La mesure des inégalités. Un point de vue de statisticiens

[eng] The example of inequalities among incomes makes it possible to present a critical analysis of the technical paraphenalia statisticians have to measure inequalities. The ratio between the meansof various social groups, the coefficients of dispersions or the synthetic indicators of concentration give various possibilities to describe the gaps between incomes among one population, or two subpopulations, in various ways. The least precaution to take is to define the concepts as well as the units between which comparisons are being made, with precision. If general reports can usually be established, the diversity of points of view as well as the existence of various instruments of measure do not always give the possibility to get the same appraisal of the range of inequalities. Things get more complex when, coming to the various fields of social life, one tries to establish a scale of « standards of living », to compare the « ways of living ». The fields which are more qualitative than incomes raise special problems. But above all, the combining of various indicators concerning various fields is technically more difficult and even raises fundamental questions : how to build only one scale of reference, how to reckon the comparative «weights» of various fields ? Last, beyond the problems of measure and combinations, thinking about the difference between inequalities as they are and inequalities as they are felt, about the notion of justice (which that of inequality implicity refers to), about the necessary conflicts and cooperations and on the specificity of the groups, which cannot always be reduced to inequalities, all this leads to the conclusion that to understand social reality on must go beyond the perspective of inequalities. [fre] L'exemple des inegalites de revenus permet une presentation critique de l'arsenal technique dont disposent les statisticiens pour mesurer les inegalites. Rapports de valeurs moyennes entre divers groupes sociaux, coefficients de dispersion ou indicateurs synthetiques de concentration, les possibilites de description des ecarts de revenus au sein d'une population ou entre sous-populations sont multiples. Une precaution elementaire est de definir avec precision les concepts ainsi que les unites entre lesquelles portent les comparaisons. Si des constats d'ensemble peuvent etre generalement etablis, la diversite des points de vue possibles et des instruments de mesure ne permet pas toujours d'obtenir des appreciations concordantes de l'ampleur des inegalites. Les choses se compliquent lorsque, abordant les divers domaines du champ social, on cherche a etablir une echelle des «niveaux de vie», a comparer des «modes de vie». Les domaines plus «qualitatifs» que les revenus posent des problemes particuliers. Mais surtout, l'agregation de differents indicateurs portant sur des domaines varies est d'une difficulte technique accrue et souleve meme des questions de principe : comment construire une echelle unique de reference, comment calculer les « poids» relatifs des differents domaines ? Enfin, au-dela de ces questions de mesure et de cumuls, des reflexions sur les differences entre inegalites reelles et inegalites percues, sur la notion d'equite (a laquelle se refere implicitement celle d'inegalite), sur les necessaires cooperations et conflits entre groupes sociaux et sur les specificites de ces groupes, lesquelles ne sont pas toujours reductibles a des inegalites, conduisent a la conclusion que la comprehension de la realite sociale doit depasser la problematique des inegalites.