The Value of Time of Car Drivers Choosing Route: Evidence from the Stockholm Congestion Charging Trial



The Value of Time of Car Drivers Choosing Route: Evidence from the Stockholm Congestion Charging Trial

Authors

M Börjesson, J Eliasson, WSP, SE

Description

Using data from the Stockholm congestion charging trial, we estimate the charging cost sensitivity compared to the fuel cost sensitivity, and the actual value of time car drivers reveal when choosing route.

Abstract

In applied travel demand forecasting (including network modeling), congestion charges are generally assumed to affect travelers in the same way as fuel costs and other travel costs do. However, there are several reasons to believe that congestion charges affect drivers differently than for example fuel costs: for exampel, it is more ?visible?, since you (as a rule) pay each time you pass a cordon (or equivalent), and it is often connected with a certain amount of hassle to actually pay the charge (calling a customer service, for example). Moreover, the way car drivers actually weigh travel time versus monetary cost when they choose route is largely unknown, since time and monetary cost are in general so correlated (as opposed to values of time in mode choice etc.).

Data from the Stockholm congestion charging trial allow us to explore this question in detail. Car drivers traveling between the northern part of the county and the southern part could choose either to go through the cordon (paying two charges ? 20-40 SEK depending on the time of day) or use the Essinge bypass, which was free of charge ? a considerable detour for some trip relations. We use a large panel travel survey conducted before and during the charging trial to estimate the charging cost sensitivity compared to fuel cost sensitivity, and also estimate the actual value of time car drivers reveal when choosing route. Note that due to e.g. selection bias or travelers? imperfect information about costs, this value of tme may differ from the average value of time for all travelers and also from the average value time for car drivers.

The results is very applicable, in that it is easily incorporated into widely used traffic models (network equilibrium models or equivalent), allowing us to make more precise forecasts of the effects of congestion charges. It should also allow us to make more precise route choice modeling in other cases where routes have clearly different monetary costs.

Publisher

Association for European Transport