This paper investigates, quantifies and ranks the factors influencing passenger cars energy demand and emissions. A vehicle stock-model approach is used for an ex-post decomposition analysis, based on administrative data, examining the impact seven underlying factors driving energy demand. The impact of methodological choice and model disaggregation are also explored. In light of the 2015 vehicle emissions scandal, the paper quantifies the difference between manufacturer-test vehicle performance and real world or "on-road" performance for a national stock model and determines the relative impact on passenger cars energy consumption. When examining the technical performance improvement, the choice of metric can lead to a distortion of 2.2 percentage points (14% overestimate) in the quantification of the efficiency improvement of the vehicle stock. The analysis pays particular attention to the influence of fuel or technology switching which is often quoted as a factor influencing energy use and emissions but rarely quantified. Even when using litres per hundred-kilometre gasoline equivalent to measure the performance improvement, changes in the makeup of the stock can lead to distortion in the efficiency measure. The results of a full decomposition analysis highlight that technical performance improvements (energy efficiency improvements for the purpose of this paper), will not provide significant energy and emission savings when the impact on-road consumption is included. The paper concludes that technology switching in conjunction with policies targeting ownership and usage are the most effective measures to control passenger car energy consumption and associated CO2 emissions.