“Beyond Human Development Index: a stochastic spanning approach”

The well-known Human Development Index (HDI) goes beyond single measure of wellbeing comparisons by obtaining a composite index by combining achievements in education, income and health dimensions. Yet, many still argue that dimensions covered by the HDI do not reflect overall well-being of societies and new indicators should be included to the HDI. This paper offers a new methodology, stochastic spanning, to test the inclusion of new dimensions to the HDI. This methodology takes the current measurement of HDI as a benchmark and tests whether spanning occurs or not by examining the distributions of composite scores obtained from a particular expansion of the feasible choice set or not. We use this methodology to test the inclusion of indicators of governance to the benchmark components of the HDI and find that inclusion of some governance indicators to the list of components of the HDI leads to welfare improvements.