The μ-values in table 6 are the threshold parameters. Since the equation does include a
constant term, one of the threshold parameters is not identified. We normalize the first to 0.
The reason for having only 2 μ values is that some of the 7 response categories (see section 3)
had to be combined to reach an acceptable data fill in each class. A further general
characteristic is that the table makes a distinction between personal, work and trip related
factors on the one hand and variables related to perceptions or behavioural changes on the
other hand.
First of all looking at the personal, work and trip related characteristics, the result that people
with a high household income seem to have a higher chance of changing the residential
location is somewhat strange. This is in contrast to some other estimation results, not
presented here. Therefore this result must be handled with care. Respondents that live alone,
own a house, work more than 35 hours per week and who get a travel cost compensation by
their employer seem to have a lower probability of changing due to the road pricing measure.
Furthermore, respondents living in a region (of Holland) suffering from traffic congestion
problems are found to have a relatively lower probability of changing house due to road
pricing. This can partly be explained by the substantial lower commuting distances in the
sample for people living within these ‘congested regions’. And toll costs off course are in case
of a kilometre charge linearly linked to distance. Respondents living in a bigger city have a
lower probability of changing due to the pricing measures. The same goes up for respondents
driving in a gasoline car. Gasoline car drivers driving fewer kilometres on a yearly basis than
diesel car drivers can partly explain this last result.
As expected, respondents that indicated they would (in general) be better of due to the
introduction of the different charges have a lower probability of changing house due to the
measure. Furthermore, a positive relation is found between the extent to which people
indicated to adapt their (short term) trip behaviour (e.g. route, departure time, mode choice
etcetera) and the probability that they are going to relocate due to a pricing measure. Next to
that, the sign of the probability of changing job due to the road pricing and the probability of
changing house is positive. This indicates that people, who have a higher probability of
changing their job due to the pricing measure, are also more willing to move house. Finally,
somewhat remarkably no significant effect of the type of price measure (i.e. type of kilometre
charge) or price level on the relocation probability has been found.
13