Advised by

This paper estimates the relationship between temperature and residential electricity consumption by exploiting variation in weather conditions across 13 Building Climate Zones in California. Cross-sectional household data and daily weather data are obtained from the Energy Information Administration’s 2005 Residential Electricity Consumption Survey and the National Oceanic and Atmospheric Administration’s Global Summary of the Day dataset respectively. For each climate zone, daily mean temperatures are sorted into seven equidistant bins based on the state’s temperature distribution in 2005. The estimated temperature bin coefficients along with two 21st-century climate forecasts are used to simulate changes in per-household consumption for a selected group of counties. The simulation results suggest that households located in regions that are expected to experience the largest transfer of days from lower to higher temperature bins may increase their consumption of electricity by more than 15 % over the course of the century.