Measuring vulnerability to multidimensional poverty in Latin America

Latin America is not the poorest region in the developing world. It is, however, a region with high inequality, precarious institutional frameworks and high exposition to covariate and idiosyncratic shocks. In this paper, with a sample of more than seven million observations, we perform estimates of vulnerability to multidimensional poverty for 17 Latin American countries at three points in time: 2005/6, 2012 and 2017. We use a multidimensional Bayesian network classifier model to estimate the conditional probability of being multidimensionally poor. We then use these probabilities and the standard downside semi-deviation as the risk parameter to identify the vul- nerable households. Our findings suggest that, despite significant reductions over the study period, in 2017, approximately 200 million people – about the size of the population of Brazil – continued living at high risk of be- coming poor or remaining multidimensionally poor. We also observe that vulnerability to poverty is reduced at a much slower rate than poverty itself, revealing that poverty reduction accomplishments can actually be quite frag- ile. Additionally, we perform a decomposition between poverty-induced and risk-induced vulnerability and find that as poverty decreases, risk-induced vulnerability becomes relatively more important than poverty-induced vulnerability. However, it is the poor-vulnerable group that still constitutes the core vulnerability group.