Measuring the Marginal Cost of Congestion
Mark Burris and Sunil Patil, Texas A&M University, December 2008, 70 pp. (473700-00088-1)
This study attempted to estimate the effect of additional vehicles joining the traffic stream when it is near capacity. The study used data from highways I-35, I-45 in Texas and I-80 in California aggregated at different time intervals. Various macroscopic traffic flow models, Catastrophe model and the bottleneck model were studied in order to identify models that best represent the speed-flow relation for the traffic data over complete range of flow.
The Catastrophe model and the bottleneck model did not fit the data well, while three macroscopic models, the Modified HCM, Newell-Franklin and Van Aerde model were found to fit the data well. Using congestion pricing theory the optimum toll rates were calculated for each of these three models. The optimum toll rates were then compared with the toll rates on some of the existing variably priced toll roads the US. The optimum toll rates estimated using Modified HCM, Newell-Franklin and Van Aerde model were $0.65, $0.81 and $0.97 per mile assuming the value of travel time savings as $20/hr for the near capacity flow. These toll rates were found to be lower when compared with the maximum toll rates on three of the existing variable toll facilities which charge about $1/mile during the hours of extreme congestion.
Keywords: Marginal Cost of Congestion, Congestion Pricing, Traffic Flow Models, Catastrophe Theory
ENTIRE REPORT (Adobe Acrobat File – 981 KB)