How to integrate real-world user behavior into models of the market diffusion of alternative fuels in passenger cars - An in-depth comparison of three models for Germany.

The future market diffusion of alternative fuels in the passenger car sector is of great interest to both carmakers and policymakers in order to decrease CO 2 emissions. The decision to buy a car is not totally objective and only partly based on cost. For this reason, those modeling the future market evolution of cars powered by alternative fuels try to include behavioral and non-cost related aspects. This paper analyzes the integration of user behavior into market diffusion models and compares three models that include this aspect. The comparison comprises three parts: first, it compares the modeling approaches, then uses a harmonized data set to model the future market diffusion of alternative fuel vehicles, with and without behavioral aspects. The most important aspects of user behavior included in the models are the use of charging infrastructure, the limited model availability, the consideration of range anxiety as a hampering factor or the willingness-to-pay-more for alternative drivetrains as a supporting factor, as well as a distinction of users' driving distances. User behavior is considered in various ways, but always has a limiting effect on electric vehicle market diffusion. While a model that distinguishes individual users and driving distances stresses the high relevance of this aspect, it is considered less important in models with a more aggregated inclusion of user behavior based on logit functions. • Integrating user behavior into models reduces sales of alternative fuel vehicles. • High market shares of electric vehicles in 2030 (up to 50%) nonetheless. • User behavior important, but policies, energy prices and purchase prices also matter. • Individual effects of user behavior can be better analyzed in disaggregated models. • Aggregated models are less sensitive to changes in user behavior. [ABSTRACT FROM AUTHOR]