Howto measure environmental performance: a new data-driven index

In this paper, we propose a new, entirely data-driven, and time-consistent index of environmental performance that relies on a reliable statistical technique to objectively assess environmental sustainability. While most indices depend on expert based or equally distributed weights, the proposed index — the Data-Driven Environmental Performance Index (DDEPI) — is built on empirical data, focusing on variables that account for the largest share of variance in the data over time. To construct the proposed index, we apply a dimensionality reduction technique, such as Robust Principal Component Analysis (Robust PCA), which captures the relationship between variables and countries for each year. The DDEPI is primarily driven by indicators related to air quality, drinking water quality, and biodiversity. An analysis of the causal relationships between the proposed index and country specific dimensions highlights the critical role of institutional quality and natural resource management in achieving sustainability goals.