par Moramarco, Domenico ![](/vufind/images/ULB/publications_list.png)
Président du jury Gassner, Marjorie![](/vufind/images/ULB/publications_list.png)
Promoteur De Rock, Bram![](/vufind/images/ULB/publications_list.png)
Publication Non publié, 2024-05-17
![](/vufind/images/ULB/publications_list.png)
Président du jury Gassner, Marjorie
![](/vufind/images/ULB/publications_list.png)
Promoteur De Rock, Bram
![](/vufind/images/ULB/publications_list.png)
Publication Non publié, 2024-05-17
Thèse de doctorat
Résumé : | The thesis is composed of four chapters.Chapter 1 axiomatically studies how to measure well-being when individuals have heterogeneous preferences and consumption bundles are evaluated by comparison with some reference bundle (a poverty line bundle, an index of needs, the average consumption in a reference group, etc.). The key requirement is that well-being remains constant if the reference bundle changes towards a bundle that the individual deems equivalent. Accounting for preferences may lead two individuals with identical consumption and reference bundles to have unequal well-being levels. To limit this inequality, four axioms based on the lattice structure of the set of indifference contours are studied. Three well-being measures are defined and characterized. One prominent measure corresponds to the ratio between equivalent income at consumption and at reference, with prices being chosen to maximize well-being.Chapter 2 proposes a framework to assess fairness in multidimensional distributions while respecting individual preferences. It introduces a simple measure - Equivalent Advantage - that captures the distance from the current outcome to the potentially individual specific norm outcome. A nonparametric approach is proposed to partially identify Equivalent Advantage via set identification of individual indifference curves. The methodology is illustrated by analysing multidimensional fairness in Belgium using the MEqIn database. Despite the set identification, the analysis of the Equivalent Advantage distribution allows for interesting insights on multidimensional inequality, poverty and opportunity distribution. Chapter 3 investigates the possibility of integrating a notion of fairness, inspired by theequality of opportunity literature, into the centralized school choice setting. By doing so, it widens the list of available tools to realize equality of opportunityin education. This chapter enriches the standard school choice setting with the notion of educational outcome or quality of a match between a school and a student. In this framework, fairness considerations are made by a social evaluator based on the match quality distribution. As starting point, the chapter investigates the compatibility between the proposed notion of fairness, a notion of efficiency based on aggregate match qualities, and the standard notion of stability. To overcome some of the identified incompatibilities, two alternative approaches are proposed. The first one is a linear programming solution to maximize fairness under stability constraints. The second approach weakens fairness and efficiency to define a class of opportunity egalitarian social welfare functions that evaluate stable matchings. This approach is complemented with an algorithm - the Stable Opportunity Egalitarian (SOE) - which finds the stable matching that maximizes social welfare.To show the implications and implementability of the proposal, this chapter concludes with a simulation based on Italian data. The results highlight the limitations of existing algorithms, stressing the need for alternative solutions, such as the proposed one proposed, to deal with complex distributive issues in matching markets.Chapter 4 investigates the problem of measuring inequality of opportunity using sample data. The chapter contributes to the literature by discussing the existence of potential biases deriving from violations of normative principles at the basis of the opportunity egalitarian paradigm. Interestingly, these biases differ from the well-documented estimation biases, which the recent literature addresses using machine learning techniques. The chapter proposes a way of measuring the severity of those normative biases and a methodology - the Opportunity Tree - to reduce them. An illustration based on the labor outcomes of Italian PhDs highlights the benefits of our proposal, which improves the normative content of inequality of opportunity estimates. |