This study targeted existing e-bike owners and investigated their future car purchase plans with the goal of identifying geographic, built-environment, socio-economic, transportation and environmental variables that influence car purchase decisions, in the past and future. This paper adds to the growing body of literature investigating motorization, focusing on household- and city-level factors, rather than national trend analysis. We found that household variables dominate the models, with few exogenous variables significantly influencing purchase decisions. Car ownership decreased the chances of purchasing a car. Duration time from first motorized vehicle purchased increase the chances of purchasing a car. Household income and number of licensed drivers increase the chances of purchasing a car. High e-bike ownership increased the chances of purchasing a car which indicates that e-bikes may not be the terminal or substitutes for cars of the motorized pathway.
There were a few regional differences as well, with weaker demand in Northeast China. Northwest China and South China have relative stronger car purchase intentions. High taxi density, bus density, population density, and urbanization reduced the likelihood of purchasing a car, meaning advanced public transit could be temper car ownership. Also, advanced public transit usually is built in the bigger cities that may often have more strict regulations or economic barriers to control car ownership. Of the environmental variables, only the number of cold days had significant effect on car purchase decisions. More restrictive e-bike policies did not influence car purchase. This could be because restrictive e-bike policies might not be strictly enforced; and cities with restrictive e-bike policies tend to have high taxi and bus density, which counters the influence of policy on car intention.
Many of these relationships corroborate other studies. However, many of the models’ insignificant variables are also important. This analysis builds on other studies and contributes to the evolving understanding of how China motorizes, particularly among a large subset of transportation users, existing owners of e-bikes. Findings in this study can assist policy makers in identifying factors that are within their control (e.g., transit service) and factors that are beyond their control (e.g., geography) when developing approaches to control motorization.
There are four main limitations to this study. First, the limited sample (based on available sampling data ability) of only e-bike owners does not allow us to estimate the true effect of e-bike ownership on vehicle purchase behavior, i.e., we cannot compare to non-e-bike owners. Nonetheless, we model motorization within a subset of total transportation system users; an important subset on the verge of adopting fully motorized vehicles. Secondly, because this survey was a telephone survey across China, our sample was limited to the dataset spanning 2008–2012 to assure reliable contact information (we attempted to contact e-bike consumers from before 2008, most of the phone numbers were disconnected). Thirdly, more data should be collected in the future to improve the models’ accuracy. The small sample size limits the power of the models. Last, the authors did not consider car ownership cost in this study as an explicit variable, which certainly has a significant influence on car ownership model. Controlling for variable car ownership costs across the sample is an area of future improvement to the model