Cover Image

Not for Sale



View/Hide Left Panel
Click for next page ( 23


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 22
22 Table 6. Approximate costs and revenues for selected tolling and cordon systems outside the United States (2005). FACILITY Tolling Systems Cordon Systems Austria* Germany* Switzerland* London* Stockholm*** Singapore Operating costs 735 m/a 7620 m/a 735 m/a 7133 m/a 740 m/a 77 m/a Toll collect (incl. capital costs) (720 m/a estimate for permanent 750 m/a BAG**** system) Average charge 70.27/km 70.12/km 70.67/km 77.4/day 72.7/day 702 per trip (40 t truck) (40 t truck) (40 t truck) (now 711.8) Fee income 7770 m/a 72 860 m/a 7800 m/a 7275 m/a 780 m/a 739 m/a Operating costs as a % 9% 16%** 4% 48% 25% 7% of revenues Annualized costs (incl. 12% 23%** 8% 55% 40% 40% capital costs) as a % of revenues * Presentation by Bernhard Oehry, Rapp Trans; data for London facility does not include the Western Extension. ** Including costs of deployment, construction, operation, and development of the infrastructure network. *** Stockholm figures for 2006. **** For enforcement under the Bundesamt fr Gterverkehr (BAG), or Federal Office of Freight Transport. Source: Conference on Road Charging Systems, Technology Choice and Cost Effectiveness, Paris, June 1, 2006 light vehicles, the primary concern is congestion. The impact vehicles are often compared with light vehicles in terms of on congestion will depend on the level of congestion, and this road capacity used by means of passenger car equivalent (PCE) is likely to vary for several reasons. First, there is systemic con- measures. The PCE will vary based on a variety of characteris- gestion, which is associated simply with the number of people tics, including the level of congestion and steepness of road using the road. This is subject to certain patterns of congestion grade, and this may be the best determination for congestion- as well as to random variation. Second, there is bottleneck con- related charges to heavy vehicles. In addition, heavy vehicles gestion, which is associated with capacity constraints on a road, cause substantial damage to road pavement based on the weight either due to physical differences, such as reduced number of per axle and certain other characteristics of the vehicle. This lanes, or operational conditions, such as on-and-off traffic. damage also varies with the ability of the road to withstand Finally, there is incident congestion, associated with accidents, heavy loads. Thus, the efficient charge for a heavy vehicle would weather, or other factors that may interfere with traffic flow also vary with the class of road, weight of the vehicle, and num- (CBO, 2009). When setting prices for congestion, there is a ber of axles (Small, Winston, and Evans, 1989). trade-off between the ability to manage congestion efficiently and the ability of the driver to make decisions based on the effi- 2.3.2 Review of U.S. Experience cient price. Prices set in advance may not accurately reflect conditions at any given time, but prices that vary dynamically VMT fees are not directly used for most road financing in may not allow the driver sufficient advance information to the United States; however, there have been a number of exper- change behavior. Charges based on congestion management iments with distance-based charges, and several states have would generate little or no revenue on low-volume roads, espe- used weight-mile taxes for heavy vehicles. Most of the experi- cially outside of urban areas. ments have been based on using GPS to determine the location Charges for externalities that vary with miles driven could of the vehicle; however, Donath et al. (2009) proposed a sys- also be included in efficient VMT fees. Externalities can be dif- tem that would rely on location data related to cell phone tow- ferent under different circumstances and may be mileage based ers as a promising near-term solution to determining where or based on other characteristics, such as amount of fuel used a vehicle is being operated. This system has not been tested, and (Parry, Walls, and Harrington, 2007). no cost estimates were provided. There is interest in odometer- For heavy vehicles, the cost imposed on the road relates to based or self-reported VMT systems, but Sorenson et al. (2009) the size of the vehicle and to the operating characteristics. reviewed near-term options for implementing VMT fees and Because of the size and possibly slower acceleration, more concluded that odometer-based systems would require major room for braking, and problems with steep road grades, heavy changes to DMV operations and databases.

OCR for page 22
23 Oregon VMT Experiment stantially to allow a system with enough reliability to be used for revenue collection, or some capabilities may have to be The Oregon mileage-charging experiment has generated omitted during the phase in. If the capabilities (e.g., for conges- substantial interest (Rufolo and Kimpel, 2008; Whitty, 2007; tion pricing) are left out of the early vehicles, then the imple- and Kim et al., 2008). In the experiment, participating vehicles mentation for congestion pricing could be further delayed. were charged a mileage fee and received a refund of the state The system relied on radio frequency (RF) communication gas tax when they fueled at participating stations. Some vehi- between the vehicle and the fuel pump. For fueling transac- cles were charged a flat fee for all miles driven in Oregon, while tions, the signal strength was required to reach a pre-specified other vehicles were charged a premium rate for driving in the level before the vehicle was clearly identified as fueling at a Portland metropolitan area during weekday rush hours and specific pump. This appears to have resulted in a substantial were charged a discounted rate for all other driving in Oregon. number of transactions that were not identified as being for Vehicles were not charged for driving outside of the state and participating vehicles. Spacing of the pumps, the level of RF did not receive rebates for gas taxes paid to other states. interference, and other factors may have affected the reliability There are several distinct advantages to this system as a of the system, and failed connections created some problems mechanism for collecting revenue. Since the fuel tax is the for the system. If not identified as a participant, the vehicle was default, the majority of revenue for the system is collected from charged the state gas tax and not the mileage fee. At the next the distributors who pay the fuel tax. This substantially reduces transaction, the mileage fee from the last identified connec- the potential for evasion and the need for enforcement mech- tion would be charged, but the refund would only include the anisms. People who do not pay the mileage fee default to pay- gas tax on the current purchase. The owner had to submit a ing the gas tax. In addition, the state has limited need to audit receipt showing the gas tax paid in the interim fueling to get the or monitor individual motorists or vehicles. It should be rela- appropriate refund of the state gas tax. Greater reliability is tively simple for a computer system to compare gas tax refunds needed for an operational system, and this is likely to increase with miles driven to flag vehicles with anomalies for audit. the actual deployment cost relative to the cost incurred in In general, the state would regularly have to audit the service the experiment. stations only with respect to the net difference between the Miles driven with no GPS signal were not charged. The mileage fee collected and the gas tax included in the whole- GPS was left on at all times to minimize the number of miles sale purchase of fuel. The system shows promise as a method driven that could not be allocated, but this resulted in battery to transition from the fuel tax to a mileage-based fee, and it problems for a large number of vehicles. could support congestion pricing at some point. Despite the Behavioral responses may not all be positive. Even with a positive aspects of the experiment, there appear to be both flat fee per mile that approximated the gas tax, some people technological and non-technological issues that deserve fur- reported reducing driving simply because they became more ther consideration. aware of short trips and cost. There was some evidence that the While the system is compatible with congestion pricing, flat fee induced people to group trips. This reduced the total congestion pricing is only feasible if the majority of vehicles are number of trips, but since drivers appeared to group these trips participating. Yet the system is projected to be installed only on with their rush-hour trip to or from work, it may have new vehicles. Since the phase-in period is expected to be fairly increased rush-hour travel. If the grouping resulted in more long, this does not seem to be a reasonable short-term system travel on uncongested local streets, it would not be a problem; for using pricing to address congestion problems. Also, the however, if the travel were on congested arterials or other system does not distinguish factors that affect the impact that roads, the flat fee pricing may have a negative effect on conges- a vehicle has on the level of congestion (e.g., class of road or tion. This could be exacerbated if it increased the amount of direction of travel), although it does charge for each mile stopping and starting (e.g., through more on-street parking) driven in the defined area. and further disrupted traffic. The effect of a flat mileage fee on The technological improvements required relate to the cost rush-hour travel should receive further analysis. and reliability of the system. In general there is going to be a Finally, major oil companies did not agree to allow their gas trade-off between cost and reliability. For the system tested, stations to participate in the Oregon experiment. Since they estimates of the mileage by zone were compared with the represent the majority of stations, the reasons for their refusal odometer mileage for some vehicles, and the differences were should be clarified and addressed. as high as 20% (Kim et al., 2008). In addition, the geographic refinement of the zones was limited. For a revenue collection Puget Sound Regional Council system, users must be convinced that the system is fair, and discrepancies in the determination of location or mileage may The Puget Sound Regional Council sponsored a project to create problems. Hence, costs may have to come down sub- equip vehicles with a device to track all road usage in the area

OCR for page 22
24 and set prices based on the class of road and time of travel. tion from the odometer will be used to generate charges, and Detailed information on all travel by a vehicle was collected the comparison between the GPS mileage and the odometer and uploaded regularly by cellular transmission. The system mileage will be used to monitor the system. Billing data will be is reported to have worked well, but complete details have not uploaded once per month to a central billing operation. The yet been released. billing center will only receive information on the total bill and A number of issues were identified that need to be addressed the apportionment among jurisdictions. During the upload, before the system can be implemented (Puget Sound Regional updated fee files could also be sent to the vehicle. The fee struc- Council, 2008). These include further refinement of the system ture could be specific to vehicle classes, with characteristics and design of enforcement and billing systems. In particular, such as fuel efficiency and emissions affecting the rate charged. there was no enforcement mechanism in the design of the experiment, and an enforcement system would have an addi- Pay-as-You-Drive Experiments tional cost. Dense road networks without access controls were identified as a concern for the pricing system. The trade-off Several of the experiments in the FHWA's Value Pricing between having data processed on the vehicle with summary Pilot program were designed to convert some of the fixed costs data sent to the billing center versus uploading of all data for of driving to variable costs. These used different types of tech- processing at a central location was identified as having privacy nology and helped to identify some of the potential methods to implications as well as communication cost implications. The collect a mileage-based user fee (MBUF) as well as some of the detailed information on travel collected did not appear to be potential drawbacks of these approaches. One such study con- a concern for the participants, but it would almost certainly ducted in Minnesota used a commercial system that plugs be a concern if participation were not completely voluntary. into the OBDII (an onboard diagnostic system) port to obtain Implementation of a full-scale system is projected to have mileage data and tracked total mileage and time for each trip. a mechanism for non-participants that could also maintain This allowed for differences by time of day in pricing but would anonymous usage. Also, the area subject to the pricing seems not allow for differences based on location of travel or class of to have been limited by the storage capacity of the system. The road. In addition, data had to be manually downloaded, and cost of the initial installation of the GPS and communication the system is not compatible with all vehicles (Abou-Zeid et al., costs were identified as key concerns, but declining costs over 2008). A separate account must be set up for each participant time for each were also noted. and there would have to be a billing or payment system. Georgia tested a system similar to the one tested in Iowa, but without the encryption, as a method to charge flat VMT fees. Iowa Pooled Fund Study and Extension An extension appears to be having delays due to instability with The Iowa Pooled Fund study designed a GPS-based system the hardware and software. that could track miles driven by area and could include a vari- ety of areas with varying degrees of overlap and separate pricing Lessons from Experiments systems. All data are maintained in a secure environment, with only total amounts owed by the vehicle to each jurisdiction gen- The costs associated with the experimental approaches erally available. The data were uploaded regularly using a smart- appear relatively high, and operational improvements would card system. If there were a dispute regarding amounts owed, be required for implementation of any of the systems on a large the vehicle owner could decrypt the data to show detailed travel scale. Nevertheless, they show that these types of systems are information (Forkenbrock and Kuhl, 2002). feasible. Since the experiments typically used prototype equip- A system similar to this is undergoing extensive testing over ment on a small scale, the expectation is that per-vehicle cost a number of years and seven locations. It will be some time for an operational system would be lower. This would be par- before the conclusions from this extended study will be avail- ticularly true if the technology is already in the vehicle for other able. However, the basic design is likely to follow that described reasons. by Forkenbrock (2008). It is likely to be somewhat different from the initial experiment. The onboard computer will have Oregon Weight-Mile the capability to store polygons so that charges can be varied by geographic area but not by road. It is expected that there will Oregon charges heavy vehicles for mileage based on the be differentiation by state but that local governments could declared weight of the vehicle and the number of axles. The also add charges for travel within their jurisdictions. The com- charge is intended to equitably allocate the cost of road puter may have the capability to use more detailed files to iden- damage to heavy vehicles since the amount of road damage tify class of road so that differential prices could be charged for increases with vehicle weight but decreases with additional different roads. For periods without a GPS signal, the informa- axles for a given weight. The system is based on monthly or