Assessing the Predictive Power of Vulnerability Measures: Evidence from Panel Data for Argentina and Chile


  • Marcelo Bérgolo CEDLAS, Universidad Nacional de la Plata, and IECON, Universidad de la Republica
  • Guillermo Cruces CEDLAS-UNLP, CONICET and IZA
  • Andrés Ham University of Illinois at Urbana Campaign


This article carries out a validation exercise of vulnerability measures as predictors of poverty at the aggregate and micro levels based on short and long term panel data for Argentina and Chile. It then compares their performance to that of deprivation indicators. The main findings indicate that while vulnerability measures are good predictors of poverty in the aggregate, the same does not occur at household level. These results imply that while useful, vulnerability estimates require incorporating shocks to attenuate biased estimates if they are to be used for targeting purposes.

Author Biography

Guillermo Cruces, CEDLAS-UNLP, CONICET and IZA

Deputy director, Center for Distributive Studies, Universidad Nacional de La Plata

Professor of Economics, UNLP Economics Dept.

Research affiliate, IZA