From Residential Energy Demand to Fuel Poverty: Income-induced Non-linearities in the Reactions of Households to Energy Price Fluctuations

The residential energy demand is growing steadily and the trend is expected to continue in the near future. At the same time, under the impulse of economic crises and environmental and energy policies, many households have experienced reductions in real income and higher energy prices. In the residential sector, the number of fuel-poor households is thus expected to rise. A better understanding of the determinants of residential energy demand, in particular of the role of income and the sensitivity of households to changes in energy prices, is crucial in the context of recurrent debates on energy efficiency and fuel poverty. We propose a panel threshold regression (PTR) model to empirically test the sensitivity of French households to energy price fluctuations—as measured by the elasticity of residential heating energy prices—and to analyze the overlap between their income and fuel poverty profiles. The PTR model allows to test for the non-linear effect of income on the reactions of households to fluctuations in energy prices. Thus, it can identify specific regimes differing by their level of estimated price elasticities. Each regime represents an elasticity-homogeneous group of households. The number of these regimes is determined based on an endogenously PTR-fixed income threshold. Thereafter, we analyze the composition of the regimes (i.e. groups) to locate the dominant proportion of fuel-poor households and analyse their monetary poverty characteristics. Results show that, depending on the income level, we can identify two groups of households that react differently to residential energy price fluctuations and that fuel-poor households belong mostly to the group of households with the highest elasticity. By extension, results also show that income poverty does not necessarily mean fuel poverty. In terms of public policy, we suggest focusing on income heterogeneity by considering different groups of households separately when defining energy efficiency measures. We also suggest paying particular attention to targeting fuel-poor households by examining the overlap between fuel and income poverty.