Incorporating uncertainty in vehicle miles traveled projections of the National Energy Modeling System

Abstract

The National Energy Modeling System (NEMS) is a computational model that forecasts the production, consumption, and prices of energy in the United States. Although NEMS is a complex and detailed model, it does not currently represent the multitude of uncertainties associated with the US energy system. These uncertainties need to be communicated to policy makers in order for them to develop better-informed decisions regarding energy policy. In this study, uncertainty is added to the vehicle miles traveled (VMT) equation of NEMS to demonstrate the importance and benefit of uncertainty in the model. The VMT model is derived and its uncertain parameters are estimated using maximum likelihood estimation. A Monte Carlo simulation is performed to model the uncertain VMT equation and demonstrate the range of possible VMT forecasts when these uncertainties are included. This simulation shows that the deterministic forecast does not adequately reflect all of the possible futures of VMT, which could lead policy makers to be unintentionally misinformed about the impacts of proposed policies. Finally, it is shown how the uncertain VMT equation could be used to help policy makers decide on the best policy to reduce transportation greenhouse gas emissions. A target is set for VMT for each of the projected years, and four decision-making techniques are used to calculate the fuel tax required to reduce VMT to this specified goal. These methods could guide policy makers to better-informed energy policy decisions, but they are only possible if some amount of uncertainty is incorporated into the model.