On the Electricity-Output Nexus across the 32 States of Mexico: Insights from a Heterogeneous Panel, 1994-2024

Abstract

The question of whether electricity consumption (ELE) influences GDP, or vice versa, at the national and international levels, has been the subject of intense analysis by the academic community without a clear answer. Furthermore, additional subnational analyses are lacking to understand the underlying heterogeneity. This paper seeks to understand the dynamics between ELE and GDP in the States of Mexico between 1994 and 2024. After applying diagnostic tests such as cross-sectional dependence, panel unit roots, slope homogeneity, and cointegration, we employ advanced panel cointegration techniques, specifically cointegrating regressions (FMOLS and DOLS) and factor-augmented techniques (CUP-BC and CUP-FM), which we use as robust estimators. The empirical findings indicate a positive long-run association between ELE and GDP in most of the States in Mexico; however, the size of the estimated coefficients varies substantially across them (1.32 for Campeche and 0.103 for Michoacán). The factor-augmented estimators confirm that ELE has a coefficient of around 0.3. In addition, this analysis adopts a heterogeneous non-Granger test to understand short-run causalities. This test shows that ELE causes GDP only in Coahuila, Tlaxcala, and Sinaloa, and GDP causes ELE in Campeche and Yucatán, with weak evidence for Chihuahua and Puebla. However, at the panel level, we obtained weak support for the growth hypothesis between ELE and GDP, whereas strong support for the neutrality hypothesis prevails. Overall, the findings suggest that subnational energy policies should account for State-specific factors relevant to long-term growth and sustainability goals. These results have important implications for State policy design, suggesting the need for State-specific approaches rather than one-size-fits-all solutions.