CLUSTER-ANALYSIS FOR MEASURING WELFARE AND QUALITY-OF-LIFE ACROSS COUNTRIES

The dream of a multiattribute approach to empirical welfare analysis is coming closer to reality because of significant advances in both theoretical and measurement areas. In this paper we propose statistical cluster analysis methods to explore different ways and levels for clustering of 23 diverse attributes such as political rights, civil liberties, life expectancy, literacy, real domestic product, etc. Several key attribute categories are identified which should be very helpful in dealing with the problem of 'double counting' in multidimensional analyses. Aggregate measures of welfare are computed for many countries, and the sensitivity of several key inequality measures with respect to clustering criteria and degrees of aggregation is studied.