Asac 2008 Conference Conceptualization of Principal Variable Selection for Computing the Water Poverty Index

This paper develops a conceptual framework resulting in an alternative method for selecting principal variables for computing a Water Poverty Index (WPI) at the national scale. To this end, a principal components variable selection strategy is applied to the 2002 data obtained for 147 countries on five key components resource, access, capacity, use, and environment compiled by Keele University. The results suggest that only three components (i.e., access, capacity, and environment) with different weights (i.e., the highest weight to capacity and the lowest weight to environment) be used for WPI calculation. It also turns out that a simpler index, based on two components (i.e., capacity and environment) with equal weights, can be computed with little information loss. The new WPIs, which are shown to correlate well with two well-known socioeconomic indicators, could help government, non-government, and business organizations set priorities for communities and countries of greatest resource needs.