Since the mid-1970s, the federal government has produced and published information on the fuel economy of new passenger cars and light trucks as required by statute (EPA, 2017a). Beginning in 1999, the Department of Energy (DOE) and Oak Ridge National Laboratory (ORNL) established a website to more efficiently provide fuel economy information to the public. This report reviews the methods and data used by DOE/ORNL to estimate the dollar and petroleum savings due to their fuel economy information efforts. We find that the methods used to estimate savings are reasonable and appropriately transparent, given the existing state of knowledge and available data. Where possible, we updated time-sensitive data. However, empirical data are lacking for many of the key steps in the estimation process which forced DOE/ORNL to make plausible assumptions about important parameters. In general, conservative assumptions have been used, reflecting a preference for under-estimating the program’s benefits. In a few cases we recommend even more conservative estimates. We carried out an analysis of the sensitivity of estimated benefits to uncertain parameters. Parameters were grouped into three classes and assigned approximate uncertainty ranges: 1) well supported by empirical data (+/- 10%), 2) supported by incomplete or indirect empirical evidence (+/- 20%) or 3) plausible conjectures (+/-50%). Ten thousand Monte Carlo simulations each were run using first uniform and then triangular probability distributions. Estimated dollar savings in 2016 due to the program’s activities in that year alone ranged from a minimum of $40 million to a maximum of $220 million. Estimated savings in all future years due to the program’s 2016 activities ranged from $230 million to $1.7 billion. Petroleum savings and estimated greenhouse gas (GHG) reductions were of similar magnitude. We consider the finding that the impacts of the program are orders of magnitude larger than the program’s cost to be reasonably robust. Sensitivity analysis indicated that assumptions about the fraction of new car buyers visiting the website www.fueleconomy.gov who changed their new car choices as a result of their visit had the greatest impact on estimated benefits. Next in importance was the increase in fuel economy resulting from information obtained from the website. We make specific recommendations concerning periodic updating of data and the use of online surveys of website customers to develop empirical support of key parameters.
