Modeling & Forecasting US Electricity Consumption

Forecasting electricity consumption has often prefered to treat socioeconomic activity and wealth as exogenous. The motivation for this research is to endogenize electricity consumption with economic activity by introducing electricity demand, and the interactions between electricity use and economic activity, into macroeconomic demand equations. In this paper, I create an energy and macroeconomic model that examines the dynamics of the US electricity markets at the sectoral level on a monthly basis. I use nearly a quarter century of monthly data to model the dynamics of several macroeconomic and electricity series, then combine these equations into a single large model. By incorporating multiple sectors of demand at once and using high-frequency data, the model provides robust feedbacks between electricity markets as well as the larger macroeconomy. The short-run is modeled simultaneously with the long-run and seasonal properties of the data using a equilibrium-disturbance technique. The equilibrium relationships are specified first at the individual equation level; these equations are then linked together to create a single large, cross-sector model. This approach allows for proper identification and specification of feedback between the electricity sectors and the economy. Using this model, I conduct several forecasting exercises to examine the impact of the 2007-2008 recession. I also conduct two sets of simulations: one scenario examines the effects of instituting a carbon price and the second examines several weather-based scenarios.