Practical limits of utilitarianism
Though welfarism, and especially welfarism based on subjective utility, is widely used in normative economics, the empirical facts hardly confirm that people are indeed utilitarian.
Nozick (1974 [1988]) imagined an “experience machine”, which has been fatal to hedonist utilitarianism. Individuals are offered the chance to be plugged into a machine for life. The latter would feed signals into their brain so that they felt wonderful experiences, irrespective of what was actually happening in the world. Since people do not (in general) prefer the happiness machine to their actual life, it seems to follow that the mental state of happiness is not everything that is valued by these people. The preference model does not do better in this regard. In their seminal article on normative experimental economics, based on a survey, Yaari and Bar-Hillel (1984) highlighted in particular that the assessment of income distributions turned on needs rather than on tastes or beliefs, as utilitarians would claim. According to an important and robust result of normative experimental economics, the way people judge the fairness of situations depends on contextual circumstances and not on the mere description of individual utilities. For example, for similar utility levels, an even split of a resource between two persons is not judged as being equal when survey respondents learn that one recipient is hard working and disabled while the other is lazy and already in receipt of benefits. Utility thus does not seem in fact to be the sole important value, and nor is it the sole relevant source of information in people’s eyes.Notably defended by J. Austin (1832 [1954]), rule utilitarianism had been clearly formulated by S.E. Toulmin (1950), J.O. Urmson (1953) and R.B. Brandt (1959). It supposes that the consequences of rules are assessed, and that actions - and also therefore classes of actions - are judged, according to rules.
Harrod (1936) raised an important problem with rule utilitarianism on the basis of a famous example. A lie is a bad thing, and a good rule should be not to lie. However, in special cases a lie is likely to induce better outcomes - although it would be judged negatively by rule utilitarianism, contrary to a priori intuition. Accordingly, act utilitarianism, where every action is assessed through its actual consequences, is favoured. This other version of utilitarianism was defended originally by Bentham, Sidgwick and Moore. Against act utilitarianism, however, some have complained that it is unable to guide practical decision making: because of lack of information, because gathering such information would be too tedious and costly, because mistakes are more likely to be made when there are so many calculations to be made, and because existing rules or norms could not be used to build expectations nor in particular to trust others’ promises. In short, rule utilitarianism requires approximations and generalizations which are likely to conflict with the utility principle in certain concrete cases, while act utilitarianism requires unworkable computations.Notice that the possibility of measuring utility and implementing utilitarian policies depends on the choice of the utility model. If utility were a mental state, it would be quite difficult to consider the different natures of pleasures on a common scale, in so far as many a priori seem incommensurable. The model of experience in play here also entails some counter-intuitive consequences. For example, as long as Mary does not know that she has been ruined on Black Tuesday of the 1929 Crash, she has not experienced her misfortune. Whereas it is impossible to deny that her utility - as experienced happiness - remains as high as when she was a wealthy lady, it is also hard to accept that she remains as well off as before. Moreover, it is impossible - or, rather, difficult - to measure, compare, and add feelings, that is, subjective utilities.
As was already underlined, the preference model supposes individual ordinal rankings that are not easier to compare nor to add. This again hampers actual implementation of utilitarian policies.A famous criticism of utilitarianism is based on Michel Foucault’s (1975) analysis of Bentham’s Panopticon. The Panopticon is a “simple idea of architecture” (Bentham 1791: 5), which determines the organization of life in a prison as a first step. Bentham’s project was to apply this in many other areas of social life, such as homes for the poor, schools, hospitals, public administration, and factories. The prison cells are arranged concentrically around a central building where an inspector is located: he personifies the otherwise fictitious impartial spectator. He can observe each person’s action, and everything that happens. Prisoners are kept separate from the others and from the spectator, of whom they know only they might be being observed. “Hence the major effect of the Panopticon: to induce in the inmate state a conscious and permanent visibility that assures the automatic functioning of power” (Foucault 1975: 234).
The inspector reflects the behavioural expectations of the whole society, and the prisoners feel in their flesh the pressure of the utility principle: hence their own autonomous desires are transformed now so as to feel and act in order to improve the greatest happiness for the greatest number; individuals gradually become their own jailer. As all that is needed here is the thought that there might be an inspector, who may be anybody: he may be changeable or even absent, such that the system seems free from any drift towards tyranny - precisely in accordance with the requirement of impartiality. It remains that individual autonomy is here being totally neglected in favour of social utility: the disciplinary system organized by the Panopticon guarantees the total submission of individuality to the collectivity, as in an authoritarian regime.
This democratic failure may be seen in the preference model as well.
In theory, an impartial spectator properly computes utilities and suggests which policy will best improve social welfare. In practice, utilitarian policies suppose that some experts have made the utility computation in the name of the other individuals. In the ideal preference model, the computation also supposes they launder individual utilities. What is the actual legitimacy of such experts? How can we be sure that such a utilitarian elite may desire and succeed in promoting the good for the mass of people? The problem would not arise if a normative demarcation of the scope of expertise were possible. Under such a demarcation, experts could be contracted within a well-defined area in which they are responsible solely for making factual observations and computations. The citizens would have previously decided that the policy should be utilitarian, such that the experts would use utilitarian models to derive their prescriptions. As a consequence, expert decisions based on utilitarianism would legitimately proceed from the will of individuals, although in fact implemented by an expert. The possibility of such a demarcation is, however, doubtful (Baujard 2013). In the end, this objection again recalls the issue of the lack of democracy in the elaboration of actual policy recommendations and the implementation of utilitarian policies. An alternative to utilitarianism hence supposes the introduction of theories of democracy within welfare economics, for example, following Sen’s more recent ideas (Sen 2009).