Une introduction à la micro-économétrie de l'évaluation

The purpose of this article is to give the simplest presentation of the methods of what is called «evaluation in microeconometrics» when two groups of observations are at disposal : the first one composed of individuals having benefited from a public policy device and the second one of individuals acting as control. Two polar cases are presented to identify the causes which would lead to an overvaluation or an undervaluation of the effects of the public policy. Then we explain in detail the methods of difference in differences, the methods of matching and eventually the methods which try to put right unobservable heterogeneity. We also show that all these methods can be understood as methods of regression. The improving of evaluation in microeconometrics lies within the increasing demand for accuracy which can be noticed throughout the fields of econometrics.