Temperature Response Functions for Climate Impact Assessments: The Case of ENSO and Electricity Consumption

We propose a simple regression-based approach to creating counterfactual temperature responses based on high-frequency temperature observations and low-frequency economic responses generated by way of a nonlinear temperature response function (TRF) or cross-TRF interacted with covariates. Our approach allows identification of components of temperature or other climate variables and hence the economic damages attributable to specific climate forcings or sources of natural variability. We apply this methodological approach to estimating the impact of the El Nino Southern Oscillation (ENSO) on electricity consumption in Washington state. By comparing the actual temperatures and responses against the counterfactuals, we find that La Nina raises and El Nino reduces electricity consumption, and hence the effect of ENSO alternates between cost and saving, which may be as high as 8.5% per month in a given month. Averaging costs and savings since 1990, ENSO presents a very small net saving (ă0.001% per month) in electricity consumption in the state.