Regional heterogeneity in the relationship between temperature and household electricity consumption across rural South Africa

In this study we examine the heterogeneous effects of temperature on household electricity consumption using combined hourly electricity consumption and temperature data across several rural areas in South Africa. We apply a panel data regression model controlling for temperature bin dummies to capture nonlinear temperature effects, as well as a set of fixed effects to capture the unobserved heterogeneity across households and over time. We find that the temperature–consumption relationship varies across geographical locations and is characterised predominantly by inverse J -shaped and inverse U -shaped patterns, which differs from the typical U -shaped curve in the literature. This study deepens the understanding of the complex and heterogeneous effect of temperature on household electricity consumption in a developing-country context, highlighting substantial variation across climatic and geographic regions within the same country.