Understanding Residential Electricity Consumption Behavior under Dynamic Pricing: Evidence from Spanish Cities

Abstract

This paper examines the short-run price responsiveness of residential electricity demand under the Spanish Voluntary Price for Small Consumers (PVPC) tariff during the 2023–2024 period, following the profound disruptions of the European energy crisis. Using hourly consumption data from 2,975 households in Madrid, Bilbao, and Valencia, we implement a Two-Stage Least Squares (2SLS) framework with nationwide wind generation as an instrumental variable to address price endogeneity. To move beyond mean-based estimates, we complement this with household-level quantile regressions, which allow us to test whether price sensitivity varies across a household's own consumption distribution distinguishing responses at low-consumption hours from responses at peak-consumption hours. Despite a 26% increase in mean prices and a threefold rise in price volatility relative to the 2016–2017 baseline, our results indicate that price elasticity of demand remains centered at zero both on average and across all quantiles of each household's consumption distribution. This finding rules out the hypothesis that households engage in price-driven peak shifting even under the significantly more volatile post-crisis price environment, and challenges the hypothesis that steeper price gradients alone are sufficient to overcome rational inattention and structural adjustment costs. Temperature is identified as the dominant driver of hourly consumption variation. Our results suggest that enabling technologies and complementary policy instruments are necessary conditions for dynamic pricing to deliver meaningful demand flexibility.

This paper examines the short-run price responsiveness of residential electricity demand under the Spanish Voluntary Price for Small Consumers (PVPC) tariff during the 2023–2024 period, following the profound disruptions of the European energy crisis. Using hourly consumption data from 2,975 households in Madrid, Bilbao, and Valencia, we implement a Two-Stage Least Squares (2SLS) framework with nationwide wind generation as an instrumental variable to address price endogeneity. To move beyond mean-based estimates, we complement this with household-level quantile regressions, which allow us to test whether price sensitivity varies across a household's own consumption distribution distinguishing responses at low-consumption hours from responses at peak-consumption hours. Despite a 26% increase in mean prices and a threefold rise in price volatility relative to the 2016–2017 baseline, our results indicate that price elasticity of demand remains centered at zero both on average and across all quantiles of each household's consumption distribution. This finding rules out the hypothesis that households engage in price-driven peak shifting even under the significantly more volatile post-crisis price environment, and challenges the hypothesis that steeper price gradients alone are sufficient to overcome rational inattention and structural adjustment costs. Temperature is identified as the dominant driver of hourly consumption variation. Our results suggest that enabling technologies and complementary policy instruments are necessary conditions for dynamic pricing to deliver meaningful demand flexibility.