Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
72 8. ANALYSIS OF MOST PROMISING VMT-FEE OPTIONS This chapter provides further analysis of the three VMT-fee options judged most promising for near-term implementation â mileage estimates based on fuel consumption, OBD II / cellular in- vehicle equipment, and coarse-resolution GPS in-vehicle equipment. For each option, the following sections present more detailed consideration of metering capabilities, costs to implement and operate (note, however, that rigorous cost estimates are beyond the scope of this project), administrative options, enforcement options, parallel charging mechanisms for certain classes of domestic vehicles, parallel charging mechanisms for foreign vehicles, legal barriers, user burdens, privacy concerns, ability to audit charges, near-term flexibility, and support for the transition to a more robust and sustainable system of VMT fees over the longer term (note that with the exception of cost and user-acceptance criteria, the OBD II / cellular and GPS options are quite similar). The chapter summarizes relevant strengths, weaknesses, uncertainties, and key design decisions for each option and notes obstacles that all of the options will face. While any of the three options could be applied to trucks as well as cars, it may be desirable to instead develop a more sophisticated approach based on high-resolution GPS to implement weight-distance truck tolls capable of metering road use by specific route. Discussion at the end of the chapter comments briefly on the strengths and limitations of this concept. 8.1. Fuel-Consumption-Based Mileage Estimates Core metering capabilities. With this option, per-mile fees could be levied on the basis of estimated mileage as a function of (a) the expected fuel economy for the vehicleâs make and model and (b) the amount of fuel purchased at the pump. If desired from a policy perspective, the per-mile fee could also vary according to such vehicle characteristics as weight, greenhouse gas emissions, or criteria pollutant emissions. Under the core configuration, it would not be possible to achieve policy goals or pricing variations keyed to the time or location of travel. This rules out accurate apportionment of mileage revenue among jurisdictions, accurate assessment of road damages caused by heavy trucks on the basis of axle-weight and specific route of travel, and any forms of congestion pricing (although, as described below, extensions to the core configuration would enable some of these options). Cost. There are two main up front costs required to implement this strategy. The first is to equip all stations with the necessary electronic gear â including an AVI reader for each pump (located in such a manner that it is possible to clearly distinguish between the vehicle located at one pump and the vehicle located at another), a service station point-of-sale system for calculating and tracking VMT fees and fuel tax rebates, and, optionally, an internet-based connection to a central data repository. Whitty and Svadlenak (2009) have estimated that this equipment would cost approximately $15,000 per station. Our research indicates that there are approximately 160,000 retail fueling stations across the United States, resulting in a total cost of roughly $2.4 billion. The other main capital cost would be for equipping all vehicles with the necessary AVI device, likely involving an RFID tag. This could be affixed to the windshield, embedded in a license plate, or even embedded in an annual registration or emissions sticker. Whitty and Svadlenak
73 (2009) have estimated an upper bound price of $50 per unit to manufacture and install a reasonably tamper-resistant AVI on existing vehicles; with approximately 250 million vehicles in the United States, this would entail a total cost of $12.5 billion. This should be viewed, though, as a worst case cost scenario. RFID tags are themselves cheap, on the order of a few cents per tag. Depending on the specific deployment strategy (e.g., embedding RFID tags within annual registration stickers), the cost could be considerably lower, perhaps negligible. The cost of administration must also be considered. Because this option would involve the expansion of an existing revenue system rather than the development of an entirely new system, administrative costs should (in theory) be lower. The key modification to the fuel tax collection system (described in more detail below) will be to provide a mechanism for resolving the difference between the amount of fuel taxes paid by the retail station, which is already included in the cost of fuel purchased from distributors, and the amount of VMT fees collected by the retail station. Depending on whether the balance is positive or negative, fuel station owners would either receive a credit or pay additional revenue to the fuel tax collection agency. In their exploration of this concept for the state of Oregon, Whitty and Svadlenak estimated that this additional administrative effort would cost about $1 million per year, roughly equal to what the state currently pays to collect fuel tax revenue. In other words, this approach would likely double the current administrative costs associated with fuel taxes. Given that fuel tax administration is quite efficient, with costs typically less than one percent of revenues, this should be viewed as a key advantage of this proposed approach. Note, however, that some participants in the projectâs expert panel expressed the opinion that administrative costs could be much higher. This assessment is therefore judged as uncertain and merits further analysis. Administration. Since 1994, the IRS has collected federal fuel excise taxes at the terminal rack, the point at which fuel is transferred from barges, ships, or pipelines to tanker trucks. The âposition holderâ â the entity that owns the fuel as it passes through the rack â is liable for the tax. As of 2003, there were about 1,400 terminal racks in the United States, and each must register with the IRS. State revenue agencies vary in their practices for collecting excise taxes; some, like the IRS, collect at the terminal rack, others collect from the distributor who purchases fuel at the rack, and still others collect at the point when the fuel first enters the state. Diesel fuel taxes, which apply to on-road vehicles but not off-road vehicles (e.g., farm equipment), may be collected by states at subsequent stages in the distribution chain. Under the proposal for levying VMT fees at the fuel pump, the IRS and comparable revenue agencies at the state level would need to interact directly with retail fuel stations to reconcile the difference between existing fuel taxes added to the price of fuel earlier in the distribution chain and VMT fees collected at the pump. If VMT fees collected exceed the amount of rebated fuel taxes, then the station would need to remit the difference to the IRS or the state collection agency. If, on the other hand, the amount of fuel taxes rebated to the customer exceeds the amount of VMT fees collected, then the fuel station would need to be reimbursed by the revenue agency (though given that the intent of VMT fees is to raise revenue, this outcome is less likely). A key point here is that the number of parties with which the revenue agency must interact could increase considerably, from a relatively small number of wholesalers and distributors to a much larger number of retailers. The IRS and state revenue agencies routinely interact with millions of
74 income tax payers each year, so the additional complexity of interacting directly with retail stations would seem to be well within their capabilities. However, one reason from moving fuel tax collection to the rack was to increase compliance, and some workshop participants expressed concern that moving the point of collection further âdownstreamâ may increase the opportunities for tax evasion. It is possible that both the IRS and states will oppose this scenario on the grounds that it both complicates their existing collection mechanisms and opens the door to increased evasion. An open question regarding administration is whether both the IRS and state revenue agencies would need to interact with retailers. One could envision, for example, the federal government collecting VMT fees from retailers and then reimbursing any states that choose to opt in to the system by levying their own VMT fees. In either case, the interaction with tax authorities may constitute an additional burden to fuel station owners, who do not currently participate directly in the collection of fuel taxes. Enforcement. This approach performs especially well with regard to enforcement. If a vehicle is equipped with the required AVI device when refueling, mileage fees will be substituted for fuel taxes. If the vehicle does not have the AVI equipment, or if the AVI device is not working for some reason (including deliberate tampering), the vehicle will instead pay fuel taxes. In other words, there is no way for a vehicle to avoid paying road use charges, either through VMT fees or through fuel taxes, when refueling. One could envision that this approach might be phased in over some period of time, such as with vehicle purchases or ownership transfers or with the routine re-issuance of license plate tags. During the transition period, vehicles not yet equipped with the device would continue to pay fuel taxes as before. When the transition is complete, the AVI device would be required for all (domestic) vehicles. At this point, the fuel tax might be raised considerably to discourage efforts to tamper with the AVI device, as this would result in a higher charge than simply paying the mileage fees. Parallel charging mechanisms for domestic vehicles. The proposed approach would work well for conventionally-fueled vehicles, which currently account for more than 99 percent of the on- road vehicle fleet (Whitty and Svadlenak, 2009). It would not, however, account for vehicles that do not need to purchase fuel at a gas station. While recent developments suggest that electric vehicles may be the first in this category to reach the stage of mass marketing, future years may witness the introduction of natural gas- and hydrogen-powered vehicles with home-fueling options as well. Pluggable hybrids present another challenge. Though such vehicles will still purchase fuel as needed, their fuel economy â the basis for estimating mileage charges â will vary highly depending on the frequency with which they recharge their batteries. In short, the pay-at-the-pump model will not handle all vehicles currently on the road, and the percentage of vehicles not covered under this model will likely increase in the coming years. This makes it necessary to consider strategies for levying VMT fees for non-conventionally fueled vehicles (likely including pluggable hybrids). Several options may be considered:
75 ⢠No charges levied. Policymakers may wish to create incentives to encourage more rapid adoption of alternative-fuel vehicles. Allowing owners to forego road use charges might be viewed as such an incentive. From a revenue perspective, this would not be problematic until alternative fuel vehicles begin to achieve significant market share (say, greater than two percent of the fleet, although EIA (2009) forecasts that this will occur by 2011). There are, however, equity implications to consider. In the early years, alternative-fuel vehicles, given their more limited production, will tend to be more expensive than conventional vehicles. As a result, early adopters are more likely to those with higher incomes. Allowing such owners to forego road use fees will therefore result in a marginal shift of the burden of the tax burden from upper- to middle- and lower income groups. ⢠Fixed VMT fee. Another option would be to simply add a fixed âmileage feeâ to the annual registration fee for alternative-fuel vehicles. Assume, for instance, that the base per-mile fee for conventional vehicles is set to 1.1 cents per mile (a roughly revenue-neutral replacement for current federal fuel taxes), and that the average car is driven 15,000 miles per year. This would result in an annual mileage surcharge, paid with registration, of $165 for alternative- fuel vehicles. Though not sensitive to actual miles of travel, the advantage of this approach would be the relative ease of administration â simply adding to the annual registration fee for a small subset of all vehicles. ⢠Fixed VMT fee with odometer-based appeal. This would be similar to the last option, with the exception that vehicle owners that travel significantly less than the assumed mileage would have the option of submitting to annual odometer readings to qualify for a reduction in total mileage fees. Though this would likely be viewed as a more equitable arrangement, it would lead to a considerable increase in administrative complexity. States that do not currently conduct vehicle inspections would need to develop the infrastructure to perform odometer readings, and many states would also need to augment their vehicle registration databases to track annual odometer readings. ⢠Odometer-based VMT fees. With this option, all alternative vehicles would submit to annual odometer readings, resulting in a more accurate assessment of VMT fees for this class of vehicles. Like the previous option, this would create additional administrative complexity for states in terms of the need to collect, track, and bill for odometer readings. ⢠Wireless communications with central billing agency. One of the assumptions underlying this project is that the near-term options for implementing VMT fees considered in the study may evolve to (or be replaced by) a more robust system that offers greater pricing flexibility (e.g., the ability to meter miles by time and location) over the longer term. The longer term approach may involve in-vehicle equipment that communicates wirelessly (e.g., via cellular communications) with a central billing agency. The idea with this option would be to develop the option for wireless communications to a central billing agency in the near term and apply the option solely to alternative-fuel vehicles (all such vehicles might be required to adopt this approach, or alternatively it could be provided as a more convenient option for vehicle owners not wishing to submit to annual odometer readings). Because the percentage share of such vehicles will remain low for some years, this could present the opportunity to develop and deploy the central billing system in a low-volume and low risk (in terms of total transactions and share of total revenue, respectively) environment. This will provide
76 operational experience that could guide the scaling and refinement of the central billing option, should the decision ultimately be made to apply this mechanism for all vehicles. Parallel charging mechanisms for foreign vehicles. Another important advantage of the pay- at-the-pump model is that it would still be possible to charge foreign vehicles for their use of the road. Unless Canada and/or Mexico also adopted a similar charging scheme, foreign vehicles would lack the necessary AVI equipment to estimate VMT fees on the basis of fuel economy. In this case, however, the drivers would simply pay applicable fuel taxes instead of mileage fees. Legal barriers. This approach would certainly require enabling federal, and potentially state legislation. One of the project panel members noted, for instance, that his state would require legislation to affix an AVI device for the purpose of federal road use fee collection to state- issued equipment such as a license plate or registration sticker. Beyond that, the authors are not aware of specific legal barriers that would prevent this form of charging. User burden. The user burden under this option would be low. Depending on the configuration of the AVI device, it might be necessary for owners to visit a station where the device would be installed, though even that would not be necessary if the choice were made to embed an RFID tag in replacement license plates or annual registration stickers. Once the device is installed, there would be no further burden for the owners of conventionally-fueled vehicles. They would simply pay mileage charges at the pump, similar to fuel taxes. Depending on the strategy used to levy mileage charges for alternative-fuel vehicles, there may be a modest burden on the owners of such vehicles (e.g., the need to have the odometer inspected on an annual basis). In the near term, though, such vehicles will likely remain a small percentage of the total fleet. Ability to audit. Auditing VMT fees should be straightforward with this option. When purchasing fuel, the receipt that it printed out could be designed to include (a) vehicle identification information, (b) rated fuel-economy for make and model, (c) corresponding mileage fee per gallon of gas, (d) gallons purchased, and (e) total mileage charge. Privacy concerns. Equipping all vehicles with some form of AVI might lead to privacy concerns. Yet the level of concern should be much less than with a system featuring GPS, since the fuel-consumption-based model will not have the capability to track either time or location of travel. Should the system be extended to include gantries that would monitor the passage of vehicles on specific routes or within specific areas to facilitate certain forms of congestion tolling (see below), privacy concerns would likely increase. Near-term flexibility. The core technical configuration under this option permits, as noted, VMT fees based on estimated mileage that may vary with such vehicle characteristics as weight and emissions class â but not with the time or location of travel. The fact that all vehicles would be equipped with AVI devices, however, creates several additional pricing options that could be pursued if desired. By adding gantries equipped with DSRC readers along heavily traveled routes, for instance, it would be possible to levy facility-based congestion tolls, as has been done with current HOT lane implementations in the United States. By placing gantries at all entrances to a crowded urban core, it would likewise be possible to implement a cordon congestion charge similar, from a technology perspective, to the Singapore program.
77 Several caveats should be mentioned. First, the installation of DSRC-equipped gantries is not inexpensive (in addition to the RFID reader, the gantry must also include a sensor to detect that a vehicle is approaching; otherwise it would not be possible to flag vehicles that lack the needed AVI device). On the other hand, congestion tolls would create a considerable revenue stream, so it should not prove difficult to recoup the infrastructure cost in a short period. Second, adding this form of road pricing would require a parallel payment mechanism. With only a simple AVI device (e.g., an RFID tag), the in-vehicle equipment would not be capable of storing congestion tolls accrued throughout the road network that could subsequently be paid at the pump. Rather, the readers attached to the gantries would need to communicate the charges owed by each passing vehicle to a central data repository such that the motorist could be subsequently billed. Third, the cost of the required AVI technology would likely increase. Standard RFID tags can be reliably read within a distance of 5 to 10 feet â sufficient for a stationary vehicle parked next to a fuel pump but not necessarily adequate for a vehicle passing under a gantry at 60 miles per hour. To increase the range, the strength of the RFID signal can be boosted through the inclusion of a small battery. This would obviously increase the cost of the AVI technology, though the added cost should not be major. Support for longer-term transition. This option should support, rather than hinder, any efforts to transition to a more robust system of VMT fees over the longer term. First, it would introduce motorists to the concept of mileage-based user fees, but in a non-threatening way. The payment mechanism would remain the same, and the privacy concerns associated with tracking the location of travel, e.g., via GPS, would not be relevant. Second, the approach leverages the existing fuel tax collection system. This should keep current administrative costs low â an important advantage â but it also obviates the need to make considerable investments in a new collection system before the optimal configuration of the longer-term system has been determined. Third, even when a robust system featuring more advanced in-vehicle technology has been implemented over the longer term, it will still be necessary to assess road use fees for foreign cars. The electronic equipment installed at gas stations, used in the near term to read vehicle identification information, can later be used to distinguish between domestic vehicles equipped with metering technology and foreign vehicles lacking the necessary metering equipment. The latter group would pay fuel taxes, while the former would pay mileage charges. Finally, though other payment options might be provided in the longer term system, many motorists will continue to purchase fuel at filling stations, and some may prefer the option of continuing to pay their mileage charges at the pump. Summary. The strengths, limitations, uncertainties, and key planning decisions for metering based on fuel-consumption can be summarized as follows: Strengths ⢠Likely low in-vehicle equipment costs for AVI ⢠Potentially low administrative costs ⢠Effective and low-cost enforcement ⢠Addition of gantries enables additional forms of pricing
78 ⢠Easily accommodates foreign users ⢠Low user burden ⢠Straightforward audit trail ⢠Minimal privacy concerns Weaknesses ⢠Core pricing flexibility is limited ⢠Inaccurate mileage estimates ⢠Not possible to accurately apportion revenue by jurisdiction ⢠High cost to equip fuel stations with electronic readers ⢠Share of vehicles requiring parallel payment mechanism increases over time Uncertainties ⢠Cost and configuration of AVI device ⢠Cost of equipping fuel stations with electronic readers ⢠Cost of modifying fuel tax collection system to interact with retail fuel stations ⢠Compliance with tax remissions Planning and Design Issues ⢠State vs. federal roles in revenue collection ⢠Payment mechanism for alternative-fuel vehicles ⢠Whether to augment metering capabilities with DSRC tolling 8.2. OBD II Port Connection Combined with Cellular Communications Core metering capabilities. By integrating vehicle speed information collected through the OBD II port over time, this approach should allow for more accurate mileage estimates than the fuel-consumption-based model. The on-board unit can also store information about the vehicleâs make and model, making it possible to vary the per-mile rate according to such factors as weight and emissions class. Additionally, the inclusion of cellular communications makes it possible to determine the area or zone of travel. This permits relatively accurate apportionment of mileage revenue among different jurisdictions and also facilitates area-based forms of pricing (e.g., paying different mileage fees in different jurisdictions or paying higher per-mile fees when driving in a particular metropolitan area during peak hours). In short, this option offers a high degree of pricing flexibility. The main capability not provided is metering travel by specific route of travel, which would be most important for weight-distance truck tolls.
79 Cost. There are several important cost components for this option. The first is the cost of the in- vehicle equipment. On-board devices featuring an OBD II port connection along with cellular communications are available on the market today (they are used in some pay-as-you-drive insurance programs, as well as in other applications). Inquiries with equipment manufacturers and insurance firms suggest that these devices currently cost about $50 per unit. It should be noted, however, that the current market penetration can be measured in the tens of thousands. Expanded to a user base of hundreds of millions of vehicles, the per-unit cost would surely decrease. Assuming a worst case scenario of $50 per unit, however, would lead to a total cost of about $12.5 billion to equip 250 million cars. Additionally, it would likely be necessary to have the units installed at a certified shop, another important start up cost (although the installation of a device connected to the OBD II port is much simpler than installing GPS equipment, discussed later). To make use of the cellular technology to determine the location of travel, it would likely be necessary to set up a cellular service agreement with one or more service providers, and the associated cost would be required on an ongoing basis. Depending on the structure of the service contract, the transmission of data (e.g., to a central billing agency) could entail additional costs. In preliminary analysis of the cost of developing and operating a central VMT fee billing agency for the state of Oregon, Bertini et al. (2002) estimated the cost of air time for data transmission at $.05 per minute. Allowing for inflation, and assuming that mileage data would be transmitted on a monthly basis and that each transmission could be accomplished within a minute, the cost would be roughly $1 per vehicle per year, or roughly $250 million annually. It should be noted, of course, that the cost of wireless transmission has decreased in the intervening years, and sending text data is even cheaper than voice communications, so the cost may well be substantially less. Given the ongoing nature of this charge, further research to update this cost estimate would be merited. A final cost to consider involves setting up and operating a central billing agency with which the on-board equipment would communicate. Note that the scale of central billing operations would depend on the selected strategy for collecting mileage data and assessing fees. One possibility would be to set up electronic readers at fuel stations, as with the previous fuel-consumption- based metering option, such that conventionally-fueled vehicles could pay mileage fees at the pump. Only alternative-fuel vehicles would then need to communicate wirelessly with a central billing agency, which could be scaled accordingly. As another option, the system could be set up such that all vehicles would communicate mileage data through the same central billing office, and this would increase the cost. Prior work by Bertini et al. (2002) provides insight into the potential magnitude of the costs of setting up a central billing agency to collect mileage data wirelessly and issue bills. As part of the initial Oregon VMT fee study, Bertini and his colleagues developed estimates for the initial capital and ongoing operational costs for a central billing agency to support Oregonâs three million drivers. For the scenario under which vehicles would transmit mileage data on a monthly basis, initial capital costs were estimated at $1 million, while operational costs were estimated at about $50 million annually. To put these numbers in context, the state of Oregon currently spends about $1 million each year collecting fuel taxes (Whitty and Svadlenak, 2009), so this would represent a substantial increase in administrative costs. It should be noted that a
80 significant share â about $18 million per year â of the estimated operational costs were based on the assumption that paper bills would be generated and sent to motorists on a monthly basis. As electronic billing systems become more widespread, this cost component would likely decrease considerably. Additionally, there would likely be some economies of scale when developing a central billing system to accommodate all motorists across the U.S. Even so, it is apparent that both the capital and operational costs of such a system would be high. Administration. The previous approach of metering mileage based on fuel consumption dovetails well with the existing system for administering fuel taxes. With this option, in contrast, there are at least three options worthy of consideration. One possibility, similar to the previous approach, would be to equip fueling stations with electronic readers capable of communicating with in-vehicle metering equipment (in this case, however, the devices would communicate mileage fees directly, as opposed to vehicle information used to estimate mileage). While requiring a higher initial capital investment to equip the stations, this could substantially reduce administrative costs. Again, however, it would be necessary to develop parallel collection mechanisms for alternative-fuel vehicles. The second option, facilitated by the use of cellular communications on the in-vehicle device, would be to establish a central billing agency that communicates with the device via wireless transmissions. As noted above, this would likely lead to higher operational costs. On the positive side, because this could handle all vehicles, it might eventually eliminate the need for parallel collection mechanisms. Additionally, it is anticipated that a greater share of alternative-fuel vehicles will enter the marketplace in the coming years, and these may need to rely on central billing. Should it prove necessary or desirable to pursue the central billing agency, the next question that arises is how to administer the billing center. Several possibilities exist, including federal administration, contracting the service out to a private firm or consortium, or developing some type of non-profit agency or joint-powers authority governed by representatives of the states. Further research to examine the strengths and weaknesses of these options is merited. The third option would be to develop debit cards that could be inserted into the OBU to pay mileage fees. This would obviate the need to report mileage data to a central billing agency, thereby reducing potential privacy concerns that might arise due to the use of cellular to meter the location of travel. Because this would increase the burden on users â who would need to purchase and periodically recharge the debit card â this would not likely be selected as the sole means of payment; rather, it would be an option for those particularly concerned with privacy. Developing this additional option would increase the costs of administration, but further research would be needed to specify the magnitude of the increase. Research would also be needed to determine the best model for producing and vending debit cards and allowing users to recharge them as needed. Enforcement. The enforcement challenge is more difficult with this option than with the approach based on fuel consumption. The key issue here is to ensure that the motorist does not accidentally or deliberately disable the in-vehicle metering device so as to reduce the number of
81 miles recorded. There are several approaches, depending on the adopted mechanism for transmitting mileage data. Under the pay-at-the-pump model, a disabled metering device would lead to a failure when the reader at the station attempts to communicate with the in-vehicle equipment. In this case, the vehicle would instead be required to pay fuel taxes. This would not help, however, with the case in which a device is disabled while driving and then re-enabled prior to purchasing fuel. An option here, described by Whitty and Svadlenak (2009), would be for the fueling station to transmit data to a central computer system on mileage fees as well as the quantity of fuel purchased for each transaction. This data could then be audited to scan for gross inconsistencies between mileage and fuel consumption, which could trigger a subsequent audit. For the approach in which mileage data are wirelessly transmitted to a central billing agency, it becomes necessary to ensure that the on-board unit is highly resistant to tampering (this would be advisable as well for the pay-at-the-pump model). Perhaps the simplest model would be to design the devices such that they could not be disabled without breaking a certification seal affixed to the device at the time of installation. The seals could then be inspected each year to ensure that no tampering had occurred. This, however, would greatly increase administrative costs. Another option would be to randomly inspect a smaller share of vehicles each year, setting the penalty high enough such that a rational actor would choose not to intentionally disable the in-vehicle equipment (similar in concept to the honor system for transit fares). Yet another possibility would be to allow police cars that stop vehicles for other reasons to check the odometer and compare that to the mileage data provided via OBU, with citations issued in the case of a gross mismatch. With all of these cases, it would be desirable to ensure that any driver that accidentally disables the device would be aware of the problem such that they could have it fixed promptly to avoid penalties. For example, the device might be designed to blink with a bright red light if it is not operating properly. As another possibility, the central billing agency could periodically send wireless queries to the on-board equipment; after a sequence of failed queries, the system could then generate an audit request. Alternatively, if the on-board device detects that it has been disabled, it could send a wireless alert to the billing agency which would likewise trigger a compliance audit. Finally, if gantries are set up throughout the network (for instance, to levy facility-specific tolls on top of the base per-mile fee), then the gantries could query the on-board unit on each passing vehicle to verify that it is functioning properly. In short, there are a range of potential enforcement options, and it may prove useful to combine several for the sake of redundancy. Further research to determine the most effective, and cost- effective, approaches would be beneficial. Parallel charging mechanisms for domestic vehicles. The OBD II port did not become standard equipment until 1996. As a result, this metering approach would not work for vehicles produced prior to that year. Fortunately, the share of such vehicles still on the road by 2015 will be small, and it will continue to decline as the years pass. Even so, it is necessary to consider parallel mechanisms for charging such vehicles for road use. There are several possibilities. ⢠No charges levied. As with the prior approach, it would be possible to allow older vehicles to travel without paying for road use. Unlike with alternative-fuel vehicles, it is difficult to
82 make a compelling policy argument for this choice (ease of administration aside). Older vehicles tend to be the least fuel efficient and most polluting, after all, so creating an incentive to keep such vehicles on the road even longer would undermine environmental goals. ⢠Fixed VMT fee. Another option would be to simply add a fixed âmileage feeâ for older vehicles. This would be inexpensive to administer, and concerns over the relative inequity between higher-mileage drivers and lower mileage drivers would diminish as the share of such vehicles on the road declines with time. ⢠Fixed VMT fee with odometer-based appeal. This would provide lower-mileage drivers with the opportunity to reduce their annual charge if they are willing to submit to annual odometer readings. On the negative side, this would also require that states develop the infrastructure to conduct odometer readings and handle mileage appeals. It may be difficult to justify such investment given that the share of pre-1996 vehicles on the road will decline with time. ⢠Odometer-based VMT fees. Under this option, all pre-1996 vehicles would submit to annual odometer readings, resulting in a more accurate assessment of VMT fees. Here again, the required administrative investment on the part of states may be difficult to justify from a cost-recovery perspective. ⢠Fuel taxes. If the pay-at-the-pump collections model is developed, it would be possible to distinguish between newer vehicles with the necessary metering equipment and older vehicles lacking the equipment. In this case, pre-1996 vehicles could simply continue to pay fuel taxes, while newer vehicles with the required metering device would pay mileage fees instead. If, in contrast, the central billing model facilitated by wireless transmissions were developed for all cars, then this option becomes more difficult, though not impossible. Under a concept put forward by Donath et al. (2009), vehicle owners with metering equipment installed in their cars would pay fuel taxes at the pump and then subsequently receive a rebate for fuel taxes paid (to record the amount of fuel taxes paid, a driverâs credit card could be linked to vehicle information; drivers who prefer to pay cash might instead be issued some type of special purpose refueling card that would indicate the vehicle identification). Vehicles lacking in-vehicle metering equipment would simply pay fuel taxes that would not be rebated. The advantage of this concept is that it would not be necessary to equip all fueling stations with DSRC readers, although it would still be necessary to upgrade the electronic payment system to link fuel tax payments with vehicle identification. On the other hand, this would lead to additional administrative costs to set up and operate the necessary bookkeeping and fuel tax rebate system, and it would also increase, to some extent, the burden on users. ⢠GPS-based metering equipment. A final option would be to install more advanced devices equipped with GPS in older vehicles. Rather than relying on the OBD II port connection to determine mileage, the device would instead rely on successive GPS readings to estimate travel distance. Though technically feasible, this application of GPS may increase the perception of privacy concerns among some owners of older vehicles. Parallel charging mechanisms for foreign vehicles. If the pay-at-the-pump model were adopted, the reading equipment at gas stations would be able to distinguish between vehicles
83 with mileage metering devices and those without. The former would have mileage fees added to the fuel purchase and fuel taxes subtracted, while the latter â including foreign vehicles â would simply pay fuel taxes. If, instead, the central billing agency option were adopted as the sole collection mechanism such that gas stations did not need to be equipped with electronic readers, and if the fuel-tax rebate system described above were not developed, then fuel taxes would presumably be phased out once the system is in place. This would make the proposition of charging foreign vehicles much more complicated. The following options, all of which face certain weaknesses, might be considered: ⢠Flat fee. With this approach, a flat fee would be assessed for all vehicles entering the country. It could be viewed as grossly inequitable in that it would not distinguish between, for example, a vehicle that crosses the border for a short day trip and one that enters for an extended several week trip. It would also increase the cost and complexity of administering border check points. ⢠Flat fee with odometer-based appeal. This would enable those that cross into the country for just a short trip to qualify for a reduced fee by checking the odometer upon entering and existing the country. This would further add to the administrative cost and complexity of operating border checkpoints. ⢠Odometer-based fees. This approach would require that all foreign vehicles have their odometers checked on entering and exiting the country and pay the corresponding road use charge. This would entail even greater expense for operating border checkpoints. ⢠On-vehicle equipment for frequent visitors. With this approach, frequent foreign visitors, such as trucks, would install the same metering equipment as domestic vehicles, thus automating the collection of road use fees and reducing administrative costs at the borders. Infrequent visitors, however, might still need to pay road use fees manually. Legal barriers. This approach would also require enabling federal, and potentially state, legislation. Beyond that, the authors are not aware of additional legal barriers that would prevent this form of charging. User burden. For this approach, the user burden depends on several technical and programmatic design issues. When the program is first initiated, users would likely need to visit a certified shop to have the in-vehicle equipment installed. Under the pay-at-the-pump model, motorists would then pay mileage fees with fuel purchases, requiring no additional burden. Under the central billing option, motorists would need to pay fees on a periodic basis, likely monthly. This would add some burden, though the process could be automated through electronic payment systems for motorists who desire this option. A final issue â and this could be significant â depends on the approach for verifying that the in-vehicle equipment has not been tampered with. If the adopted approach relies on annual inspections, the additional user burden (along with administrative cost and complexity) would be considerable. If it proves possible to develop effective enforcement strategies not reliant on periodic equipment inspections, the user burden would be much reduced.
84 Privacy concerns. This approach would allow mileage metering by time of travel and by area or jurisdiction. The privacy concerns are thus higher with this option than with estimating miles based on fuel consumption. The level of concern, however, may be less than with a GPS-based system. In part due to inaccurate or inflammatory press coverage, the mere mention of GPS often conjures imagery of âbig brother watching,â and this perception may be sufficiently strong as to render a GPS-based approach politically infeasible. Deriving the location of travel based on cellular communications provides accuracy on par with lower-resolution GPS receivers, but it may not create the same stigma in the public eye. Ability to audit. With any OBU that meters location information, there is a desire to both protect privacy and provide enough data to enable the motorist to verify that charges have been correctly applied (Whitty and Svadlenak, 2009). While it was noted at the workshop that a system could be designed with high degrees of both privacy protection and auditability, the trade-off is that the system would become much more expensive. This is an issue that could benefit from further research attention. Near-term flexibility. A key advantage of this approach is that it offers very flexible pricing options. The in-vehicle equipment enables accurate metering of mileage by jurisdiction or area, making it possible to vary charges by vehicle characteristics, by jurisdiction, or by time and area of travel (e.g., area-based congestion tolls). By adding gantries at strategic points throughout the road network, it would also be possible to levy facility-based or cordon congestion tolls. Each time a vehicle passed a gantry, the in-vehicle equipment would store the relevant charge, and that would be added to base mileage when computing the fees owed. As noted earlier, gantries could also query to verify than on-board units are functioning properly, thus supporting the enforcement function. Support for longer-term transition. Under the assumption that gantries are added to enable facility-based or cordon congestion tolls, this configuration allows for almost any conceivable pricing option. The sole exception would be fees that vary by specific route across the entire road network. This latter capability, however, would be mainly relevant for weight-distance truck tolls intended to capture the greater damage imposed by heavily laden trucks traveling on lightly- engineered surface streets. For all other vehicles, this approach could be implemented in the near term and could also support a robust and sustainable system of VMT fees over the longer term. The longer-term task, then, would not be one of transition, but rather one of evolution and refinement (for example, finding ways to make the on-board device even more tamper-resistant over time). Summary assessment. The strengths, limitations, uncertainties, and key planning decisions for the OBD II / cellular option can be summarized as follows Strengths ⢠Core configuration offers flexible pricing ⢠Addition of gantries enables additional forms of pricing ⢠Allows for accurate apportionment of fees by jurisdiction
85 ⢠Comparatively low in-vehicle installation costs ⢠Percentage of vehicles requiring parallel payment mechanism declines over time ⢠Can support longer-term sustainable revenue system; no transition required Weaknesses ⢠Use of cellular to meter location unproven in road pricing applications ⢠Higher costs for in-vehicle equipment ⢠Potentially higher administrative cost if central billing agency used to collect fees ⢠Difficult to charge foreign users unless pay-at-the-pump collection model employed ⢠Enforcement more difficult, perhaps more costly ⢠Potentially higher user burden depending on enforcement strategies ⢠Greater privacy concerns given cellular metering of location Uncertainties ⢠Practical feasibility of metering location via cellular service ⢠Cost of in-vehicle equipment manufactured at scale ⢠Ongoing cost of cellular service ⢠Costs associated with pay-at-the-pump system, if implemented ⢠Costs associated with central billing agency system ⢠Costs associated with debit card payment option ⢠Cost and effectiveness of alternate enforcement options ⢠Whether cellular location will raise privacy concerns comparable to GPS ⢠How to enable billing audits while protecting privacy Planning and Design Issues ⢠Strategy for procuring technology (see next chapter) ⢠Method(s) for collecting revenue â pay at the pump, central billing, and/or debit cards ⢠Public vs. private administration of central billing agency ⢠Strategies to enforce compliance ⢠Parallel payment system for vehicles manufactured prior to 1996 ⢠Method for charging foreign vehicles, if fuel taxes phased out ⢠Whether to augment metering capabilities with DSRC gantries
86 8.3. Coarse-Resolution GPS Option Core Metering Capabilities. The metering capabilities of this option are identical to that of the OBD II / cellular combination, and can likewise be extended through the addition of DSRC gantries at strategic points throughout the road network. The only capability not possible under this configuration is metering by specific route across the entire road network. Note that because the GPS signal may not be available in some areas (e.g., in tunnels or in urban canyons), it may still be necessary to include a connection to the OBD II port to ensure accurate mileage metering at all times. Cost. Relative to the OBD II / cellular OBU configuration, this option would likely cost more on a per-unit basis due to the need to include a GPS receiver (although depending on the selected method for collecting fees, it may be possible to omit the cellular component). A cursory review of currently available consumer GPS applications (e.g., personal navigation devices) suggests that the per-unit cost should not exceed one or two hundred dollars. Additionally, because GPS requires connection to a power source (likely the vehicleâs battery), the initial installation costs for this option are likely to be higher as well. It is important to stress, however, that there are factors that could result, over the next five years, in much lower cost. First, the per-unit cost should come down considerably if production is scaled to hundreds of millions of units, as would occur with a national system. Second, many vehicle manufacturers are beginning to include GPS as standard equipment on their models. Provided that the OBU could be designed to plug into the vehicleâs GPS signal, this would reduce the cost of the OBU (which would no longer require a separate GPS receiver) as well as the cost of installing the OBU (which would no longer require a battery connection). This technology configuration would present the same payment options as with the OBD II / cellular device, so the cost of collecting revenue would be comparable. Administration. The same administrative options possible with the OBD II / cellular device â pay at the pump, central billing, and debit cards â are also relevant for this configuration. The relative advantages and limitations of these three options also remain the same. Enforcement. This option also presents the same enforcement challenges as the OBD II / cellular device â namely, finding a way to make sure that the motorist is unable to temporarily disable the OBU to avoid applicable mileage fees. The same core options, including use of the odometer as a redundancy check, tamper-resistant OBU design, and external verification that the OBU is functioning properly, are possible. Parallel Charging Mechanisms â Domestic Vehicles. One key advantage of this option is the technology can be configured to work with any vehicle. Thus it is not strictly necessary to create a parallel charging mechanism (although one might choose to do so to allow a technology phase- in period). Parallel Charging Mechanisms â Foreign Vehicles. Unless foreign vehicles are equipped with the same in-vehicle metering equipment, it will be necessary to determine a method for charging those vehicles for road use. The same options available for the OBD II / cellular platform would
87 again be relevant here. If current fuel taxes were left in place, foreign vehicles would simply pay the fuel taxes when refueling. This would be the case if (a) the pay-at-the-pump collection mechanism were implemented, or (b) the central billing agency approach were designed with a mechanism for tracking and rebating fuel taxes paid. If, on the other hand, fuel taxes were phased out, then it would be necessary to employ a more cumbersome approach for charging foreign vehicles for travel in the U.S. The options, as described under the OBD II / cellular approach, include a flat fee, a flat fee with optional odometer reading, odometer-based fees, and the installation of in-vehicle equipment for frequent foreign visitors. Legal Barriers. Federal and state legislation required for this option would be similar to that for the OBD II / cellular configuration. User Burden. The burden on users would be comparable to that for the OBD II / cellular combination, including the initial time required to install the device and, possibly, paying bills on a periodic basis (or setting up automated bill payment) if the central billing agency approach is pursued. Privacy Concerns. Privacy concerns are likely to be considerable with this option (Forkenbrock and Kuhl 2002, Whitty 2003). Though it would not meter location at a greater level of precision than the OBD II / cellular option, the use of GPS appears to be clearly linked with privacy concerns in the public debate. To achieve a minimum level of public acceptability, it will be imperative to employ one or more strategies to ease privacy concerns. Ability to Audit. This faces similar challenges as the OBD II / cellular option with regard to the issues of protecting privacy and allowing motorists to audit their charges. Further research on this topic would be beneficial. Near-Term Flexibility. The flexibility of this approach is comparable to the OBD II / cellular option. The in-vehicle equipment enables accurate metering of mileage by jurisdiction or area, making it possible to vary charges by vehicle characteristics, by jurisdiction, or by time and area of travel (e.g., area-based congestion tolls). By adding gantries at strategic points throughout the road network, it would also be possible to levy facility-based or cordon congestion tolls. Support for Longer-Term Transition. As with the OBD II / cellular combination, this configuration allows for almost any conceivable pricing option. The sole exception would be fees that vary by specific route across the entire road network, applicable mainly for weight- distance truck tolls. Thus it would not be necessary to transition to a more robust VMT-based revenue system over the longer term; this platform would provide all required capabilities. Summary. The strengths, limitations, uncertainties, and key planning decisions for the coarse- resolution GPS option can be summarized as follows: Strengths ⢠Core configuration offers flexible pricing
88 ⢠Addition of gantries enables additional forms of pricing ⢠Allows for accurate apportionment of fees by jurisdiction ⢠Potential ongoing cost savings if cellular service not required ⢠Accommodates all vehicle types; no parallel payment mechanism required ⢠Can support longer-term sustainable revenue system; no transition required Weaknesses ⢠Likely the highest costs for in-vehicle equipment ⢠Likely the highest costs for equipment installation ⢠Potentially higher administrative cost if central billing agency used to collect fees ⢠Difficult to charge foreign users unless pay-at-the-pump collection model employed ⢠Enforcement more difficult, perhaps more costly ⢠Potentially higher user burden depending on enforcement strategies ⢠Highest level of privacy concerns Uncertainties ⢠Cost of in-vehicle equipment manufactured at scale ⢠Cost implications if GPS becomes standard vehicle equipment ⢠Costs associated with pay-at-the-pump system, if implemented ⢠Costs associated with central billing agency system ⢠Costs associated with debit card payment option ⢠Cost and effectiveness of alternate enforcement options ⢠Whether privacy concerns associated with GPS can be overcome ⢠How to enable billing audits while protecting privacy Key planning decisions ⢠Strategy for procuring technology (see next chapter) ⢠Method(s) for collecting revenue â pay at the pump, central billing, and/or debit cards ⢠Public vs. private administration of central billing agency ⢠Strategies to enforce compliance ⢠Method for charging foreign vehicles, if fuel taxes phased out ⢠Strategies to enforce compliance ⢠Whether to augment metering capabilities with DSRC gantries
89 8.4. Shared Obstacles Despite their promise, the three VMT fee options discussed in this chapter share several important limitations to consider â most notably increased cost and administrative complexity. Evidence to date suggests that any of the options will involve higher â likely much higher â collection costs than current fuel taxes. At the same time, moving the point of collection from a relatively small number of entities (fuel wholesalers) to a much larger number (either retail fuel stations or individual motorists) raises the possibility of increased tax evasion. All three would also be far more complex administratively; depending on the specific option, it may be necessary to develop or secure new tax collection channels; a new national agency or expanded state powers; cooperation from entities not currently involved with fuel tax collection, such as cellular providers and retail fuel stations; participation of the IRS; national technological specifications and a system for certifying compliance; and enabling or conforming state legislation. Finally, while VMT is projected to grow more quickly than fuel consumption, it would still be necessary to either index VMT fees or institute periodic raises to offset the effects of inflation. This would likely engender the same level of political resistance as the prospect of raising fuel taxes does today. These factors should receive due consideration in the debate over whether, and at what pace, to pursue a transition from fuel taxes to VMT fees. 8.5. GPS-Based Weight-Distance Truck Tolls The three metering concepts explored in this chapter could apply to trucks as well as cars. Given, however, that a key issue in charging trucks for road use pertains to the maintenance costs they impose, and that this in turn depends on their weight as well as their specific routes of travel, it could be beneficial to develop a separate system for implementing VMT-based fees for trucks. For the required level of locational specificity, such a system would need to rely on precise GPS receivers as well as digital road maps more accurate than those currently available. To properly account for wear and tear imposed on the roadway, the rate structure would also include some measure of the vehicleâs weight (i.e., the per-mile rate would be higher for heavier trucks than for lighter trucks). The current research is intended to focus on near-term options for general-purpose VMT fees, and so excluded a comparable depth of consideration for a GPS-based charging mechanism specifically for trucks. It is nonetheless useful to consider briefly the potential advantages and drawbacks of near-term implementation of electronic weight-distance truck tolls reliant on GPS- enabled equipment. Advantages. There are several advantages â some significant â worth mentioning. These can be summarized as follows. ⢠Near-term feasibility. One of the principal motivations for this project stems from concern that it would not be possible to implement a flexible, general-purpose system for levying VMT fees reliant on GPS technology in the near term (although, as indicated by our inclusion of the coarse-resolution GPS option above, our research suggests that this option merits further consideration). Limiting factors include the cost of retrofitting the existing fleet of vehicles as well as public concerns over potential privacy issues. While not entirely absent, these two concerns are much less relevant for the trucking industry. First, in
90 comparison to passenger vehicles, trucks would face higher per-mile fees based on their weight, and they travel many more miles each year. Thus the cost of the in-vehicle equipment could be amortized much more quickly. Additionally, the trucking industry, as a commercial enterprise, is already subject to a much stricter regulatory regime than other motorists. This reduces the privacy concerns associated with metering travel by time and location. ⢠Higher revenue. With respect to road maintenance, the current system of fuel taxes creates a cross subsidy between lighter passenger vehicles and heavier trucks. That is, the percentage of road damage attributable to trucking, particularly on lightly engineered surface streets, exceeds the share of fuel taxes paid by trucks. Developing a weight-distance truck tolling system that adequately captures road damages would therefore likely lead to an overall increase in revenue from this segment of road users (Small et al., 1989). ⢠Reduced road damages. Establishing a weight-distance truck toll could also lead to a reduction in road damages. If structured on the basis of axle-weight, the fees would create an incentive for truckers to adopt rig configurations with more axles to reduce road use fees. This in turn would lead to a reduction in actual road damages (Small et al., 1989). Disadvantages. There are several disadvantages that should also be considered. ⢠In-vehicle equipment cost. To meter travel by specific route, relatively precise GPS receivers would be required, likely entailing a cost of several hundred dollars per unit. As noted above, however, even this higher cost could be recouped quickly. ⢠Cost of developing accurate digital road maps. Work by researchers at the University of Minnesota (e.g., Trach et al., 2005) demonstrates that currently available digital road network maps are not sufficiently accurate to enable the detection of route-specific travel. It would therefore be necessary to develop (and maintain over time) a more accurate digital road map for the entire United States â entailing a considerable expense. ⢠Resistance among the trucking industry. To the extent that weight-distance truck tolls place a greater financial burden to truckers, the trucking industry can be expected to resist this concept. To gain greater buy-in, it will likely prove necessary to ensure that the program also creates benefits for the trucking industry. Allocation of revenue will be key. For example, revenue could be allocated to relieve congestion along important trucking corridors or to create new truck-only lanes. The potential of automated weight-distance truck tolls to reduce the burden of regulatory compliance among trucking firms should also be highlighted. For instance, electronically-enabled weight-distance truck tolling systems could automate the collection and reporting of data under IRP and IFTA. They would also reduce the effort associated with the current, manually implemented weight-distance truck tolls in Oregon, Kentucky, New York, and New Mexico. ⢠Privacy concerns. Because the trucking industry is already subject to significant federal and state regulations, privacy concerns may not be as high as with general passenger traffic. On the other hand, the specific pattern of truck pickups, routes, and drop offs may be viewed as proprietary information that supports a firmâs competitive advantage. Firms will therefore want to ensure that the privacy of such information is maintained.
