Who is sensitive to DSM? Understanding the determinants of the shape of electricity load curves and demand shifting: Socio-demographic characteristics, appliance use and attitudes

Highlights: • A cluster analysis on smart-meter data identifies typical load profiles. • Regression models are used to identify determinants of load profile and shifting. • Occupancy presence, number of wet appliances and age influence the load profile. • Financial incentives are more effective than information feedback for load shifting. • Occupancy availability during noon is a major factor enabling load shifting. -- Abstract: To date, research on demand side management has mostly focused on the determinants of electricity consumption and stated preference experiments to understand social acceptability. Further experimental research is needed to identify the determinants for demand response schemes. This paper contributes to addressing this gap by making use of data from a randomised control trial which contains 15 months of smart meter electricity data combined with household characteristics and differences in incentives to shift their electricity use between 11am and 3pm. Cluster analysis performed on electricity data identified three distinct electricity daily load profiles. Each cluster was then linked to household characteristics by means of a multinomial logistic regression to identify the determinants of the load curves' shapes. Findings show that occupancy presence at home, age and appliance ownership were strong predictors. Finally, this paper is among the first to provide experimental evidence on the determinants of load shifting. We find that households with head aged above 65, households who belong to the cluster exhibiting a load profile characterised by a relatively high peak at noon and a low peak in the evening, and those who received money incentives were more likely to shift electricity use towards middle of the day (11am-3pm).