Does The Specification of A New Class of Poverty Measures Matter? Evidence from Tunisia

The evaluation of the horizontal inefficiency of an anti-poverty design is often reduced to the determination of the type I errors, which occur where eligible individuals are not awarded benefits.  Because under-coverage ratio does not consider the social cost resulting from unequal treatment of like individuals, it is irrelevant in assessing the severity and the depth of horizontal inefficiency.  Also, when the cost of inequality approach is adopted to derive from a poverty measure, respecting the transfer axiom, a cost of inequality that is decomposable into two components, corresponding to vertical and horizontal inequality respectively, it is no longer possible to have different aversions toward these two forms of inequality.  We follow then the cost of inequality approach after specifying a new class of poverty measures, which are parameterized by two coefficients allowing so different preferences toward these two equality principles. When these two coefficients are identical, the new poverty measures class reduces to the Foster, Greer and Thorbecke’s (1984) class, whose poverty measures imply the same aversion to vertical inequality and horizontal inequity.  Further, for a given poverty line, the new class enables to characterize the set of poverty measures in which policymakers are indifferent between the post-reform poverty alleviation program and the status quo. \