What Sustains Nuclear Power? A Joint ESG Analysis of the Electricity Mix in Twenty-Nine Nuclear Economies

Résumé fourni par la source

Nuclear power is almost universally treated in the empirical literature as a determinant of emissions rather than as an outcome to be explained. This study reverses the direction of analysis and asks which environmental, social and governance conditions sustain nuclear generation within a national electricity mix. Using a balanced panel of the twenty-nine countries operating civil nuclear capacity between 2004 and 2020, comprising 466 observations, we estimate a single equation in which all three ESG pillars enter jointly rather than the pillar-by-pillar specifications common in this field, and we document that separate estimation reverses the sign of three of nine coefficients. Six estimators are compared: Pooled OLS, weighted least squares, random effects, two-way fixed effects, Between and correlated random effects. Fossil dependence is negative and significant at the one per cent level in every specification, consistent with technological lock-in. Water stress carries a positive coefficient, reflecting industrial concentration rather than resource abundance. Population ageing displays an inverted-U profile with a turning point at 14.3 per cent, and female parliamentary representation is negatively associated with the nuclear share, providing the first aggregate evidence consistent with the gender gap in nuclear risk perception. Electoral democracy reverses sign across dimensions, positive between countries and negative within them. All three pillars are jointly significant in five of six estimators. An unsupervised classification identifies four national configurations accounting for half of the cross-country variance, and country-blocked cross-validation shows that the linear specification outperforms every flexible algorithm, whose apparent superiority under random partitioning is an artefact of panel leakage.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.