Enrico Mattia SALONIA will defend his thesis on Monday 16th June at 10:00 AM (Auditorium 6, bâtiment TSE)
Title: On Choice, Belief, and Distribution : Axiomatic Studies in Behavioural Economics
Supervisors: Professors Ingela ALGER and François SALANIE
Memberships are:
- Ingela ALGER : Senior Researcher, CRNS/TSE-R Supervisor
- François Salanié : Senior Researcher, INRAE/TSE-R Co Supervisor
- Marcus Pivato : Professeur of Economics, Université Paris 1 Panthéon-Sorbonne Rapporteur
- Marciano Siniscalchi : Professeur of economics, Northwestern University Rapporteure
- Pierpaolo Battigalli : Professeur of economics, Bocconi University Examinateur
Abstract :
This thesis studies principles underlying individual behaviour, information processing, and resource allocation. It focuses on foundational issues in behavioural economics, where the concepts investigated in this thesis are introduced.
The first chapter studies individuals with preferences for universalisation—that is, they consider what would happen if everyone were to act as they do. Universalisation has been shown to have evolutionary foundations, to align with observed behaviour, and to lead to desirable allocations under various normative frameworks. Existing models, such as Homo Moralis preferences and Kantian equilibrium, lack choice-theoretic foundations, limiting their generalisability. To address this, I develop an axiomatic model characterising preferences for universalisation. The main challenge is that universalisation is a non-consequentialist attitude, which is difficult to capture using standard choice-theoretic tools. A key behavioural prediction of my model is that the independence axiom holds only among actions that are universalised in equivalent ways. My framework unifies previous models, introduces a broader class of universalisation preferences, and offers guidance for empirical studies.
The second chapter studies the concept of meritocracy, widely discussed both publicly and in the economics and philosophy literature. An allocation is meritocratic if more meritorious individuals obtain more rewarding outcomes. Each instance of meritocracy is characterised by two components: a merit criterion, which determines what counts as meritorious behaviour, and a reward criterion, which specifies which outcomes are more rewarding. By examining whether the allocation choices of impartial spectators align with particular merit and reward criteria, one can test the extent to which individuals adhere to different meritocratic principles. I consider two motivations for supporting meritocracy: rewarding merit as intrinsically fair—interpreting meritocracy as an end—and using meritocracy as an instrument to achieve other goals, such as efficiency—thus treating it as a means. I show that these two justifications are equivalent in terms of the rules they imply. Different assumptions about the merit and reward criteria accommodate various instances of meritocracy. I
characterise and examine two meritocratic principles found in the literature: Pareto meritocracy, in which merit derives from generating a Pareto improvement, and proportional meritocracy, in which consumption increases proportionally with effort. I conclude by distinguishing meritocracy from responsibility-sensitive egalitarianism.
The third chapter studies individuals whose well-being—their preferences over outcomes—is directly shaped by their beliefs. Such belief-dependent preferences explain a range of behaviours that deviate from expected utility theory. A growing body of evidence suggests that individuals selectively avoid or distort information, consistent with a preference for holding particular beliefs. When beliefs influence preferences over outcomes, belief formation itself may be endogenously shaped by those preferences. This interdependence complicates the task of inferring tastes and beliefs from choice data. The main contribution of the chapter is to present a model of belief-dependent preferences combined with non-Bayesian updating, and to provide choice data sufficient to test and identify the model’s components. I introduce a novel form of choice data, which generalises the notion of a menu in the menu-choice literature, and introduce axioms over preferences on such menus.
