ESTIMATION PAR INTERVALLE DU SEUIL ET DU TAUX DE PAUVRETE PAR LES METHODES DE LA SIMULATION

Abstract

Until recently, the academic papers and researches that had focused on the evaluation of the threshold and poverty rate had only considered getting a punctual estimator of these two parameters. This evaluation was done by extracting a sample from a given population. Actually, statistics takes into consideration the reliance/trust that one can have in such estimators that one has gotten starting from a sample and the using them instead of unknown parameters. Statisticians consider that one cannot rely on punctual estimators; thus, they suggest using interval estimators instead. The problem that arises when calculating an interval of reliance for these two parametersthreshold and poverty rate- is the fact that one needs a large number of samples to calculate estimators for the above mentioned parameters. Using more samples implies higher costs, especially when the researches look into finding out the financial status of a large group of people. On the other hand, it’s not clear how the Algerians will declare their real assets. In this article, we suggest a new methodology to resolve these issues. This methodology is based on the simulation theory and its main strength is the fact that it only builds on one sample. Therefore, the probabilistic laws that generate more samples through computer simulation derive from this one sample. For each individual sample one can estimate the threshold and the poverty rate. Thus, more estimators, that can be useful in evaluating an interval of reliance for the threshold and the poverty rate of the studied population, are obtained.