Multidimensional Poverty with Missing Attributes

Abstract:     
Suppose that n quantifiable attributes of an individual - e.g. income, life expectancy, education etc. - are important for deciding whether or not that individual is living in poverty. If his profile of attributes falls below a certain poverty frontier then the individual is considered poor. Consider the problem of a policy-implementer who can only observe k < n of these attributes. The bureaucrat wishes to choose a poverty frontier only in those attributes that are observable, so as to minimize the expected number of misclassified individuals. We characterize the solution to this minimization problem.

Number of Pages in PDF File: 11

Keywords: multidimensional poverty

JEL Classification: D63, I32

working papers series