Dedicated Revenue Mechanisms for Freight Transportation Investment (2012)

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.

34 Concept A VMT fee system would charge vehicle owners based on the number of miles traveled on public roads and highways. VMT is a direct measure of vehicle activity. VMT fees (also called mileage-based user fees, or MBUFs) are the leading candidate to replace fuel taxes as the major source of federal highway funding and are discussed extensively in the literature. The proposed switch from fuel-based taxation to mileage-based fees is motivated in part by the projected erosion of fuel-based tax revenue due to improved fuel efficiency and alternate sources of power. National VMT for passenger, freight, and service vehicles, in contrast, is expected to grow indefinitely. VMT fees would also facilitate coverage of hybrid, alternative fuel, and electric vehicles that would otherwise make little or no contribution under the current fuel tax regime. Mileage-based fees have been addressed throughout the reviewed literature and are generally perceived as the leading candidate for funding a freight infrastructure program. There are as yet no detailed proposals for U.S. VMT fees. There have been a small number of tests and pilot projects, but most dis- cussions of VMT fees or MBUFs remain at the conceptual level. Moreover, analysis to date of VMT fees has focused almost exclusively on passenger cars. There is a truck VMT system in operation on the German Autobahn network, but information on freight VMT system concepts is otherwise very sparse. While there are numerous variations on VMT fees, they have in common the concepts of fees based on vehicle type and miles traveled and of automated collection via on-board devices of some kind (often referred to generically as OBUs). Proposals for VMT fees can be split into two basic types: • Distance/vehicle VMT fees, in which the fee is based on the miles traveled, regardless of where the mileage is incurred. For freight applications, VMT fees would presumably vary by vehicle type and weight class. • Time/location VMT fees, in which the fee would vary by some combination of time and location, as well as by vehicle type or class. Time/location VMT fees would allow incor- poration of TDM options such as congestion pricing, road pricing, and tolling. The National Surface Transportation Infrastructure Financ- ing Commission has recommended that the United States move to such a system. If a VMT fee system were implemented for all vehicles, a higher rate could be charged for medium- and heavy-duty trucks with the additional revenue used for freight infrastructure funding. The scope of this study, however, is focused on separate, dedicated funding sources for freight infrastructure. Distance/Vehicle VMT Fees Distance/vehicle VMT fees would be a revenue-generating mechanism with minimal potential for congestion pricing or other TDM options beyond creating a mild incentive to min- imize distance traveled. By varying the fee with the vehicle type, a distance/vehicle VMT system could collect a weight- distance tax similar to that levied in Oregon. Proposals to date for VMT fees do not address the differ- ences between vehicle types, nor are there any concrete pro- posals for VMT fee amounts. Some VMT fee proposals implicitly or explicitly anticipate varying the fee according to the size or weight class of the vehicle affected. This approach may be contrasted with VMT fees for passenger vehicles or private automobiles, which may or may not vary with vehicle type or size. There are no known proposals to assess fees for freight vehicles on the basis of distance traveled only (which would be a pure VMT fee), but neither are there concrete proposals for graduating VMT fees by vehicle type or weight. Time/Location VMT Fees Time/location VMT fees that can vary by location, route, or time create the option for congestion pricing, tolling, road pricing, or other TDM measures. The means by which time, C h a p t e r 4 Fees for Vehicle Miles Traveled

35 location, or route might be determined include GPS, auto- matic vehicle location (AVL), DSRC, and other technolo- gies. It is generally assumed that time/location VMT systems would include all the features of distance/vehicle VMT sys- tems, as well as incremental capabilities to vary fees on the basis of factors beyond distance traveled or vehicle class. A key challenge is identifying the additional cost, complexity, and privacy issues inherent in time/location VMT proposals with TDM capability and balancing them against the ben- efits. A true cost/benefit analysis of TDM and road pricing strategies is beyond the study scope, as is any evaluation of road pricing or TDM concepts themselves. For a time/location VMT system to be operational, vehi- cles need to be fitted with equipment capable of tracking the number of vehicle miles traveled by location, such as installa- tion of a GPS receiver. The GPS receiver would be connected to a geographic information system (GIS) database that records the number of miles traveled within delineated juris- dictions. Each jurisdiction could have a fee schedule based on functional road class, vehicle type, or other factors that is used to calculate fees the vehicle owner owes based on the mileage traveled. Time/location VMT fees would have much higher capital, deployment, and collection costs. In the most elaborate form, time/location VMT could have federal, state, and local com- ponents (National Chamber Foundation of the U.S. Cham- ber of Commerce, 2005, p. IX). Time/location VMT fees would raise issues of privacy and confidentiality. GPS-based or equivalent systems would have the inherent capability to reveal detailed information about where the vehicle has traveled and when. For some, this capa- bility raises concerns regarding personal privacy, although there are means by which privacy concerns can be addressed. In the commercial sphere, GPS data could reveal the location of a company’s customers, the frequency with which those customers are served, the home locations of drivers, and ele- ments of the company’s business strategy. The incremental benefits of time/location VMT fees would mostly consist of opportunities for congestion pricing or road pricing on routes not otherwise equipped for tolling. Unless a national congestion management or toll system emerges, all such benefits would be at the state or regional level and will depend on state or local initiatives for implementation. There would be no significant nationwide benefits to time/location VMT systems. Area-Based VMT Congestion Pricing Area-based congestion pricing would likely be simpler to implement than route-based road pricing, and could use “consumer-grade” or coarse-resolution GPS rather than high-resolution GPS. Area-based VMT fees would entail a surcharge for entering a defined area, typically a congested urban center. The surcharge could be a flat fee, a mileage- based fee, a time-based fee, or some combination. The imple- mentation of area-based congestion pricing would entail knowledge or recognition of vehicle location, thus raising privacy issues. VMT Road Pricing Road pricing, defined here as fees that vary by route, would require high-resolution GPS or an equivalent technology to distinguish routes with sufficient accuracy. High-resolution GPS would be an added cost for consumer systems. High- resolution GPS technology, however, is built into many com- mercial truck location systems, so the marginal cost of adding location capability to the commercially equipped fleet could be relatively low. Only a small portion of the fleet is equipped this way, however. The administrative and collection cost would be substantially higher than for a distance/vehicle VMT system. The marginal benefit of road pricing would depend on how and where it is applied. A system that charged different rates for the Interstate Highway System, or for the designated National Highway System, would be applicable to most of the country. Such a system could, however, encourage diversion to secondary routes less capable of accommodating heavy-truck traffic. Moreover, such a system would turn the Interstate system into toll roads, which is contrary to current rules. Tolling Interoperability Tolling interoperability, often mentioned as a benefit of time/ location VMT systems, is apparently being achieved through other means. A substantial degree of coordination through electronic toll collection has been achieved through E-ZPass, which covers 15 states (see Figure 5). Vermont and Con- necticut have not joined E-ZPass, but those states have no toll facilities. A similar system, FastTrak, covers toll bridges and highways in California. The benefits of a truly national tolling system are questionable, as the frequency with which the same vehicle would use toll facilities in multiple states that are not part of E-ZPass is probably minimal. Parking Fee Collection Parking fee collection is another proposed application of time/location VMT fees (Grush, 2010). There apparently is significant potential to reduce the cost of collecting money from parking meters and issuing tickets and reduce evasion and embezzlement through GPS-enabled fee collection. The vehicle owner (primarily a passenger-car owner) would ben- efit from added convenience, but may also “lose” from more

36 consistent fee collection. The major beneficiaries would be municipal parking authorities. Moreover, the use of time/ location VMT systems to handle parking raises serious pri- vacy issues since it would entail detailed knowledge of vehicle location and time. Finally, the benefits would mostly be in large metropolitan areas. Realization of Benefits The congestion management capabilities of time/location VMT systems are of no benefit unless congestion management systems themselves are implemented. Time/location VMT systems would enable congestion management pricing in any well-defined area and would further enable fees to vary by time and vehicle type. This ability would dramatically reduce the implementation barriers and costs of congestion pricing. The practical feasibility of congestion pricing itself, however, is open to question. Congestion pricing has been discussed for decades, but no serious proposals for implementation have surfaced in the United States. Worldwide, congestion pricing remains rare. For those cities that are the prime candidates, such as Manhattan and downtown San Francisco, bridge, tun- nel, and highway tolls are already being used as rough con- gestion management tools. Perhaps most basically, if VMT fees were only applied to medium- and heavy-duty trucks, the scope of congestion pricing or road pricing would be likewise limited. VMT Fee Pilot Projects There have been three significant U.S. VMT fee pilot projects: • The Oregon Department of Transportation (DOT) Road User Charge Pilot, which took place in 2006–2007. • The National Road User Charge Pilot, which began in 2008 and is continuing under the University of Iowa’s Public Policy Center. • The New York Truck VMT Fee Project, completed in 2011. The Oregon DOT Road User Charge Pilot is the best known and most completely documented of the studies. Vehicles in the study were equipped with GPS-based OBUs that could transmit VMT data at specially equipped gas stations. Almost 300 motorists participated, but no commercial trucks were involved. The Oregon pilot demonstrated the technical fea- sibility of the collection system and successfully dealt with privacy issues by not transmitting or storing detailed move- ment information. The National Road User Charge Pilot has involved par- ticipants in multiple states and cities. No actual VMT fees are collected, and participants are compensated. The major objec- tive of the tests is to determine whether OBUs can perform as expected and whether they are acceptable to drivers. The 4-year project began in 2008 and is expected to be complete in 2012. Figure 5. State coverage of E-ZPass.

37 While published findings are not yet available, the NCFRP Proj- ect 29 study team was able to interview the project manager. The New York study, conducted by Delcan Corporation, is the only pilot project concerned with trucks. The project objective is to “establish a revenue-neutral, mileage-based tax model that generates support from [the] motor carrier indus- try and [the] public sector” (Mudge, April 20, 2010). Pre- liminary findings and observations from this study include confirmation that many large trucks are already equipped with telematics that can track VMT. The contractor antici- pates developing a fee structure that can vary by type of road (e.g., thruways versus local roads) and by time of day (e.g., peak vs. off-peak)—essentially a time/location VMT system. The preliminary proposed base fee is $0.106 per non-thruway mile and $.051 per mile plus existing tolls for thruway travel. The project will also investigate the use of IFTA filings for VMT fee payment. The Oregon DOT Road User Charge Pilot and the National Road User Charge Pilot are both revenue neutral. The New York study includes examination of different fee structures for trucks but it, too, is revenue neutral. None of these stud- ies addresses public acceptance of VMT fees in addition to current fuel taxes. The New York pilot is focused on trucks and should yield more insight into the workability of VMT fee systems for truck fleets. The principal results of these pilot projects to date has been to establish technical feasibility, gain some limited insight into costs, and determine that participants are more favor- ably disposed toward VMT fees and privacy issues after they have experienced the system. International Experience Truck VMT fees are in place in Switzerland and Germany. Swiss road pricing dates back to 2001. The current system uses on-board GPS-based units that record vehicle route and mile- age. The data are transmitted monthly to the Swiss Customs Agency, who then bills the truck owner. More than 60,000 trucks participate. The German truck toll system is based on each truck’s emissions classification, axle loading (weight per axle), and mileage. The GPS-based, electronic system began on January 1, 2005, and its major goal is to raise revenue. The system uses multiple technologies. The Swiss and German sys- tems, like the Oregon and national pilot studies, demonstrate the technical feasibility of VMT systems. The German system also provides some precedent for fee differentiation based on emissions characteristics and axle loading. A third European system proposed for the Netherlands was recently cancelled. The Dutch system, which would have applied to both passenger vehicles and trucks, was deemed too controversial by a new government and the project was shelved. In retrospect, it appears to close observers that insuffi- cient attention was paid to institutional and political questions while technical problems were being solved (Jongman, April 21, 2010). Congestion Pricing Examples Examples of congestion pricing can be seen in Singapore, Stockholm, and London: • The Singapore example is essentially an automated toll sys- tem. Entry fees are charged to all vehicles entering central Singapore. The area-based system is implemented through DSRC tags on all vehicles. DSRC readers on the entrance routes to central Singapore are placed on overhead gantries. The system charges the fee to a debit account, but does not identify the vehicle, driver, or time. No data are stored in the OBU. There are no fees for vehicles that operate solely within the toll perimeter. • Stockholm, like Singapore, is an “island” city with limited access routes. Tolls are charged on those routes for vehicles entering the central city. As in Singapore, vehicles that stay within the toll perimeter are not charged. • Access to central London is priced using license plate rec- ognition (LPR) technology. Time/location VMT fee systems would be a more flexible means of implementing such TDM measures. In effect, the Singapore, Stockholm, and London systems charge vehicles for using routes that lead into congested areas. GPS-based or equivalent time/location VMT systems would allow more variability in fee structure by location, time of day, and so forth. The cost of collection would be reduced for the con- gestion management agency, but the capital cost to vehicle owners would be much higher. Concept of Operations Collection Requirements The complexity of distance/vehicle and time/location VMT fee systems requires more elaborate collection mechanisms than other revenue-generation strategies. For distance/vehicle fees the collection mechanism should do the following: • Identify the vehicle and type (e.g., VIN) and link it to a fee account and the vehicle owner. • Track mileage and determine the total since the last reading. • Communicate automatically with a stationary reader, com- municate via cell phone or Internet, provide downloadable data, or allow for manual reading. • Provide for auditing and enforcement. • Resist evasion or tampering.

38 Dedicated On-Board Units Most analysis of VMT options to date assume or con- clude that vehicles would be equipped with some kind of device (generically an OBU) to automatically track, record, and report mileage for distance/vehicle VMT systems. In their simplest form, VMT-compatible OBUs would record mileage traveled and report that mileage when queried elec- tronically (via DSRC, RFID, or cellular) or read manually. In effect, such devices would be remote-reading odome- ters. The device would also have to identify the VIN, which would be linked to a VMT fee account. The need for the VIN would be twofold if distance/vehicle VMT fees vary by vehicle type. For an OBU to obtain mileage data from the vehicle itself, it would have to be either retrofitted to a wide variety of mechan- ical or electronic odometers or built into new vehicles by the original equipment manufacturer (OEM). GPS technology is not necessary for distance/vehicle VMT fee implementation. As multiple studies have pointed out, the OBD II data link, standard on newer automobiles, provides electronic access to odometer readings. Transmission or down- loading of these readings is sufficient for distance-based systems. Similar connections provide access to electronic odometer readings for trucks. Older trucks, however, do not have such connections and would require more complex installations or manual readings. One preliminary observation of “The National Evaluation of a Mileage-based Road User Charge” is that passenger-car retrofits are unlikely to be a practical strategy due to the com- plexity of OEM electronic systems on newer cars. Moreover, retrofitted OBUs would likely be far more vulnerable to tam- pering than units integrated with OEM systems. Commercial OBUs or EOBRs for trucks are designed as either retrofits or OEM installations and can successfully interface with most truck electronic systems. “The National Evaluation of a Mileage-based Road User Charge” has distinguished two OBU types or approaches with markedly different implications for privacy and col- lection cost. “Thin” OBU clients (see Figure 6) transmit raw information on vehicle identity, miles traveled, location (rel- ative to taxing jurisdictions), and perhaps route and time. Information is compiled, and VMT fees are determined and assessed at a centralized location. “Thick” OBU clients (see Figure 7) download information on fee rules periodically. The “thick” OBU client then calcu- lates corresponding VMT fees for vehicle activity and trans- mits only fee information and vehicle identity at designated billing intervals. While a thick OBU technology may be more expensive, the thick OBU approach would drastically reduce bandwidth and central processing needs, as well as preserving privacy. In this respect, a thick OBU is akin to a digital taxi meter that applies a predetermined rate structure to cumulative vehicle activity and returns a calculated fare. The thick OBU client would effectively be an electronic third party, taking infor- mation from the operator and calculating (and possibly pay- ing) fees owed. The public fee collection agency would then be in the position of establishing rules and procedures, dis- seminating rules to thick OBU clients and other users, receiv- ing payments, and auditing and enforcement. Commercial Telematics The use of commercial telematics systems coupled with third-party VMT fee calculation, reporting, and payment Figure 6. “Thin” client data flow.

39 could greatly facilitate implementation of a VMT fee system for trucks. The generic term “telematics” covers a wide range of commercial or in-firm systems installed in trucks to com- municate with central dispatchers and information systems. A substantial portion of the commercial vehicle fleet (esti- mated by the study team at roughly 10%) is already equipped with EOBRs or GPS-based or cellular-based location and communications systems such as those offered by Qual- comm, Teletrac, or numerous other vendors. These telemat- ics systems have been installed for the fleet operator’s own reasons, including monitoring hours of service compliance and driver performance. As the costs of these systems decline in real terms, their advantages become more apparent. As the need for mandated compliance monitoring increases, these systems will become more common. The VMT-compliant OBUs proposed for use in VMT fee collection, the EOBRs used by many truck operators, and the broader classification of “telematics” have a complex relation- ship. Appendix B addresses this issue in detail. The functions of these systems can overlap, with most EOBRs and telematics systems in use having the capability to record and report vehicle VMT. More elaborate commercial trucking systems with GPS or other AVL capabilities might also meet the requirements of time/location VMT fee collection. If VMT system implemen- tation provides options for compliance, many truck operators would likely choose to use or upgrade their existing systems rather than incurring the cost of additional technology. It is reasonable to expect that commercial or in-house EOBR and telematics systems developed in the future would support any implemented VMT fee collection system, given the rapid turnover of telematics technology and the multi- year implementation expectation for VMT systems. Although EOBRs and telematics systems can be costly in terms of acqui- sition, installation, service fees, and upkeep, the incremental cost of VMT compliance should be low for distance-based fees. Thick-client capabilities would be somewhat more costly than thin-client capabilities. Time/location VMT capabilities would add high incremental costs to EOBRs that simply recorded hours, miles, and vehicle conditions for download, but low incremental costs to GPS/AVL telematics applications. Another option would be to require OEM VMT-compliant OBU installation for new trucks. However, the long life of heavy-duty diesel vehicles means that the truck population turns over very slowly, limiting the effectiveness of such a strategy. While it is common for aggressive carriers to turn over their fleet in 5 to 10 years, the vehicles retired by those carriers typically “cascade” into less demanding services where they often last another 10 to 15 years. As port drayage truck replacement and clean truck programs demonstrate, 10- to 20-year-old trucks still operate in large numbers, and 30-year-old trucks are not exceptional. Trucks can last even longer in seasonal or agricultural use. The concept of thick-client monthly fee reporting in place of thin-client vehicle activity reporting raises the possibility of fleet-based reporting and payment for operators of multiple vehicles. Truck fleets equipped with Qualcomm, Teletrac, or similar electronic event recording systems already have the ability to generate detailed monthly reports of vehicle activity. These monthly reports could be the basis for monthly VMT fee submissions for the entire fleet, completely resolving any routine privacy issues and streamlining the collection process. Appendix C provides additional detail on the use of on- board devices in trucking. Mobile Device Alternatives Recent pilot studies in Oregon have demonstrated the feasibility of using mobile “smartphone” applications to Figure 7. “Thick” client data flow.

40 automate the Oregon weight-distance tax. The Oregon Department of Transportation’s (ODOT’s) Truck Road Use Electronics (TRUE) test used five modified BlackBerry smartphones with a custom computer application to calcu- late mileage, combine the mileage with data on truck con- figuration and number of axles, calculate the tax, and create a tax bill for payment. The test was successful, and, although it was not a direct test of VMT applications, it suggests that there may be ways to implement VMT systems without vehicle OBUs. The use of smartphones in this application is also consistent with technology trends in the trucking industry, in which low-cost cellular-based communica- tions are being used for new purposes. Recent revelations that iPhones were retaining location data reinforce the conclusion that non-GPS smartphone sys- tems are a possibility. The adverse public and government reaction to the iPhone revelation, however, illustrated public antipathy to the concept of location tracking. Fuel Consumption Basis for Collection A VMT fee based on fuel purchase and estimated miles per gallon would be a fuel tax that varied by vehicle type. Presum- ably, estimated miles per gallon would be designated for each vehicle based on size, weight, engine type, or some combina- tion of factors. To vary fuel taxes by vehicle, however, requires either a means of automatically identifying the vehicle and adjusting the tax rate at the time and place of purchase, or retroactively adjusting the tax rate through refunds or credits. This mechanism, then, is essentially the same as the proposals for fuel tax surcharges or selective rate increases. Vehicle fuel consumption can vary greatly depending on the individual vehicle, the route traveled, and the load carried. Two identical Class 8 diesel tractors could have very different average fuel consumption if one were used for urban gasoline deliveries (traveling short distances in congested conditions and hauling empty trailers half the time) and the other were used by an interstate truckload firm (traveling long distances at highway speeds with lighter trailers but loaded 90% of the time). Fuel consumption under such circumstances would not be an accurate proxy for VMT. Self-Reporting Self-reporting of odometer readings as a VMT implemen- tation strategy has been considered; however, it has been dismissed in the literature as being cumbersome, costly, unreliable, and prone to evasion. Self-reporting, however, is already universal practice for all truck fleets that operate and purchase fuel across state boundaries and for other fleets and drivers subject to hours-of-service rules. Interstate operators are required to track and report vehicle miles traveled and fuel purchased by state so that state fuel tax revenue can be apportioned according to miles driven. For such carriers, self- reporting of fleet total VMT is already in place. Many other fleet operators record mileage for their own management purposes. Figure 8 is an example of a standard truck driver’s log used to record hours, miles, and other activity. For-hire and private fleet operators may collect this data to monitor driver hours of service, fuel use, vehicle utilization, main- tenance needs, and other operating factors. Firms that do Figure 8. Sample driver’s log form.

41 so already have VMT records that could be used as a basis for fee calculation and auditing, although precision may be lacking. There is a legitimate concern over accuracy and com- pleteness in self-reporting options. Drivers can and do fal- sify logs and may not have strong incentives to accurately report miles traveled There are no readily available, reliable data on the prevalence of this practice, partly because there is rarely an independent audit trail against which logs can be checked. Self-reporting would probably have to remain as a fallback option during any implementation period for automated reporting. Operators who have not installed OBUs capable of tracking and reporting VMT or have not implemented other solutions could be afforded a manual tracking and self- reporting option, with appropriate auditing. Periodic Inspection Some proposals suggest obtaining odometer information as part of routine vehicle inspections. This option is consid- ered to be very costly in terms of time and labor. States that do inspect vehicles typically do so annually, which would mean a long time between large payments. A number of states, notably California, do not inspect vehicles annually. California has a biennial emissions certification process in which mileage information could be taken, but there is no verification by any public body. One collection option would be a prepayment system with annual adjustments made through periodic inspections. This option would be similar to estimated income taxes, where quarterly tax payments are based on the previous year’s income and reconciled with annual filing. Periodic inspection could also be used for compliance auditing and enforcement. By linking reported VMT to vehicle registration, it would be relatively easy to “flag” reg- istrations for vehicles reporting suspiciously low VMT. Such vehicles could have their odometers checked before registra- tion was renewed and any mileage discrepancies (and any penalties) addressed at that time. While odometer tamper- ing is all too feasible at the time of a vehicle sale, it would be more difficult to consistently alter odometer readings at regular inspections. Passing Readers Passing inspection via overhead gantries or wayside read- ers is commonly used for DSRC/RFID toll card systems such as E-ZPass. Overhead gantries and wayside readers are only practical, however, when vehicle traffic must pass through a limited number of check points. Successful applications include road pricing systems in Singapore and Stockholm, island cities whose entrances and exits are limited. Such installations are unlikely to be suitable for rural and subur- ban areas or sprawling metropolises, such as Chicago or Los Angeles. License plate optical character recognition (OCR) readers have been used in Canada and London for tolling systems. All of these passing reader options may be suitable for triggering payments for congestion pricing, road pricing, or other TDM functions, but would not be feasible for col- lecting VMT data. The problem of physical data collection is most challeng- ing for freight and service vehicles that spend most of their operating life in one locale. A substantial portion of the total truck population has a routine operating range of 50 miles or less. Such vehicles might go for long periods without passing a data collection point, or these vehicles may be able to avoid them indefinitely. Examples might include propane delivery trucks in rural areas, farm trucks that are rarely driven into urban areas, local delivery vehicles used on a fixed route, and local tradesman’s vehicles and construction equipment. Gas Station Data Collection One common proposal is to “read” electronic odom- eters at gas stations and to add the VMT fee to the amount charged for fuel. This system was used in the Oregon DOT Road User Charge Pilot. Conceptually, the system would function as follows: • Vehicles would be equipped with OBUs. • Gas stations (including unmanned “card-lock” fueling sites used by truck fleets) would be equipped with electronic OBU readers. • At the station, the OBU would communicate the mileage and vehicle identity to the user fee system, and the fee sys- tem would determine how many miles had been driven since the last reading and add the fees to the fuel bill. This option raises some implementation questions. Most discussions of a system based on gas station data collection appear to assume a credit card transaction in which the driver can simply add the VMT user fee to the fuel bill. In such a case, the driver could fill the tank at whatever cost and also pay the VMT fee (assuming a sufficient credit card limit). A cash customer could face a problem. For example, a customer with $2.75 due in VMT fees can only buy $17.25 worth of fuel with a $20 bill. Thus, the fuel pump may have to display the VMT fee due before the purchaser decides how much gas to buy. Such an eventuality would require modification of fuel pumps, a costly proposition. Although it may be uncommon for cash to be used to pay for enough fuel to fill the 250-gallon tank on a Class 8 tractor, cash transactions for smaller medium-duty trucks are probably more prevalent.

42 Bulk Fuel The proposed collection of VMT fees at gas stations or with fuel cards raises the question of vehicles that do not frequent gas stations. These vehicles include fleets whose fuel is pur- chased in bulk and dispensed at operating terminals. Table 12 shows the distribution of fueling locations for medium and heavy-duty trucks using 2002 VIUS data. About 34% of the total trucks use either their own facilities or other facilities not open to the general public to refuel, and they account for about 35% of the VMT. For freight transportation, the category of non-exempt bulk fuel purchasers is the most serious issue. Bulk-fueled fleets likely include major parcel or less-than-truckload (LTL) companies such as UPS, FedEx, or Yellow Freight that maintain terminals as operating bases for their truck fleets; construction companies; and utilities. The issue is most serious for the largest fleets. As Table 13 shows, the share of trucks using their own fuel facility or other fuel facilities not open to the public grows with fleet size. The widespread use of non-public fueling facilities would complicate any VMT reporting system tied to commercial fuel purchases. The options for obtaining odometer readings for such fleets include using EOBRs or telematics, installing OBU readers at company terminals, requiring periodic visits to gas stations or other reader sites, and periodic self-reporting. One mitigating factor is that very large freight and service fleets are probably the most likely to use OBUs or EOBRs for their own communications, dispatching, and management purposes. Fleets using such devices would not need to check odometer readings at fueling locations. There are also commercially available fuel management systems that address the bulk fuel issue. A February 2010 Municipal Fuel Management Advisory by the Massachusetts Office of the Inspector General reviewed a wide range of issues related to control of fuel purchases and delivery by public agencies and recommended the use of care- fully managed fuel cards and electronic management sys- tems for bulk fuel (Massachusetts Office of the Inspector General, February 2010). Research into available systems for managing bulk fuel located three types applicable to VMT fees: • Keypad entry systems, in which the driver enters vehicle identification and odometer readings (FuelMaster is one commercial system). • Automated systems that use wireless communication between vehicle ID tags and dispenser systems to control Fueling Location Trucks (thousands) Trucks % Miles (millions) Miles % Gas Station 2,441 49% 36,263 26% Truck Stop 862 17% 52,809 38% Own Facility 1,565 31% 43,791 32% Other Non-public Facility 146 3% 4,448 3% Total 5,014 100% 137,311 100% Source: 2002 VIUS (U.S. Census Bureau, 2004, Table 3a). Table 12. Medium- and heavy-duty truck fueling locations. Truck Fleet Size Gas Station Truck Stop Own Facility Other Facility Total No Fleet 96.4% 1.6% 1.7% 0.3% 100.0% 1 to 5 73.8% 6.1% 18.2% 1.9% 100.0% 6 to 10 55.3% 5.7% 35.5% 3.4% 99.9% 11 to 20 41.1% 5.1% 48.9% 4.9% 100.0% 21 to 50 48.3% 6.3% 44.4% 1.0% 100.0% 51 or More 48.3% 6.3% 44.4% 1.0% 100.0% Overall 93.9% 1.8% 3.7% 0.5% 99.9% Source: U.S. Bureau of the Census, "2002 Vehicle Inventory and Use Survey," Microdata file, January 2005. Data generated by Stacy Davis, Oak Ridge National Laboratory, March 2005. Table 13. Share of trucks by fleet size and primary fueling facility—2002.

43 access and record usage (Easy Fuel and Fleet Control are two commercial systems). • Mobile fueling trailers with monitoring equipment. These systems, like the fuel card systems, offer regular reporting of vehicle and fleet mileage that could be used to calculate VMT fees. These systems may be justified for fleet operators, such as public agencies, construction firms, utilities, and trucking companies. They are unlikely to be cost-effective for farmers or others with small fleets. Fuel Cards Fuel companies offer fleet credit card options that can be used to differentiate VMT fee vehicles from fuel tax vehi- cles, to track fuel taxes paid, and to track manually entered odometer readings with fuel purchases. One simple approach would be to issue VMT-compliant fleets with fuel cards that automatically add taxes to fuel purchases. This approach would implement “gas station” payment without any addi- tional technology. Gasoline credit cards already offer features that could greater facilitate fuel tax/VMT transition. Chevron and Texaco Business Card Services, for example, offer multiple options to fleet managers, including links between cards and specific vehicles, and a requirement for the driver to enter odometer readings before receiving fuel. Tax-exempt cards are available for government fleets. Cards can be linked either to drivers or to vehicles. Fleet fuel cards—such as CSI, Wright Express, Fuel Express, or Fleet One—are intended for commercial or private truck fleet operators and are accepted at most fueling stations regardless of brand or ownership. These services offer extensive reporting options. Drivers can be required to enter vehicle ID, personal identification numbers (PINs), and odometer read- ings when buying fuel. The resulting reports detail mileage, fuel use, and miles per gallon by vehicle and for the fleet. These cards can thus be used to calculate and pay vehicle and fleet VMT fees automatically without any additional technology. Fuel cards are, in theory, available to any vehicle owner. In combination with third-party reporting and payment, fuel cards or gasoline company cards could provide a non- technological implementation of VMT fees. For fleets using these fuel cards, implementing VMT fees would be a nearly costless accounting change. Fuel cards are credit cards, and their transaction cost is borne by the retailer as a percentage of sales. Reliance on fuel card systems for VMT fee collection does, however, have some potential drawbacks. The difficulty is linking the credit card to the vehicle and ensuring that the vehicle being fueled is indeed paying VMT fees. Public agen- cies sometimes have problems with employees using govern- ment credit cards to fuel private vehicles. The potential for evasion could be significant if tax-exempt credit cards are used to fuel non-VMT vehicles in a mixed fleet or vehicles in an entirely different fleet. Many truck drivers carry and use multiple fuel cards. If a driver used Card A, reported 750 miles since the last reading, and then switched to Card B, he could be charged duplicate VMT fees for the 750 miles on Card A. The issue becomes more complex if drivers use a mix of VMT-compliant fuel cards, conventional fuel or gasoline company cards, other credit cards (e.g., Visa or MasterCard), debit cards, company charge accounts, cash, and bulk fuel. Many fuel retailers already offer lower prices for cash payment, and an oppor- tunity to evade VMT fees would increase the incentive to use cash for drivers or fleet operators. It would be extremely costly to attempt to audit cash fuel sales to determine whether VMT fees had been paid. International Fuel Tax Agreement Reporting Trucks that operate and purchase fuel in more than one state already reconcile taxes paid and fleet VMT by state under the IFTA system. IFTA is an agreement among all states (except Alaska and Hawaii) and all Canadian provinces (except Northwestern Territories, Nunavut, and Yukon) to simplify the reporting of fuel used and tax paid by motor car- riers operating in more than one jurisdiction. Persons or organizations that operate qualified motor vehicles are subject to IFTA licensing. A qualified motor vehicle is defined as a motor vehicle used, designed, or main- tained for transportation of persons or property and having one of the three following characteristics: • Having two axles and a GVW or registered GVW exceed- ing 26,000 pounds (i.e., GVW Class 7–8). • Having three or more axles regardless of weight. • Used in combination, when the weight of such combina- tion exceeds 26,000 pounds gross vehicle or registered gross vehicle weight. Vehicles enrolled under IFTA display IFTA decals issued by their home state. In 2009, state DOTs issued 2.5 mil- lion sets of IFTA decals for trucks, so approximately that many vehicles were covered by the IFTA system. That total is roughly 25% of the medium- and heavy-duty fleet. There were about 290,000 IFTA accounts, suggesting an average of a bit less than 10 vehicles per account (2009 IFTA Annual Report). Table 14 indicates that about 11% of the truck pop- ulation routinely operates interstate. Together, the 25% with IFTA decals and the 11% routinely operating interstate sug- gests that many vehicles occasionally operate interstate and have IFTA decals for that purpose or that fleet operators may

44 routinely enroll the whole fleet in IFTA even though some of the trucks usually remain within one state. If a VMT system took only mileage into account, the IFTA system could be used largely as is, simply by calculating the VMT fee. If the VMT fee varied by vehicle type, the fleet oper- ator would have to total the mileage traveled by each type or class of vehicle and then calculate a total amount due. Note that an IFTA-like system would enable a fleet opera- tor to make a single payment for all trucks in the fleet. This approach would obviate the need for a public agency to query individual OBUs for each truck. Because the IFTA return also shows gallons of fuel purchased, the IFTA-like return could be audited for compliance using standard ranges of fuel consump- tion. The company’s internal records would then become the audit trail, if needed. In this regard, an IFTA-like return would function like a “thick” OBU client, except that it would be for an entire fleet rather than for an individual vehicle. IFTA is a self-reporting system, whether the reported mile- ages are obtained from OBUs or manual logs. Since the pri- mary purpose of IFTA is allocating fuel taxes already paid, the fleet owner does not have a major financial stake in accuracy. If an IFTA-like system were used to determine VMT fees owed, fleet operators would have an incentive to under-report mileage and auditing may be necessary. Distance/vehicle VMT fees would depend only on total mileage accuracy by vehicle type, but implementing time/location VMT fees may not be feasible through an IFTA-like system. Third-Party Reporting Third-party reporting could reduce the number of truck- ing industry fee payers and the accompanying government collection cost. There are multiple options for third-party VMT reporting and fee payment including the following: • Commercial telematics or EOBR providers who calculate and pay the VMT fees on behalf of the fleet operator. • Fuel card providers who offer third-party reporting and payment. • IFTA, which is itself effectively a third-party tax revenue allocation system, and could become a VMT revenue allo- cation system. Owned versus Leased Vehicles Implementation of VMT systems will have to take the dif- ference between owned and leased units into account. While it is operation of the vehicle that accumulates VMT, it may be necessary to assess and collect VMT fees through the reg- istered owner. Freight and service fleet operators use a mix of owned, long-term leased, and short-term rental vehicles to meet their needs. Table 15 gives some examples of large fleets and the reported ownership and leasing mix. Many major leasing companies, notably GE Capital, offer comprehensive fleet management services that include payment of taxes. Primary Jurisdiction Trucks (thousands) Trucks % Miles (millions) Miles % Intrastate 4,196 89% 84,974 68% Interstate 496 11% 40,901 32% Total 4,692 100% 125,875 100% Source: 2002 VIUS (U.S. Census Bureau, 2004, Table 3a). Table 14. Intrastate versus interstate operations. Fleet Total Vehicles Owned Leased/Rented AT&T 77,480 94% 6% UPS 72,633 100% 0% Verizon 58,768 100% 0% FedEx 36,701 100% 0% Coca-Cola Ent. 22,455 100% 0% PepsiCo, Inc. 19,297 92% 8% ServiceMaster 15,198 32% 68% Salvation Army 14,600 0% 100% Qwest Communications 13,700 0% 100% DHL Express 18,586 93% 7% Xerox Corp. 10,000 0% 100% Interstate Brands 8,238 75% 25% Aramark Services 8,208 30% 70% Source: Automotive Fleet, 2009. Table 15. Ownership/leasing mix of large fleets.

45 Time/Location Collection Requirements Time/location VMT fees that can vary by location, route, or time create the option for congestion pricing, tolling, road pricing, and other TDM measures that are ordinarily absent in distance/vehicle VMT proposals. The means by which time, location, or route might be determined include GPS, AVL, DSRC, and other technologies. For a time/location VMT system to be operational, vehi- cles need to be fitted with equipment capable of tracking the number of vehicle miles traveled by location, such as a GPS receiver. The GPS receiver would be connected to a GIS database that records the number of miles traveled within delineated jurisdictions. Each jurisdiction could have a fee schedule based on functional road class, vehicle type, or other factors that is used to calculate the fees that the vehicle owner owes based on the miles traveled. The popularity and features of satellite-based GPS and cellular-based AVL systems have led to their consideration as VMT fee implementation options. Because such systems could track and report mileage accurately by time and loca- tion without being integrated with mechanical or electronic vehicle systems, they are the leading candidates for VMT fee implementation. Such devices have been used in some of the pilot studies. Once it is assumed that a GPS/AVL-enabled OBU will be required for distance/vehicle VMT fees, the incremental cost of implementing more elaborate time/ location VMT fees appears low. The use of GPS systems for congestion pricing, road pric- ing, tolling, or other applications is limited by system resolu- tion. So-called “consumer grade” or “coarse resolution” GPS may have a claimed accuracy of ±5 meters, but the available literature suggests that this accuracy is not always achieved, particularly in adverse conditions. Moreover, an accuracy of ± 5 meters is not sufficient to distinguish lanes on a highway, to distinguish operation on a frontage road from an adjacent freeway, or to distinguish operation on a public road from off-road operation on an adjacent construction site. Research on GPS-enabled vehicle systems under NCFRP Project 14, “Truck Drayage Systems,” found that false readings could be generated by vehicles operating just outside geofenced perimeters. At best, it appears that consumer-grade or coarse- resolution GPS could support area-based congestion pricing under which fees are charged for entry to a well-defined cen- tral business district (e.g., Manhattan) during specified times. So-called “military grade,” or “high-resolution,” GPS sys- tems are capable of extremely fine resolution, but are much more expensive to acquire and operate. Many individual drivers of passenger vehicles find GPS systems attractive and justified for their own purposes, basically location and navi- gation. Those needs are met by a consumer-grade, coarse- resolution GPS. Requiring high-resolution GPS for tolling or road pricing would impose substantial incremental costs, even for those already using GPS. Area-based congestion pricing would likely be simpler to implement than route-based road pricing and might use consumer-grade or coarse-resolution GPS rather than high- resolution GPS. Area-based VMT fees would entail a sur- charge for entering a defined area, typically a congested urban center. The surcharge could be a flat fee, a mileage-based fee, a time-based fee, or some combination. The implementation of area-based congestion pricing would entail knowledge or rec- ognition of vehicle location, thus raising serious privacy issues. Evasion and Enforcement The incentive for evading VMT fees could be significant. The annual Class 7–8 VMT fee estimates shown in Table 25 range from $1,759 to $1,848. By way of comparison, typi- cal owner-operator earnings could range from $25,000 to $40,000 per year, so VMT fees could be 5% of an owner- operator’s annual income. The most serious evasion problems would probably stem from fleet owners who systematically underreported mileage or otherwise manipulated the system. A fleet owner with 10 Class 8 vehicles would have roughly $18,000 annually at stake; a fleet of 100 Class 8 tractors could pay roughly $180,000 in annual federal VMT fees. Methods of Evading VMT Fees A VMT fee system may offer substantial opportunities and incentives for evasion, including the following: • Disconnecting, tampering with, or resetting an OBU. • Disconnecting or resetting a mechanical or electronic odometer. • Underreporting in a manual system. • Underreporting or manipulating a fuel card system. • Avoiding fuel-linked reporting by purchasing fuel with cash. • Operating an unregistered vehicle. Tampering with OBUs The notion of creating a tamperproof OBU system is illu- sory. The simplest tampering method would probably be to cut power to the OBU. GPS-based devices can be defeated by shielding the device from a GPS signal or removing the antenna. GPS signals can also be jammed to avoid revealing GPS location, using readily available and inexpensive devices. These devices are illegal to own or use in the United States, but can be easily obtained through online vendors. While some proposals include “tamperproof” seals on OBUs, research on “tamperproof ” seals for international

46 cargo security and other purposes has found that such seals can be readily defeated (Johnston, 2003). Moreover, the cost of physically inspecting OBUs for seal integrity would be pro- hibitive. There will be over 5 million vehicle operators with incentives to beat the system. A VMT system would therefore entail either a substantially higher rate of evasion or a very high cost of enforcement. Tampering with Odometers Evasion through tampering with electronic vehicle odom- eters is a distinct possibility for a VMT system. Odometer fraud is already considered a significant problem in used- car sales, where “rolling back” the odometer to show lower mileage can raise the sale value. Hardware and software for resetting mechanical or electronic odometers are openly sold. A digital “reprogrammer” capable of changing odometer settings in multiple vehicles sells for $200 to $400, and it is claimed that the procedure takes only a few minutes once the cable connection is made. Instructions for resetting odometers of many kinds are available on the Internet. For example, see the following (as of May 23, 2011): • http://www.youtube.com/watch?v=eCnK3v9PLPM& feature=related. • http://www.youtube.com/watch?v=q4vr1LIOhuI&feature =related. • http://www.youtube.com/watch?v=sRKCr4Nh-tQ& feature=related. At present, odometer tampering appears to be aimed at one-time adjustment for selling used vehicles or avoiding mileage penalties in vehicle leases. Frequent tampering to foil monthly or random VMT fee readings would be more difficult, but not impossible. It is possible to imagine a fleet owner investing in a reprogrammer or having dishonest fuel- ing station operators or employees offer to reset odometers for a small fee. The extent and impact of tampering might be reduced in three ways: • Audits to compare reported mileage and fuel purchased with typical fuel consumption averages. • More frequent or random OBU readings. • Substantial penalties. Manipulating Fuel Cards The gas station or fuel card options for collection may be subject to systematic evasion by drivers who falsify the odom- eter readings they enter or who switch cards between vehi- cles. For the evasion to persist, the driver (on his own or at the direction of his employer) would have to keep a second log book. Substantial underreporting would, however, appear as exceptionally poor fuel mileage (low miles per gallon) that could signal the need for verification or audit. To pass an audit, the driver or owner would have to alter the odometer reading on the truck itself. Cash Fuel Purchases If VMT payments are linked for reporting or auditing purposes to fuel purchases, there may be a particular prob- lem with drivers who pay cash, as the transaction would not link the quantity of fuel to the vehicle or its odometer read- ing. A truck operator could escape reporting or avoid the appearance of buying too much fuel for the reported VMT by periodically supplementing fuel card purchases with cash transactions. Most of the 160,000+ fuel retailers in the United States are also convenience stores that sell a variety of mer- chandise. Fuel cards and some bulk fueling systems can be made to refuse to dispense fuel without an acceptable odom- eter reading, but commercial fueling stations will probably have to continue to accept cash. Moreover, a side payment of half the applicable VMT fees in cash in return for the lack of a transaction record would leave both buyer and seller better off until caught. Unregistered Vehicles An unregistered vehicle would still need fuel, but it would not pay VMT fees. Some vehicle owners already avoid regis- tering their vehicles to avoid the registration fees or the costs of bringing their vehicles up to state standards for safety or emissions. There would be an increased incentive to avoid registering the vehicle or to file a false certificate of non- operation if by doing so the owner could save hundreds or thousands of dollars in VMT fees. There are already penalties for operating an unregistered vehicle, and there may have to be additional penalties for not paying VMT fees. Enforcement Options Three enforcement options are discussed below—linking VMT fees to registration, fuel card auditing, and direct VMT fee payment by vehicle drivers or owners. Linking VMT Fees to Registration To ensure that the owner of every applicable vehicle is paying VMT fees, each vehicle will need a separate, unique account that records payments and can be audited. Each vehi- cle already has such an account—the state vehicle registration

47 record. It should be a straightforward process to link the VIN to a record of VMT fee payments and flag anomalies for audit. It would also be possible for the tax-collecting authority to bill owners of non-paying vehicles a default minimum annual levy, plus a penalty for non-payment. Vehicles without a satisfac- tory payment record could be denied annual re-registration, or outstanding payments could be collected as part of re- registration (following the example of video rental stores). Vehicles that have paid suspiciously low VMT fees or no VMT fees at all could be flagged for odometer reading verification as part of the annual registration process, even in states (such as California) that do not have regular annual inspections. Own- ers who have underpaid their VMT fees could be required to pay the difference, plus a penalty, as part of the registration process. It would also be possible to send invoices for esti- mated VMT fees to owners of vehicles for which fees have not been paid or for which low fee payments suggest evasion. States already send out registration invoices, so the processes could be combined. The cost of adding VMT fee tracking to these registration databases, however, is unknown. Fuel Card Auditing Auditing and enforcement would be relatively simple for users of fuel cards. Fuel card providers can deliver reports with mileage entered, fuel purchases, and equivalent miles per gallon for each vehicle. The tax-collecting agency could request or obtain fuel usage data for any vehicle for which abnormally low fees were paid. Within limits, the accuracy of mileages entered at each fuel purchase would not matter much. A driver who underreported mileage on one occasion to reduce VMT fees would either eventually pay the fees when he reported a correct mileage later, or he would have to go on underreporting indefinitely. It would be possible to block fuel purchases for drivers entering implausibly low mileages or at least to require the driver to verify the mileage before pro- ceeding. An alternative would be to issue warnings on suspi- cious mileage entries and require official verification if under- payments continue. Requiring odometer readings would also make it more difficult for drivers to switch fuel cards between vehicles. If the VMT fee rate for a Class 4 truck is lower than the rate for a Class 6 truck, there would be an incentive to use the fuel card for the Class 4 truck to fill the Class 6 truck. While it is possible that some truck owners may be able to “game” the system in this way, significant evasion of this kind would become progressively more apparent the longer it persisted. Direct VMT Fee Payment by Vehicle Drivers or Owners A significant part of current fuel tax evasion is attributable to fuel dealers who may collect fuel taxes from customers but underpay them to the federal government. Direct VMT fee payment by vehicle drivers or owners may reduce this prob- lem, but at a much higher collection cost. Federal Collection Requirements Each collection option described above has a correspond- ing federal requirement for receiving and processing pay- ments. For VMT fees, the number of transactions will vary with the number of vehicles reporting, the frequency of reporting, and the degree to which owners report on a con- solidated fleet basis. Table 16 provides illustrative estimates of potential trans- action volumes for the various VMT fee alternatives. An aver- age fleet size of five vehicles is used for these estimates (based on an industry rule of thumb, as there are no solid data avail- able), as well as a working assumption that 50% of the truck population would use fleet-based reporting and payments. A higher average fleet size and greater prevalence of fleet-based reporting would reduce the transaction volume. The estimates in Table 16 range from a minimum of 2.5 mil- lion transactions for annual reporting for the smallest popu- Table 16. VMT transaction volume estimates.

48 lation with a fleet-based report option to 936.7 million for biweekly reporting by individual OBU-equipped vehicles in the broadest coverage. Relatively frequent reporting and payment may be desir- able to maintain steady federal cash flow and to avoid single large annual payments that could threaten truck owner finances. With the very large number of transactions shown in Table 16, one logical approach might be to adopt stag- gered quarterly billing. This would provide a steady federal revenue stream with one-third of the transactions required for monthly billing. Comparison of Distance/Vehicle VMT Collection Options The alternatives discussed above range from a mixed system of existing options (self-reporting, commercial OBUs, IFTA, and fuel cards) to nationwide deployment of new technology (mandatory VMT-compatible OBUs). These alternatives can be usefully considered as progressions in time and scope. The fastest, least costly, and least disruptive approach to implementing distance/vehicle VMT fees would be to establish a rate schedule and payment requirement and allow multiple methods for meeting the requirement. Truck fleet opera- tors who participate in IFTA or who use fuel cards could use information from those systems to calculate and pay VMT fees. Users of commercial telematics systems would likely be able to do the same. Owner-operators and operators of small fleets could use manual odometer readings or driver logs to self-report. As the VMT fee became a routine part of truck- ing operations and management, owners of fleets too large for efficient self-reporting would likely migrate to one of the other payment options. Fleet-by-fleet calculation and pay- ment of fees would be supplemented or replaced by third- party payment through IFTA, fuel card companies, telemat- ics providers, or independent firms. Under this scenario, eventual transition to telematics or OBU deployment would be motivated by industry needs and preferences rather than by mandate. It would be far more efficient for telematic or OBU applications to be consolidated at the fleet level so that, for example, AT&T would submit one fleet payment rather than 77,480 individual vehicle pay- ments (see Table 15). It is also possible to envision a VMT system with a target time horizon for electronic reporting and payment (5 years, for example). In that scenario, the federal government would cer- tify one or more OBU/EOBR systems as VMT-compliant, and the trucking industry would be required to purchase, install, and employ one of those systems for VMT fee payment. A mandatory OBU system could impose a significant finan- cial burden on truck fleet operators that did not find telemat- ics systems cost-effective for their own purposes. As explained in Appendix B, commercial telematics systems typically cost $500 or more per vehicle, creating a high threshold for small fleets for which communications and record-keeping ben- efits would be minimal. VMT compliance would likely add little to system cost, but the minimum price of a single- purpose VMT-compliant system could easily approach that $500 figure. The marginal value of a nationwide OBU system for trucks appears small when there are multiple, less costly options for fleet-based reporting and payment. The substantial public- sector investment in OBU communications and processing would likewise appear less cost-effective than submission of electronic records and payments from a small number of third-party reporters and individual reports from small fleets or owner-operators. These comparisons could change substantially if a national VMT system were implemented for all vehicles. With the required communications, processing, and auditing/ enforcement system set up for the roughly 240 million private passenger vehicles, the incremental cost of accommodating 10 million medium-duty and heavy-duty trucks would be far less than in a truck-only system. The unit cost of dedicated, VMT-compliant OBUs would also likely decline. Costs and Revenues VMT Fee Structure As of early 2001, there had been no concrete, detailed pro- posals developed for either distance/vehicle or time/location VMT fees. The studies reviewed for this project have included conceptual amounts, such as $0.01 per mile, to illustrate the revenue potential or as a basis for surveying public opinion. Pilot projects have been intentionally “revenue neutral” so that participants pay the same amount in VMT fees that they would have paid in fuel taxes. “Flat” Distance-Only VMT Fee The literature reviewed to date contains little analysis of the VMT rates or rate structure that would be required to meet various funding targets. Table 17 provides a starting point by estimating the “flat” distance-based VMT fee that would be required to generate $5 billion in gross fee revenue from the roughly 227 billion truck VMT in 2008. The cor- responding VMT fee would be about $0.022 per mile. Table 17 applies that flat, revenue-neutral VMT fee to a range of vehicle types and compares the outcome with estimated cur- rent federal fuel tax payments. As Table 17 indicates, medium-duty trucks traveling 12,000 miles annually would pay about $264 in VMT fees in addition to $344–371 in estimated fuel taxes, and a Class 8

49 diesel tractor running 80,000 miles annually in long-haul service would pay about $1,759 in VMT fees in addition to $3,425 in estimated fuel taxes. Vehicle Weight Adjustments To account for vehicle weight, a VMT system would have to vary by weight class or operating weight. The literature on VMT fees has thus far been concerned almost exclusively with passenger cars and light trucks used for personal trans- portation. The gross operating weight of passenger cars and light trucks used in personal transportation ordinarily varies from about 3,000 to 7,000 lb. In that relatively narrow weight range, it may be a reasonable policy to charge all such vehicles the same VMT rate under the implicit assumption that they all have nearly the same infrastructure impact per mile trav- eled. That assumption breaks down, however, when applied to medium- and heavy-duty trucks, whose gross operating weight can range from 10,000 lb (the minimum for Class 3) to 105,000 lb (the maximum for Class 8 in some states). While the impact of weight is controversial, there is no doubt that a loaded 105,000-lb tractor and semi-trailer has a greater impact than a 10,000-lb delivery van. To reflect user impacts, a VMT system would have to sub- divide the national truck fleet into groups with roughly simi- lar highway impacts and assign an appropriate fee scale. There are a number of vehicle classification systems in use. The most common is the GVW system shown in Table 18. Table 19 takes a simplified approach to an illustrative weight- indexed VMT structure. For Classes 4–6 and for Classes 7–8, a “vehicle weight index” is based on the ratio of the weight range midpoint. The “indexed $/VMT rate” is then set to distribute a $5 billion gross revenue goal between the two vehicle groups according to estimated VMT in each group. The approach in Table 19 effectively assumes that infrastructure impact is a lin- ear function of vehicle weight, a fairly conservative approach. Compared to the annual “flat” fee estimates in Table 17, this fee structure would, of course, shift some of the burden from medium-duty trucks to heavy-duty trucks. State Weight-Distance Tax Systems Four states—Oregon, New York, New Mexico, and Kentucky—have “weight-distance” or “ton-mile” taxes. A number of other states have had ton-mile taxes or some varia- tion in the past, but have discontinued them due to collection expense, compliance costs imposed on carriers, legal challenges, and concerns over the impact on state economic development and competitiveness. For the states that have weight-distance taxes, implementing them has required coping with the same weight-related issues that a national distance/vehicle VMT fee faces. GVW Class and Fuel Type GVW Limit Est. Annual VMT* Est. MPG** Est. Annual Gallons Fuel Tax $/Gal Fed Fuel Taxes "Flat" $/VMT Rate Annual "Flat" VMT Fee Diesel 4 16,000 12,000 8.5 1,412 0.244 344$ 0.022 264$ 5 19,500 12,000 7.9 1,519 0.244 371$ 0.022 264$ 6 26,000 15,000 7.0 2,143 0.244 523$ 0.022 330$ 7 33,000 40,000 6.4 6,250 0.244 1,525$ 0.022 879$ 8 33,000 80,000 5.7 14,035 0.244 3,425$ 0.022 1,759$ * Rough estimate - needs better data. ** 2002 Data. Table 17. “Flat” VMT fee example. Table 18. Truck classifications.

50 Oregon’s “weight-mile” tax replaces the state fuel tax for vehicles over 26,000 lb (nominally GVW Classes 7 and 8). Oregon defines a “declared weight” as follows (Oregon Department of Transportation, October 2008): Weight-mile tax is paid on a declared weight basis. The TAX DECLARED WEIGHT is the weight of the vehicle and the load that represents the heaviest weight the vehicle will actually carry in a given configuration for the reporting period. Table 20 shows the applicable tax rates for vehicles with declared weights of 26,001 to 80,000 lb. The rate schedule is not linear—rates rise with weight at an additional .0024 cents for each additional 2,000 lb at the lowest level and rise faster at higher weights. Note that the Oregon tax system does not attempt to ascertain the actual operating weight of the vehicle and its load. New York’s tax (see Table 21) is imposed in addition to state fuel taxes and at lower rates than Oregon’s tax. The New York rates are also more complex, with different rates for laden and unladen operations. The distinction is critical because New York is faced with the task of attempting to tax many trucks that are only in the state for a short dis- tance due to its geographic position among the northeast states. New Mexico’s taxes are much lower and have not been increased in over 20 years. The numbers in Table 22 are in mills, each mill being 1/10 of a cent or 1/1000 of a dollar. Kentucky’s rate is the simplest: $.0285 per mile for trucks over 59,999 lb. Table 23 uses the inherent relationship in the Oregon weight-mile and fuel tax structure to develop a third pos- sible VMT fee scenario. The fuel-tax-equivalent VMT fee for Classes 1–6 and the Oregon weight-mile tax for Classes 7–8 (based on the same midpoint weights as in Table 19) yield a set of annual VMT fee estimates. Those estimates closely resemble existing federal rate equivalents for Classes 1–6 (which is to be expected since the Oregon state fuel tax rates are similar to the federal rates). For larger trucks, the “Ore- gon” rates would be 39% higher for Class 7 and 110% higher for Class 8. Table 19. Example of weight-indexed VMT cost structure. Table 20. Oregon weight-mile tax rates.

51 Source: New York State Department of Taxation and Finance, November 2009. Table 21. New York state highway use tax table. Table 22. New Mexico state weight-distance tax tables.

52 Passenger-Car Equivalents Ideally, VMT fees for trucks or other vehicles would be based on passenger car equivalents (PCEs). PCEs measure the impact of other vehicles on infrastructure and conges- tion compared to passenger cars. Nominal PCEs for trucks are around 2.0 because, while a truck may be much longer than an automobile, a truck does not require a multiple of the intervehicle spacing. As trucks get larger and heavier, the PCE increases with the stopping distance, because more space must be allowed between vehicles. The use of PCEs to scale VMT fees would require a stan- dardized PCE methodology. Although there are suggested practices for estimating PCEs, they typically vary with ter- rain and traffic levels. In heavy traffic or on steep grades, the effective PCE of a tractor-trailer may be 4.0 or higher. Table 24 provides examples of truck PCEs from FHWA’s HPMS Field Manual (Office of Highway Policy Informa- tion, May 2005). The literature review located no discussions of the use of PCEs for truck VMT fees. It would appear that substantial technical work would be required to construct a standardized PCE methodology for that application. Comparison of Potential Fee Structures The various distance/vehicle fee structures postulated above would have somewhat different impacts on trucks. Table 25 displays the possible VMT fee rates by vehicle class and the resulting annual fee totals for the annual mileages used in previous tables. Note that Table 25 does not attempt to take other federal or state fees or taxes into account. The major difference is between the “flat” fee option and the two weight-adjusted fee options; the two different meth- ods of adjusting for weight yield similar results. The cost impact to vehicle operators would depend on their annual mileage (as expected for a VMT fee). The largest cost impacts would affect Class 8 trucks and tractors. Con-way Inc. reportedly operates around 12,000 Class 8 tractors. If each one averages about 50,000 annual miles, as assumed in the analysis, the most aggressive weight- indexed VMT fee schedule would cost the company around $22 million annually (Transport Topics Top 100 For-Hire Carriers—2011, 2011, p. 4). At the other extreme, a local owner-operator who used a Class 8 tractor for a local sand and gravel business might only travel 20,000 miles annually. Under the less aggressive “Oregon” scenario, that owner- operator would see an annual increase of $540, or about $45 per month. While that amount seems less impressive than the J. B. Hunt example, it would be an unwelcome burden to a local owner-operator. Source: Tioga Group analysis. Table 23. Graduated VMT fees based on Oregon rates. Two-Way Flow Rates (passenger cars/hr) Level Rolling Mountainous 0-600 1.7 2.5 7.2 >600-1,200 1.2 1.9 7.2 >1,200 1.1 1.5 7.2 Type of Terrain Source: Office of Highway Policy Information, May 2005. Table 24. Examples of PCE factors for trucks. GVW Class and Fuel Type Fuel Tax Equivalent $/VMT Annual Fuel Tax "Flat" $/VMT Rate "Flat" Annual VMT Fee Wt. Indexed $/VMT Rate Annual Wt. Indexed VMT Fee "Oregon" Eqv. VMT Fee Annual "Oregon" VMT Fee Diesel Class 4 0.029 344$ 0.022 264$ Class 5 0.031 371$ 0.022 264$ Class 6 0.035 523$ 0.022 330$ Class 7 0.038 1,525$ 0.022 879$ Class 8 0.043 3,425$ 0.022 1,759$ 0.013 0.027 159$ 1,770$ 0.011 133$ 0.029 1,848$ Source: Tioga Group analysis. Table 25. Potential distance/vehicle fee structures.

53 Time/Location Congestion Fee Structure The research team did not investigate or analyze possible time/location fee amounts above those involved in distance/ vehicle fees. The additional fees for TDM measures would almost certainly be set by local, regional, or state toll authori- ties. None of the studies or discussions to date envision a national TDM program. Revenue Potential Revenue potential varies with VMT fee rates and the num- ber of vehicles affected. Table 26 shows the relationship for an annual gross revenue target of $5 billion. The fee rates would be the same for $5 billion gross rev- enue whether collected via mandatory OBUs or through some combination of reporting and payment options. The required average VMT fee would decline from $0.048 per mile, if applied to just Classes 7 and 8 freight trucks, to a low of $0.022 per mile for all Class 4–8 vehicles. The actual rate should vary by GVW class or some other means to reflect size and weight differences. Federal Collection and Enforcement Costs Many studies and reports express collection costs as a per- centage of revenue received. While this expression of revenue received is revealing, it is of limited value since collection costs will remain roughly the same if tax or fee rates change. VMT fee collection costs would be a function of the num- ber of vehicles, collection points, and transactions, and col- lection costs would not be expected to vary significantly with the actual fee rate. The exception would be if credit cards are accepted as payment by the collection agency itself, whether on a transaction-by-transaction basis or as the basis for a pre- paid account (similar to toll card accounts). At present, credit card purchases of fuel result in credit card fees payable by the retailer. The collection agency (IRS) actually collects the taxes from the wholesalers, who do not pay credit card fees. If a federal collection agency accepts credit cards or deducts from prepaid accounts linked to credit cards, the federal col- lection agency will incur the credit card fees. Credit card fees typically range from 2 to 4%, although a federal agency might be able to negotiate a favorable rate. Most VMT fee proposals are “pay-as-you go,” mean- ing that fees are collected at frequent intervals rather than monthly or annually. This feature, however, would entail a very large number of transactions. Weekly transactions for 10 million medium- and heavy-duty trucks would yield over 500 million annual transactions. At $0.10 per transaction, that annual cost would be $52 million. This is a small number compared to the totals handled by credit card or telephone systems, but it is a huge increase over the present fuel tax system, which involves roughly 24,000 annual transactions. Annual collection costs for a distance/vehicle VMT fee are variously estimated to range from about $10 to $100 per vehi- cle. These figures are mostly informal “guesstimates” that are not backed by empirical data or formal estimate techniques. The most appropriate estimate obtained by the research team is a figure of $35 per vehicle for the administrative and collec- tion costs of a state E-ZPass system. That estimate is used in later cost comparisons of revenue options. Higher estimates were obtained from the proposed Dutch VMT fee system. Bidders on that system estimated about $75 per vehicle for annual costs without depreciation and $100 per vehicle with depreciation. The highest estimate was derived from data and announcements of the German Autobahn truck fee collec- tion system, which had average annual collection and over- head costs of $804 per vehicle. The cost of enforcement will depend on the incentives and opportunity for evasion; the cost of auditing to detect evasion; and the costs of investigation, prosecution, fine assessment, or other measures. These factors will vary depending on how VMT fees are implemented. In the Oregon pilot project and proposed implementation, for example, vehicles that did not Table 26. VMT fee rates for $5 billion gross revenue.

54 or could not pay VMT fees would be charged fuel taxes as a default, so the incentive for evading VMT fees would be small. The research team was unable to locate any estimates of enforcement costs for VMT fees, in part because enforcement costs are usually included in administrative cost estimates. Information on the enforcement costs for IFTA reporting or for state weight-distance taxes is not readily available and therefore provides little guidance. Enforcement costs are therefore an unknown and a can- didate for further research. The range of possible evasion methods and the evasion incentives for truck fleet operators, however, suggest that substantial resources may have to be devoted to enforcement activities. The average collection cost per vehicle shown in Table 27 is based on collection and enforcement cost estimates of $35 per vehicle, plus an amount to reflect the use of credit cards and the fees charged by credit card companies. The estimate assumes that half the payments are made by credit card and that the aver- age credit card fee is 2%. The result is to add $50 million to each collection cost estimate for VMT fees. The average fee per vehicle thus includes the $35 base amount plus a portion of the credit card fee that declines with the number of affected vehicles. Implementation and Compliance The federal implementation cost estimates for VMT fees draw on the Oregon pilot project estimates. In this case, the options differ in the level of industry implementation cost but are identical in federal data infrastructure requirements. According to the Convenience Store Association, there are over 160,000 retail fuel sellers in the United States. To col- lect VMT fees at these locations, they would all have to be equipped with wireless OBU readers and appropriate com- munications equipment. The communications equipment could probably be shared with lines used for credit card transactions. The cost of equipping fuel dealers with OBU readers, however, would be substantial. The Oregon pilot project estimated a capital cost of roughly $17,556 per sta- tion in 2007 dollars or about $18,300 in 2010 dollars. For the 160,000+ locations in the United States, the capital cost would be roughly $3 billion. Table 28 provides estimates of federal government and freight industry implementation costs. Federal and industry implementation costs are converted to annual equivalents. The industry costs are estimated at $500 per vehicle, which covers OBU/EOBR equipment cost and installation. The OBU/industry options version assumes that half the affected trucks use OBUs/EOBRs at $500 each, yielding an average of $250 per vehicle. The mandatory OBU option assumes that the entire affected fleet would be equipped. Of the roughly 9 million Class 4–8 trucks in the United States, about 6.7 million were in fleets of four vehicles or more in 2005. (Bureau of Transportation Statistics, 2010, Table 1-14) Based on a rough average fleet size of 100 vehicles for fleets of 25 or more and 10 vehicles for fleets of 4 to 24, there were roughly 300,000 fleets that could report on a consoli- dated basis. At an estimated annual cost of $300 each, their annual compliance cost would be $90 million. The remain- ing 3.3 million trucks are either operated singly or in fleets of less than four vehicles. If reporting individually at $300 each ($25 per month), their annual compliance cost would be $990 million. The annual total for the 10 million trucks would be $1.08 billion, or an average of $108 per truck. VMT fee compliance costs are therefore estimated in Table 29 at $108 per vehicle per year for the estimated 2008 truck fleet. This is a very rough estimate, and should be supplanted by future research. Note, however, that economies of scale for report- ing and compliance are dramatic and would favor larger fleets. Net Revenue and Revenue Efficiency Table 30 provides estimates of total net annual federal rev- enue from VMT fees ($5 billion less collection and annual- ized implementation costs) and total annual industry cost Table 27. VMT fee collection and enforcement costs.

Table 28. VMT fee implementation cost estimates.

56 ($5 billion plus annualized implementation and compliance costs) and shows these costs broken down to net annual rev- enue per vehicle and annual cost per vehicle. Table 30 also shows the ratio of net federal revenue to annual industry cost. As with the fuel tax surcharge options, the more technology- dependent options have lower efficiency ratios due to their higher implementation and compliance costs. Privacy Concerns The proposed approaches to assessing VMT fees typically involve using OBUs to record, store, and transmit road use data. These proposals have raised privacy concerns regard- ing the government’s ability to track the movements of indi- viduals. Whether these concerns are legitimate as a practi- cal matter depends on the nature and specificity of the data recorded. There are numerous potential VMT fee system designs, ranging from systems that collect no information about the movements of individuals to systems that collect detailed GPS data about every movement a vehicle makes. As a legal matter, it appears likely that virtually any system design would comply with existing constitutional and statu- tory privacy protections. The Fourth Amendment to the U.S. Constitution protects individuals’ “reasonable expectations of privacy.” However, the U.S. Supreme Court has found elec- tronic monitoring of vehicle movements to be consistent with the Fourth Amendment because, according to the Court, an individual cannot have a reasonable expectation of privacy in his or her movements in public. These precedents suggest that collecting information regarding vehicle movements via a VMT fee system would not violate the Fourth Amendment. One potential exception is a system that provides the govern- ment with detailed GPS route data. Federal courts have upheld the use of limited GPS vehicle tracking based on the Supreme Court precedents permitting electronic vehicle monitoring. However, some courts have indicated that they might view continuous or widespread detailed GPS tracking as constitu- tionally distinguishable from the type of limited and relatively imprecise tracking approved by the Supreme Court. The recent (2011) uproar regarding revelations that Apple iPhones were tracking and retaining user location data reveals the depth of feeling over such privacy issues. Apple was promptly sued, and several governments began immediate inquiries. Apple was effectively forced to modify the iPhone operating system to eliminate retention of location data. To avoid potential constitutional challenges, a road user fee system should therefore be designed so that it does not pro- vide detailed route information to the government or third- party collector of the data. The “thick-client” system meets this criterion. The above conclusions do not change if the government uses a private third party to collect the data, since a third party collecting the data on behalf of the government will be subject to the same constitutional standards. However, if a third-party collector were involved, it would be critical for the authorizing statute to make the third party subject to strict data security and protection obligations of the same sort that would apply were the government collecting the data itself. The privacy and data confidentiality issue may be easier to manage in the commercial trucking sector. Fleet operators assume that they have a right to know what their vehicles are doing at all times, including vehicles operated under contract by subhaulers. The widespread use of telematics and EOBR event recording and location systems has already made much of the trucking industry comfortable with third-party han- dling of vehicle activity data. Existing safeguards and proto- cols appear to be satisfactory. Fleet operators engaged in inter- state operations and those in states with weight-distance taxes already track and report fleet miles traveled by state, building that data up from electronic or manual vehicle records. There are no statutory privacy protections that would pro- hibit a VMT fee system. There are federal statutory provi- sions regarding the use of “tracking devices.” However, it is not clear whether these provisions would apply to most VMT fee systems. Even if the provisions did apply, they would not Table 29. VMT fee compliance cost estimates.

Table 30. VMT fee net revenue and efficiency.

58 impose any standards beyond what is required by the Fourth Amendment. There are also some state laws prohibiting the use of tracking devices. To the extent that these laws would conflict with a federally mandated VMT system, they pre- sumably would be preempted. A more detailed discussion of these privacy issues prepared by the research team’s consult- ing legal counsel is attached as Appendix D. Behavior Incentives It has been argued that a direct measure of highway impact (a VMT fee) will create better incentives for efficient routing (i.e., minimizing VMT) than a fuel tax. This may be less impor- tant for the medium- and heavy-duty truck sector than for passenger vehicles. Truck fleet operators already have strong incentives to minimize miles and hours, and many do so aggres- sively. A VMT fee roughly equivalent to the current fuel tax would change truck operating costs by no more than 2%. Such a small change is unlikely to change operations or driver behavior. A VMT fee is generally expected to be a stronger incentive for efficiency than the fuel tax because it relates directly to the decision of whether or not to make a given trip and what route to take on that trip. Freight movements are a func- tion of customer demand. Freight carriers choose whether or not to take a given piece of business and make the trips associated with that business. Given the close attention freight carriers pay to cost and the small part that fuel taxes or VMT fees would play in that cost, it seems unlikely that the difference between a fuel tax and a VMT tax would be decisive. Time/location VMT fees that incorporate TDM or tolling may have greater behavioral impacts than distance/vehicle VMT fees, but the behavioral response of freight carriers and service providers is still constricted by demand. Both freight carriers and service providers may be able to choose whether or not to serve a particular time, location, and route combi- nation, but ultimately the customer needs to have the freight delivered and the service provided. It seems likely that carriers and service providers would seek to pass along congestion fees, tolls, or other similar expenses in the form of price increases or surcharges. Their price increases or surcharges would then become incentives for the customer to reduce shipments or receipts, to reschedule, or to relocate. The PierPass program at the Ports of Los Angeles and Long Beach is essentially a congestion fee program and has resulted in up to 40% of the applicable business being handled in off- peak hours. Much of the shift, however, resulted from cus- tomer choices rather than trucker choices. The greatest long-term impact of time/location VMT fees may therefore be manifested in the ways and places in which business is done. Initially, time/location VMT fees would encourage freight-dependent and service-dependent busi- nesses to relocate outside the fee zones and off toll routes. These incentives, however, are primarily functions of con- gestion pricing, road pricing, or tolling features and are not attributable to the VMT fees per se. Implementation Legislation and regulations to implement such a VMT system in place of fuel taxes would have to address the following issues: • The scale of VMT charges for increasing vehicle weights. • The fee indexing method. • Dedication of VMT revenue to highway uses. • The method and options for fee collection and compliance. • Penalties for evasion and means of enforcement. • Provisions for international trucks. • Requirements, if any, for vehicle manufacturers to install OBUs as OEM equipment. • Technical specifications of VMT-compliant OBUs, or a means for developing such specifications. • Privacy safeguards. • Implementation timeline and budget. Faced with these kinds of challenges, most government and industry stakeholders contacted in the course of this study have responded with some variation on “Why don’t we just fix the fuel tax?” Implementation of VMT systems in the freight and ser- vice trucking sectors involves key policy issues that have been only peripherally addressed in the literature to date. To date, almost all the research and study has been focused on the VMT fee concept, the collection technology, and privacy issues. Passenger-focused VMT fee discussions have usually assumed or concluded that each vehicle would have to be equipped with an OBU capable of automatically recording and reporting VMT. This viewpoint is based in turn on the assumption or conclusion that passenger-vehicle operators do not track and report their mileage by other means and could not do so efficiently. There is a need for additional research on methods for varying distance/vehicle VMT fees by vehicle type. There are many different possibilities besides the examples pre- sented above. As noted above, there are two relationships at issue: • The relationship between vehicle type (however classified) and actual or average operating weight. • The relationship between actual or average operating weight and infrastructure requirements and costs. Additional research would be required if an attempt is made to incorporate axle weights or PCEs in the methodology. Table 31 shows the research team’s working assumptions regarding the implementation phase-in time and pass-through

Revenue Option Years to First Revenue Federal Implementation Cost Phase-In Years Tax Burden Passed to Customers First Year Tax Burden Passed to Customers Second Year VMT Distance/ Vehicle Fee - OBU/Options Class 7&8 freight 5 5 33% 66% Class 7&8 all types 5 5 33% 66% Class 4-8 freight 5 5 33% 66% Class 4-8 all types 5 5 33% 66% VMT Distance/ Vehicle Fee - OBU Only Class 7&8 freight 5 5 33% 66% Class 7&8 all types 5 5 33% 66% Class 4-8 freight 5 5 33% 66% Class 4-8 all types 5 5 33% 66% Tax Burden Passed to Customers Third Year Implementation and Compliance Cost Passed to Implementation and Compliance Cost Passed to Customers Second Year Implementation and Compliance Cost Passed to 100% 33% 66% 100% 100% 33% 66% 100% 100% 33% 66% 100% 100% 33% 66% 100% 100% 33% 66% 100% 100% 33% 66% 100% 100% 33% 66% 100% 100% 33% 66% 100% Third Year Customers First Year Customers Table 31. VMT fee implementation phase-in schedule.

60 of tax burden and compliance costs for VMT fee options. The implementation period is assumed to be 5 years due to the need to develop a system for fee collection, auditing, and enforce- ment and to develop and deploy the necessary technology. The federal government would incur costs, but receive no revenue in this 5-year period. For purposes of modeling economic impacts, the research team assumed that about one-third of the industry tax burden and compliance costs would be passed on in the first post- implementation year, two-thirds in the second year, and all by the third year. This 5-year implementation timeline for the medium- and heavy-duty trucking sector is much shorter than the timelines estimated in the literature for the passenger-vehicle fleet. Advantages The key advantages of VMT fees include a strong link- age between road use and fee structure, meeting the policy preference for user fees as opposed to taxes. VMT fees would facilitate coverage of hybrid, alternative fuel, and electric vehicles that would otherwise make little or no contribution under a fuel tax regime. VMT fees are user fees to the degree that use equates to vehicle miles traveled. VMT fees could differentiate between vehicle types and classes and could thus approximate vehicle tare weight, but would not reflect the net weight of freight or other vehicle loads. VMT fee systems also hold out the possibility for linkage between federal, state, regional, or local programs, particularly where time/location VMT systems or other means can dis- tinguish miles operated by jurisdiction. If cost-effective time/ location capabilities can be added to a distance/vehicle VMT system, a multipurpose system could emerge for both collect- ing revenue and implementing TDM measures. Such a sys- tem could also accumulate data on highway activity for plan- ning purposes. The potential for simultaneously establishing national interoperability standards for tolling systems is also an attraction. National freight truck VMT is expected to continue to grow while fuel consumption plateaus or grows more slowly due to improved efficiency. VMT fees would still require indexing, however, to retain buying power in the face of inflation. A freight or truck VMT fee system could also be viewed as a pilot for eventual replacement of fuel taxes with VMT fees for all vehicles. Disadvantages Drawbacks to a distance/vehicle VMT system include com- plexity and high implementation, collection, compliance, and enforcement costs. There would be substantially greater opportunities for evasion relative to fuel taxes. Another drawback of VMT fee proposals is the very long implementation time. An AASHTO report, for example, anticipates a 20-year timeline to allow for additional studies, pilot projects, technology advances, legislative and regulatory steps, and eventual implementation with a phase-in period (AASHTO, September 2007, p. 13). Time/location VMT fees would have substantially higher implementation costs than distance/vehicle fees and raise privacy issues that probably can be resolved legally and technically, but may still affect political feasibility. The cost/ benefit balance of time/location VMT systems is open to question, as they would benefit only a small portion of the vehicle population, and the advantages of these systems over existing toll systems are unclear. VMT concepts would be difficult to apply to non-highway modes. The incremental behavioral incentives envisioned for VMT fees in the passenger sector may not be as strong in the trucking sector, where mileage is already often tracked and minimized. The revenue potential and user equity features of VMT fees would be diluted to the extent that government vehicles are exempted. Exemptions would also complicate the collection system. The public acceptance barriers to a distance/vehicle VMT system are substantial. The concept can, with reason, be char- acterized as more costly and invasive than fuel taxes with few compensating benefits.