Between-mode Differences in the Value of Travel Time: Self-selection or Strategic Behaviour?
M Fosgerau, K Hjorth, S V Lyk-Jensen, Danish Transport Research Institute, DK
The paper tests whether between-mode differences in the value of travel time can be explained by self-selection or result from respondent?s strategic behaviour. We find that self-selection seems to explain most of the observed differences.
Our point of departure is the finding of large differences between modes in the value of travel time (VTT) found from a stated choice experiment involving travel time and cost with more than 6000 respondents. From a series of within-mode experiments for the current mode of respondents, we find a VTT for car drivers several times larger than the value for bus passengers, while the value for train passengers is in between. Such results give cause for concern but are not unique to our data. We find the differences cannot be explained by differences in income and comfort.
The paper advances two competing explanations for the observed differences: Strategic behaviour and self-selection. The self-selection story is convenient, since it allows us to accept that we are in fact measuring preferences and hence the VTT that we intend to measure. The strategic behaviour story is very inconvenient, since it implies that the data do not reveal preferences. The paper aims to provide empirical evidence to discern between these two competing explanations.
Under strategic behaviour, respondents think outside the context of the experiment and consider their ability to influence political decisions. For car there is no established mechanism whereby respondents could actually pay for reduced travel times. Respondents may therefore feel that it is a free lunch to express a wish to pay for increased speed. Conversely, public transport passengers pay fares set by political decision while travel times may be deemed difficult to change as they are determined by traffic conditions and not politically. Passengers may hope that expressing a low willingness to pay may influence the setting of fares. In both cases, choices cannot be seen as an expression of preferences. Under self-selection we hypothesise a distribution of the VTT in the population with the individual VTT being only partly explained by observed variables. Those with high VTT may, ceteris paribus, choose the fast modes, car and then train. Those with low VTT tend to choose slow modes. So the differences in VTT that we see may be due to self-selection.
In our data, respondents are interviewed to identify a recent trip. They then carry out a within-mode stated choice exercise for this trip in the current mode. This is where we find the large differences between modes mentioned above. Respondents are then asked to identify an alternative mode for their current trip and a similar experiment is carried out within this alternative mode.
Under the self-selection hypothesis we would expect respondents to carry their unobserved VTT with them to the alternative mode. We would thus expect current car drivers to have a high VTT also in bus as an alternative and conversely for bus passengers. On the other hand, if responses are strategic, we would expect a large decrease in the VTT of car drivers as they go to bus and the converse for bus passengers. Our data thus allow an empirical test of the two competing hypotheses.
After controlling for the choice of mode we find that self-selection into modes seems to explain most of the observed differences. Expectations concerning comfort are also present as current car drivers have larger VTT in train and current train users have lower VTT in car.
Association for European Transport