Dynamic modelling of household car ownership

Abstract

All over the world, governments are facing sustainability challenges, which are among others related to car use and ownership. Car ownership research is traditionally mainly focussed on the effect of socio-demographic characteristics, like age, income and having a driving license. More recent research showed that life events related to work, family and residential location, and built environment factors are other promising explanatory factors of changing car ownership. However, the Dutch car ownership models do not incorporate these factors adequately, due to a lack of data. Next to that, it was found that it is theoretically better to model car ownership in a dynamic way (taking into account car ownership of last year). Since such dynamic models are still a rarity in the literature, this thesis examines to what extent the forecasts of household car ownership in the Netherlands are affected by the inclusion of life events and built environment factors in a dynamic model. To this end, the dynamic model is enriched in two steps: first in terms of structure (static versus dynamic), and secondly regarding substance (without or with the effect of life events and the built environment). The step from a static to a dynamic model (by using a transaction choice model) resulted in a significantly higher model fit (p<0.001), which was true as well for including life events and built environment factors. A variety of life events were found to affect car transaction behaviour, especially those with a spatial component (relocation and job transition). Furthermore, the absence or free nearby parking is found to reduce car ownership.The forecasts of the dynamic model were especially on a short term better in line with the historic car ownership trend than that of the static model. Including the effects of life events and built environment factors in the dynamic model resulted in a lower car ownership trend (up to 1.4% lower), while a similar effect was found of increased paid parking on a national level as well as on the level of Rotterdam. Therefore, reduced free parking at the street – especially near new business and residential areas – and campaigns aimed at people experiencing the habit-breaking effect of life events, are policy measures with the potential to reduce car ownership and to work towards a more sustainable and accessible future.