Abstract

Comparison between revealed and stated preference household vehicle choices

David Brownstone
University of California, Irvine

A new panel survey of California households' vehicle purchase and
utilization behavior provides a unique opportunity to test the consistency
of stated preferences with actual behavior. 2857 households were surveyed in
early Fall of 1993 and again in late Fall of 1994.  Detailed socioeconomic
and vehicle transaction and utilization data were collected along with their
responses to two different stated preference tasks.  The stated preference
tasks were designed to elicit households' preferences for alternative-fuel
vehicles, but there were enough gasoline-powered vehicles in the choice set
to identify stated preference models of choice between existing gasoline
vehicles.  I first compare predicted vehicle transactions from these stated
preference data with actual transactions.  I also compare willingness to pay
estimates from similar models estimated on stated preferences and actual
transactions observed between panel waves.  Although the power of these
comparisons is hampered by the small number of observed transactions, the
stated preference models exhibit few systematic biases.