Fractional integration and energy demand : A time series analysis for Latin America

Share

Abstract

In this paper, we examine energy demand in a group of Latin American regions using a fractional integration approach. Employing annual data from 1965 to 2023 on primary energy consumption in exajoules (EJ) and per capita consumption in gigajoules (GJ), we investigate the persistence and mean-reverting properties of energy demand over time. The application of fractional integration techniques allows us to capture both short- and long-term dependencies, offering a more flexible framework compared to traditional time series models. Our findings indicate that energy demand in Latin America exhibits long-memory characteristics, implying that shocks to consumption may have prolonged effects, with some countries displaying a slow mean-reverting process while others show evidence of permanent shocks. These heterogeneous results suggest that structural factors, such as economic development, energy policies, and technological advancements, play a crucial role in shaping consumption patterns. Additionally, the study highlights the importance of considering long-run dynamics in energy demand forecasting and policymaking, particularly in the context of economic growth and environmental sustainability. The results emphasize the need for adaptive energy strategies that consider the varying degrees of persistence across countries, aiming for a balance between economic development and the transition towards cleaner energy sources.