Working Paper
Measuring Vulnerability to Poverty

Many argue that poverty is intimately linked with ‘vulnerability’. Still, there is no consensus about how to define and measure ‘vulnerability’. We review theory and describe strengths and limits of recently proposed measures. We then propose a definition of vulnerability and develop a general empirical framework that combines Monte Carlo and bootstrap statistical techniques. The approach estimates the expected distribution of future expenditures for each household and then calculates vulnerability measures as a function of those distributions. The approach addresses weaknesses in existing methods, and can be implemented with panel data. An application to Côte d’Ivoire in 1985–86 shows that by our definition there was considerable vulnerability in the cities outside of Abidjan, a finding obscured by existing methods.