How global energy efficiency varies from the perspective of 136 countries: A spatial‒temporal analysis considering game crossings

Understanding the spatial‒temporal evolution of global energy efficiency (EE) and its influencing factors is critical for achieving sustainable development. However, existing studies lack comprehensive global-scale analyses and fail to address interregional competition in efficiency assessments. This study addresses these gaps by applying an undesirable game cross-efficiency data envelopment analysis model to 136 countries (2000–2019). Spatial autocorrelation, the Malmquist index and the Tobit model are used to reveal spatiotemporal patterns, productivity drivers and influencing factors. The key findings include the following. (1) Global EE average of 0.470. European and Oceania countries had higher levels, North America had medium and stable levels, South America showed a gradually increasing trend, and Asian countries had lower levels. (2) Spatial clustering revealed persistent high-high clusters concentrated in Western Europe, whereas low-low clusters dominated Eastern Europe and Asia. (3) Malmquist index decomposition indicated that technological progress (annual growth: 2 %) is the primary driver of total factor productivity improvement but that technical efficiency is offset by other factors. (4) Tobit regression highlighted GDP per capita and the renewable energy share as key enablers, whereas nighttime light intensity, urbanization, the industrial structure and foreign investment significantly inhibit EE. In addition, each influencing factor has heterogeneous characteristics at the intercontinental level. These findings provide a data-driven foundation for cross-regional energy policy design while equipping policymakers with spatially targeted strategies.