Methods for forecasting the market penetration of electric drivetrains in the passenger car market

Abstract

This is the author's version of a work that was published in the following source: Jochem, P.; Gómez Vilchez, J. J.; Ensslen, A.; Schäuble, J.; Fichtner, W. (2018).Methods for forecasting the market penetration of electric drivetrains in the passenger car market .Current car technologies will not solve upcoming challenges of mitigating greenhouse gas emissions in road transport. Projections of the market penetration by alternative drive train technologies are controversial regarding both forecast market shares and applied scientific methods. Accepting this latter challenge, we provide a (so far missing) overview of methods applied in this field and give some recommendations for further work.Our focus is to classify the applied methods into a convenient pattern and to analyse models from the recent scientific literature which consider the electrification of light-duty vehicles. We differentiate the following bottom-up approaches: Econometric models with disaggregated data (such as discrete choice), and agent-based simulation models. The group of top-down models are subdivided into econometric models with aggregated data (e.g. vehicle stock data), system dynamics, as well as integrated assessment models with general equilibrium models. It becomes obvious that some methods have a stronger methodological background whereas others require comprehensive data sets or can be combined more flexibly with other methods. Even though there is no dominant method, we can identify a trend in the literature towards data-driven hybrid approaches, which considers micro and macro aspects influencing the market penetration of electric vehicles.

Simulating Electric Vehicle Diffusion and Charging Activities in France and Germany

“…Literature on the diffusion of PEV is a broad research field which has been evolving considerably over the last decade. Reviews exist for methods used to model the market penetration of PEV [16,17]. The review by Gnann et al (2018) [18] focuses on international PEV market diffusion models and compares corresponding research questions, assumptions, and results to find that there are country-specific differences in the importance of input factors.…”

“…Combining bottom-up and top-down approaches in models for vehicle diffusion and adoption has become more popular in recent years [16,[22][23][24]. Not only personal preferences influence adoption decisions, but also macro-economic parameters.…”

“…In particular, better designs of interfaces between models and surveys could lead to an improvement of PEV penetration models [17], which is, for example, addressed by Wolinetz and Axsen (2017) [25]. Disaggregated survey data facilitate forecasts of potential future market developments already in early market phases [16] and enable the analysis of effects of varying input parameters on market developments [25]. PEV penetration models based on aggregated data, on the other hand, are suitable for medium-to long-term forecasts, as long as sufficient market development data is available [16].…”

“…Disaggregated survey data facilitate forecasts of potential future market developments already in early market phases [16] and enable the analysis of effects of varying input parameters on market developments [25]. PEV penetration models based on aggregated data, on the other hand, are suitable for medium-to long-term forecasts, as long as sufficient market development data is available [16].…”

“…We use a hybrid PEV diffusion model in this study, i.e., we combine a bottom-up and a top-down approach [16] in order to better capture adopters' personal preferences as well as shifting macro-economic framework conditions [17]. As suggested by [16], the model presented in this study takes into account economic and social information (based on the bottom-up binary logistic modeling approach) as well as market development information (based on the top-down Bass diffusion model). By applying the binary logistic model, our modeling approach identifies early PEV adopters within representative mobility data sets, and their corresponding PEV charging behavior is simulated for France and Germany [41].…”