Augmenting efficiency of public investment in the Brazilian agrarian reform

Abstract

In this paper we apply the Multiple Data Envelopment Analysis (MDEA) approach to evaluate efficiency of 90 settlements within the Brazilian agrarian reform. Previous study of these data using standard DEA approach yielded only modest discrimination among the DMUs, where roughly half were found efficient, reflecting the curse of dimensionality problem. It is shown that the novel MDEA approach breaks the tie among the DMUs that were previously found efficient, while ranking of inefficient DMUs is generally preserved.