Is there a Clearly Identifiable Distribution Function of Individual Poverty Scores?


  • Valerie Bérenger
  • Franck Celestini


The goal of this paper is to define a multidimensional poverty score for each household belonging to the same society in order to answer the following question: is it possible to characterize poverty, as it is income, by analyzing the nature of an associated probability density function? Based on the “Totally Fuzzy and Relative” (TFR) approach, the method proposed permits obtaining individual poverty scores lying between 0 and infinity. We apply the method to the data from the 1986-1987 and the 1993-1994 French Surveys of Living Conditions. In both cases, the probability density function of the poverty scores follows an exponential distribution characterized by a single parameter. We then examine the relationship between our multidimensional poverty score and income. The intuitive negative correlation is recovered and fully analyzed. In particular, using our poverty score distribution we estimate the poverty line that gives the best agreement between income-based and multidimensional measure of poverty.