Climate Change Impacts on US Electricity Demand: Insights from Micro-Consistent Aggregation of a Structural Model

B. Coffey, A. Stern, and I. Sue Wing

Working paper (30 January 2015)


We use a large dataset of 2.3 million hourly observations of load across three regional power grids to estimate the effects of weather on the demand for electric power, and use the results to draw implications for the impacts of climate change. In a novel approach, we develop a micro-founded model of individual electricity demand for space conditioning that is rooted in the thermodynamics of building energy transfer, and demonstrate how it can be aggregated up to the weather zones that constitute the smallest geographic unit within regional power pools. What results is a deceptively simple reduced-form specification that can be estimated using standard panel data econometric techniques. Compared with existing semiparametric approaches in the empirical climate economics literature, our model generates marginal effects of heat on electricity demand which are smaller at moderate temperatures but substantially larger at the extreme temperatures. The difficulties faced by semiparametric approaches in capturing such tail impacts suggest that they may underestimate the increase in demand for peak power from climate-driven heatwaves, and with this, the generation and transmission investments necessary to ensure adequate electricity supply.

cite: BibTeX | EndNote | RIS