Residential Space Heating Determinants and Supply-Side Restrictions: Discrete Choice Approach

This paper provides an empirical analysis of the supply-side constraints that impact household choices of space heating systems in French dwellings. Based on data from the 2006 National Housing Survey and the 2013 Household Survey, we estimate discrete choice models using socio-demographic, dwelling and spatial characteristics as determinants of choice. In order to capture the supply-side constraints, we perform a post-estimation clustering based on the Expectation-Maximization algorithm, which allows us to determine the groups of households that are most likely to choose each considered space heating system. The results suggest that for 2006, households living in individual houses preferred to heat space using an individual boiler, whereas those living in apartments opted for direct electric heating. In 2013 all households preferred direct electric heating, and wood heating became their second choice. The discrete choice models do not show a significant change in behavior from 2006 to 2013. The post-estimation clustering indicates that many households were strongly constrained by the conditions and characteristics of their living space in 2006. The mean probability of choosing direct electric heating or a boiler was more than 0.8 for 21% of households living in houses and for 36% living in apartments. Thus, these households did not seem to have a choice. In 2013 the supply-side constraints were weaker. The highest mean probability (0.755) is observed in the group opting for an individual boiler. We also observe an increase in the popularity of heat pump and wood heating systems.