There has been interest among transport policy makers in initiatives that would increase the uptake of Electric Vehicles (EVs). These initiatives are obtained from output from relevant car demand models. Among the core inputs in these models are car and household variables that result in the variations in car fuel type preferences by creating the trade-offs between market alternatives, and variations on how long they would keep the cars they have chosen. In this study, the car preferences and car holding duration decisions by households have been examined by deploying several model structures. The model structures included discrete choice models for the car preferences, hazard-based duration models for the car holding duration, and a novel copula-based joint discrete choice and hazard-based duration model for the dependent decision of car preferences and car holding duration. To test the effectiveness of these models in capturing the two different decision-making behaviours by households, several primary and secondary data types were input in these models. The data types deployed to these models included retrospective preference and holding duration, stated preference and prospective holding duration data on cars. The prospective holding duration data is a novel data type that was found useful in predicting future car transaction timing decisions. The general findings from testing the different model structures demonstrated that households make decisions on car type dependent on how long they want to keep their cars. The joint model was used to test different retrospective policies that would have increased the uptake of EVs at the time of carrying out data collection.
For the policy options scenario analysis, the joint model with a component of SP added (scaled to the RP component of the joint model) was deployed to carry out retrospective assessment of policy options. The results showed that changes to the Value Added Tax (VAT) on the Battery Electric Vehicles (BEV) purchase and leasing costs result in significant additions of BEVs to the car fleet. Changes to the fuel duty on petrol and diesel cars and VAT on the purchasing and leasing cost of BEVs result in significant removal of CO2e (equivalent carbon dioxide) from the passenger cars.
As far as novelty is concerned, this thesis has contributed to two major innovations. The first novelty addresses the gap on whether households make decisions on type choice dependent on how long they plan to keep their cars. This was achieved by developing a model that combines the decisions on type choice and transaction timing (car holding duration) using copula theory. Secondly the research addresses the gap on the usefulness of data on future decisions on cars such as when to replace, dispose or add another car to their household – a data type that has been termed as ‘prospective.’ This was achieved by deploying the prospective duration data to the different model structures to test its applicability in such decisions that are likely stable due to the long-term planning associated with transaction timing decisions on cars.
In sum, this study found that four policy variables contributed to the variations on car fuel type choices. These variables are purchase cost, lease cost, annual running cost and fuelling cost. Household income plays a significant role in the addition of other cars to the existing car fleet of consumers. The decisions by car consumers on choosing a particular transaction (to replace, acquire or dispose a car) are dependent on the fuel type of the car. A 10% reduction of the purchase or lease cost of BEV cars would add about 10% of new, used or lease BEV cars. Furthermore, a similar percentage reduction of the purchase or lease cost of BEV cars would remove 0.1MtCO2e on average from private car use. That said, these findings can point policy makers into tailoring appropriate policies bearing in mind the context of this study.
