Some approaches to the construction of a sustainable development index for the Philippines

Abstract

Principal Component Analysis (PCA) is a convenient tool used in aggregating the indicators of sustainable development and providing indices where different weights are assigned to the various indicators. There are, however, problems in interpreting of indices, especially if time series data are used. This study explores the feasibility of applying recent developments in PCA of time series using Philippine data. We present the comparative advantages of SPCA (Sparse Principal Component Analysis) relative to averaging of an adequacy/inadequacy index and PCA in index construction from various indicators of sustainable development in the Philippines in terms of usefulness and validity of indices being developed. SPCA can attain sparse and non-overlapping loadings without losing a large amount of explained variance compared to PCA. Because of the non-overlapping contribution of variables in SPCA components, indices can have clear and mutually exclusive meanings, facilitating interpretation. Even with a more complicated algorithm, reduced dimensions and simpler interpretation of indices justify the advantages of SPCA over PCA in index construction. The indices are interpreted in terms of the milestone of sustainability in the Philippines. The resulting indices provide an adequate summary of the sustainable indicators and evidence of the importance of leadership and political will in sustainable development.