A traffic flow simulation program is used to estimate the external time costs that an additional vehicle using a congested city street imposes on other motorists on that street. The traffic flow on two street networks in Toronto is simulated for the morning rush hour and a mid-day period. After simulating the actual traffic load in these two periods, traffic volumes on individual streets were varied one at a time by one hundred vehicles per hour. The incremental delay to other vehicles from the addition of these vehicles is calculated by the program and the number of vehicle hours of delay per additional vehicle mile travelled is determined. Assuming a value of time for all motorists, the incremental external time cost attributed to the added traffic on each street in each direction during each of the two periods can be determined. The stimulated traffic variations show that the marginal external social cost of an added vehicle mile considering time costs alone ranges from zero on some roads to over one dollar per vehicle mile in the heavy direction in the morning rush hour. The average of this external congestion cost in the suburban area for inbound motorists in the morning period, weighted by the volume of traffic, was 38 cents per vehicle mile. This study demonstrates the usefulness of a traffic simulation program for estimating congestion costs, and identifies some problems inherent in previous empirical approaches to this problem.
Read full abstract