Resumen: This paper studies the worldwide evolution of energy intensity for a large sample of countries during the period 1990–2015, differentiating between non-renewable, non-clean and total energy intensity. This division allows us to establish more precise policy recommendations which, along with the use of the Phillips and Sul (2009) methodology, provides the novelty of the analysis. Our results refute recent evidence favouring the hypothesis of global convergence for all types of energy intensity. Grouping countries either by regions or by income level, the evidence against this hypothesis remains overwhelming, with very few exceptions. Nonetheless, we can observe the presence of several convergence clubs, whose creation strongly depends on energy prices as well as on external energetic dependency. In any event, results relating to the different types of energy intensity are varied, suggesting that previous policy recommendations aimed at tackling climate change based on total energy intensity analyses are questionable. Idioma: Inglés DOI: 10.1016/j.eneco.2020.104738 Año: 2020 Publicado en: ENERGY ECONOMICS 87 (2020), 104738 [12 pp.] ISSN: 0140-9883 Factor impacto JCR: 7.042 (2020) Categ. JCR: ECONOMICS rank: 12 / 375 = 0.032 (2020) - Q1 - T1 Factor impacto SCIMAGO: 2.5 - Energy (miscellaneous) (Q1) - Economics and Econometrics (Q1)