This paper estimates residential electricity and gas price elasticities during the recent energy crisis, a period of exceptionally large price increases, using VAT-inclusive prices that reflect both market dynamics and policy-induced variation. We compare a standard constant-elasticity (log-log) specification with two regime-switching approaches: a structural break model distinguishing pre- and post-crisis elasticities, and a threshold model allowing elasticities to vary with the price level. Because prices are nearly flat outside the crisis, the constant elasticity estimate is dominated by crisis-period variation and may poorly describe responses under normal market conditions. The two regime-based models reveal different patterns. In the structural break model, responsiveness is larger after the onset of the crisis for both electricity and gas, and for gas it is undetectable before the crisis; this cannot be separated from other crisis-related factors. In the threshold model, elasticities are significant in both regimes but smaller in magnitude at high prices for both markets, consistent with weaker responsiveness during high-price periods. These results suggest that the two specifications capture distinct dimensions of price responsiveness (price level versus price variance), and that the choice of regime-identification strategy shapes which type of response is detected.
Hindriks, J., Serse, V., & Paré, L. (2026). Revisiting Energy Demand Elasticity: The Power of Regime Switch (LIDAM Discussion Paper CORE 2026/06). https://hdl.handle.net/2078.5/275758