Sensitivity of environmental performance index based on stochastic dominance

MEHMET PINAR*

*Corresponding author for this work

Research output: Contribution to journalArticle (journal)peer-review

1 Downloads (Pure)

Abstract

The Environmental Performance Index (EPI) is a popular sustainability index. It is a composite index which ranks 180 countries based on their environmental performance in 32 indicators. Using the EPI data and stochastic dominance efficiency methodology, this paper examines the sensitivity to the subjective weights assigned to the indicators and categories of environmental performance. The findings show a remarkable variation in environmental performance based on alternative weights which are selected using Stochastic Dominance criteria. Except for 2020, the environmental health category in EPI gets relatively higher importance in the optimal scenario, and ecosystem vitality gets relatively higher weights in the inferior scenario, suggesting that the environmental health category achievements have been relatively higher for most countries over time. The ranking analysis also shows major variations in country rankings with alternative weights. Two countries, Maldives and Gabon, would have experienced more than 100 position changes in their rankings with alternative weights. Furthermore, 67 countries would have experienced 30 or more position changes in their rankings and 37 countries experienced an EPI score change of more than 50 (out of 100) with alternative weights. Overall, the results illustrate the importance of sensitivity analysis of composite indices to increase reliability and transparency.
Original languageEnglish
Article number114767
Pages (from-to)114767
JournalJournal of Environmental Management
Volume310
Early online date22 Feb 2022
DOIs
Publication statusPublished - 15 May 2022

Keywords

  • Composite index
  • Environmental performance
  • Indicators
  • Stochastic dominance

Fingerprint

Dive into the research topics of 'Sensitivity of environmental performance index based on stochastic dominance'. Together they form a unique fingerprint.

Cite this