The object of this paper is to question the widespread use in economics of the figure of the “social planner” to ground normative assessments and justify paternalism. This perspective treats individuals as mere (more or less efficient) automata, whose sole objective is the satisfaction of exogenously given preferences. The conception of human societies for conventional economics is described as a “hive”–in which each individual is producing welfare–scientifically designed by a social planner so as to maximise the total production of welfare. This approach justifies paternalism as a form of trade between the individuals and the omnipotent and benevolent social planner: by designing adequate incentives, the planner can steer individual behaviours into the right direction, and then maximise the amount of welfare in the society. This is a beneficial outcome both for the individuals and for the planner, since the planner is assumed to be benevolent.

This approach however implicitly claims that what matters for the society is the satisfaction of individual preferences, but without justifying why preference satisfaction matters, and where those preferences come from. This perspective denies in particular individuals' autonomy and their ability to form their preferences, and therefore their own normative assessments. I argue that this difficulty results from the dual dimension of the social planner as the Sovereign of the society (choosing what matters)and as the Government (implementing the measures wanted by the Sovereign), and that economists are not entitled to define what should be the goal of public policies –i.e. to endorse the social planner's perspective as the Sovereign. I illustrate the risks of this approach with the case of behavioural welfare economics and the common claim that boundedly rational individuals should be helped by the “choice architect” so as to achieve their own goals –although the definition of the true ends of the individuals is generally arbitrarily imposed by behavioural welfare economists (such as saving more for one's retirement).

I finally argue that normative economics should be reoriented by placing at its core individual autonomy and democratic processes: instead of designing optimal institutions to reach an exogenous normative objective, normative economists should design institutions to allow the individuals to engage in public debate and hence define what matters for themselves. I illustrate this argument by discussing Ostrom's institutional design principles on the successful management of common pool-resources.