Measuring Poverty: An Interval Based Composite Indicator Approach

Abstract

The measurement of the poverty is a fundamental theme in Social Science. Poverty as a multidimensional concept can be measured by means of composite indicators which are relevant for their capacity to synthesize complex statistical information. However, these indicators could be affected by subjective choices. In order to avoid the problem, we propose interval based composite indicators which allow, in this context, to obtain robust and reliable measures. In particular, we will consider all the different factors identified on the basis of a relevant conceptual model of the poverty, we have identified. Then we will compute a different composite indicator considering a different random configuration of the different factors. At the end we are able to obtain a different interval for each region based on the different factor choices. In this sense, we will construct an interval based composite indicator based on the outcomes computed. The different intervals can be compared and different rankings for poverty can be obtained. In fact, the poverty interval composite indicator can be considered and compared for their parameters like centre, minimum, maximum and range. The results show a relevant and consistent measurement of the indicator, but also a relevant impact of the shadow sector on the final measures.

Keywords: poverty, composite indicators, interval data, symbolic data

JEL Classification: I32, I3, C43, C02