DIW Discussion Papers 2175, 40 S.
Sophie M. Behr, Till Köveker, Merve Kücük
2026. A previous version of this paper was published 2025 as DIW Berlin Discussion Papers 2112.
get_appDownload (PDF 1.19 MB)
The 2022 energy crisis led to sharp energy price surges and prompted public appeals and government programs to reduce energy consumption. This paper isolates energy savings driven by retail price increases from non-price driven savings attributable to other crisisrelated factors such as public appeals. Using German residential energy billing data and a DiD-PSM approach, we identify average price-driven savings. We employ machine learning to estimate building-level price-driven and non-price-driven savings, and analyze their variation with socio-economic characteristics using census data. Our findings reveal that energy savings were driven by non-price factors instead of price increases, and homogeneous across socio-economic characteristics.
Topics: Business cycles, Energy economics, Digitalization
JEL-Classification: Q41;Q48
Keywords: Energy crisis, Energy policy, Causal inference, Double machine learning