary behaviour. The incentives to solve my own behavioural problems are much larger if I expect society to hold me responsible for the consequences of my actions’ (153). The problem, of course, as Saint-Paul spends a good deal of the book explaining, is that society's view of the individual as unitary and responsible may be empirically unfounded, and its treatment of the individual as unitary under such circumstances would seem profoundly unjust. Indeed, the suggestion that my expectation of being held responsible for my actions provides me with an incentive to behave responsibly makes sense only if we already endorse a conception of the self as unitary. Thus, Saint-Paul should offer a defence of both his philosophical commitment to (virtually) unlimited human freedom and his empirical suggestion that the state can encourage non-unitary individuals to become unitary.

Saint-Paul's contribution to the current policy debate regarding paternalism is timely, and society's embrace of new restrictions of individual liberty – coupled with its diminishing emphasis on individual responsibility – is troubling, particularly as such restrictions are done in the name of pursuing an ill-defined and problematic conception of individual welfare. Although we ought to distinguish between debates about the value of a utilitarian state and the value of a paternalistic state, Saint-Paul's call to slow down is welcome to those of us who felt that the move from behavioural economics to state paternalism happened much too quickly.