Car Purchasers Preferences for Hydrogen and Plug-in Electric Vehicles
J Jordal-Jorgensen, COWI A/S, DK
Stated Preference survey of car purchasers demand for Hybrid, Hydrogen and Plug-in Electric vehicles. Besides traditional WTP results, the paper also compares different methods for including attitudes and perceptions as latent variables in the model.
The paper presents the main findings of a Stated Preference survey of Danish car purchasers' preferences for future technology vehicles. The objective of the study was to obtain household preferences for alternative-fuelled vehicles and to identify trade-offs and incentives for households to switch to cleaner vehicle technologies. The survey included five technologies:
- Conventional gasoline or diesel
- Bio fuel
- Plug-in Electric.
In each choice experiment, respondents were asked to choose between a car similar to the one they recently purchased and a hypothetical car with new technology and slightly different attributes. The attributes were purchase price, annual fuel cost and maintenance cost, operating range, acceleration, service agreement, local air pollution and greenhouse gas emissions.
The monetary attributes were customized to each respondent. The annual fuel cost was computed as the product between the annual number of kilometres travelled by the vehicle, as specified by the respondent earlier in the survey, and the fuel efficiency of the vehicles in the experiment.
The samplings of alternatives were based on a random variation in attributes based on a uniform probability. This may not be the most efficient design. However, since we asked a large number of respondents, parameter estimates were estimated with high precision and significance. Furthermore, using this approach we were able to analyse all parameters and interaction effects.
The preferences revealed in a stated preference experiment are manifestations of the underlying preferences and thus serve as indicators of preferences. Beyond that, we have collected indicators of attitudes and perceptions, such as responses to attitudinal and perceptual questions in the survey. These indicators were helpful in model identification and increased the efficiency of the estimated choice model parameters.
There are several methods for including latent variables in the analysis of the choice model. The paper compares the results of four different methods for including attitudes and perceptions as latent variables in model formulation:
- Inclusion of indicators directly in the utility function
- Factor analysis followed by a Choice Model (segmentation)
- Sequential latent variable model
- Integrated Choice and Latent variable model.
Estimations were based on a mixed logit model with panel specification developed by Arne Rise Hole and the GLLAMM procedure developed by Skrondal, A. & Rabe-Hesketh.
The study was based on a stated preference survey among a large number of randomly selected new car purchasers. Approximately 18,000 randomly selected new car purchasers received an invitation to participate in the survey. Approximately 6,000 filled in the questionnaire on the Internet, corresponding to a response rate of 32%, which is slightly higher than the average response rate in such surveys. A smaller number (approximately 50) requested the paper version of the questionnaire.
The main results presented in the paper are the car purchasers' willingness to pay for cars with new technology, the willingness to pay for increased operating range of plug-in electric vehicles. Another interesting finding discussed in the paper is the variations in the willingness to pay for environmental-friendly vehicles across different segments of car purchasers.
Finally, the paper report on different methods of including attitudes and perceptions as latent variables in the model formulations. In order to obtain detailed information about views on new technologies, respondents were asked about their perception of the new technologies. To validate the survey and parameter estimates, an investigation was made of the possible correlation between respondent's perception of the new technologies and their rationale for choosing between the new technologies. Model estimations show that there is a very clear correlation between respondents' perception of the technologies and their choice in the choice experiments.
Association for European Transport