Measuring price elasticities with a difference-in-difference design: Investigating the North-South electricity spot price gap in Norway

Abstract

This study uses spot price and municipality-level electricity consumption panel data spanning around three years to derive short run estimates of the price elasticity of electricity demand (PEED) for the Norwegian residential electricity market. A difference-in-difference model is applied on an exogenous spot price shock concentrated in the Norwegian South during 2021 and 2022 which has caused a strong deviation in electricity prices between the Norwegian price areas. A sample of municipalities adjacent to both sides of the North-South price border is assigned to a control (North) and treatment group (South). The main identifying assumption relied upon is that electricity demand in both groups would follow parallel trends under the absence of treatment, i.e., the spot price shock. After having validated that assumption, the study estimates daily, weekly, and monthly values for the spot PEED. Given indication for a lagged response when using an extension to the baseline regression model, we infer that our weekly and monthly PEED estimates better capture the response horizon in this study. Our baseline estimates converge to -0.02 for both residential housing and cabins. This implies a purchase price elasticity of demand of -0.12 for residential electricity consumption and -0.05 for cabins. The purchase PEED estimates are derived by an algebraic approximation using mean spot prices before and during the major price shock phase, averages for additional purchase price components, and a government support scheme for residential electricity consumption. The estimated inelastic demand patterns confirm previous research for the Norwegian electricity market and are consistent with the response in aggregate electricity consumption observed in this paper. Our findings are further discussed with regards to potential non-linearity of price elasticities which may depend on households’ electricity expenditure-to-income ratio. Our baseline estimates remain robust for rural Norwegian areas while providing some evidence for slightly more elastic electricity demand in urban areas. In the light of the energy transition, an increase in the efficiency of electricity use is the most straightforward answer to more natural variation in renewable power generation. However, inelastic residential electricity demand, as shown in this paper, underlines the urgency for new policy which incentivizes more flexibility in demand without causing distortions in utility and welfare.