RAND/TRB plan for AASHTO's VMT charges ignores toll technologies and expertise
By Peter Samuel
AASHTO, the state DOTs lobby in Washington DC has a report from the Academies of Sciences' Transportation Research Board (TRB) that determinedly avoids toll technologies and expertise in the search for a quick fix to declining gas tax revenues. Written by a RAND Corp led team the report lists three options for estimating vehicle-miles traveled (VMT) charges:
- derived from fuel consumption and an assigned miles per gallon and levied at the gasoline pump (MPG)
- on board diagnostics (OBD2) speed data and cellular communications as proposed by Max Donath (Donath)
- a GPS unit in every vehicle (GPS)
The report (RAND report) written at the behest of AASHTO urges federal funding for the planning, analysis, research and development on these three approaches to allow one to be used to collect VMT charges nationwide from 2015. It doesn't provide any costing.
It proposes a 1.1c/mile VMT charge on all vehicles (or 0.8c/mile cars, 3.4c/mile trucks) as the revenue-neutral equivalent of the current gasoline/diesel fuel tax (gas tax) to garner about $36b in 2015. Attraction for AASHTO is that the principal revenue stream would automatically grow with VMT whereas fuel tax revenues have been growing at less than half that rate due to improvements in vehicle mileage per gallon. The study was prompted by the prospect of stagnation of fuel tax revenues based on further improvements in mileage, and the introduction of electric cars, together with legislators reluctance to hike the gas tax.
In six years VMT revenues would rise about 10% to close to $40b whereas the gas tax lingers around $36b, barely growing at all. (see graph "National VMT Fee Revenue Forecasts" nearby)
Earlier this year the gas tax funded Highway Trust Fund went insolvent, and was bailed out with a one-time grant of $8b from the US Treasury.
The RAND team describes the three methodologies on which it urges development work:
(1) MPG: Vehicles would be equipped with an RFID transponder perhaps embedded in the license plate or registration sticker. At a gas station to purchase fuel, readers at the gas pump would read the vehicle ID, and determine the vehicle's fuel-economy rating based on the make and model. The
VMT charge would be based on gallons of fuel purchase and expected miles per gallon. The VMT could then be added to the fuel purchase price, while fuel taxes (already paid by the wholesaler level and therefore built into the retail price) would be subtracted. Vehicles not yet equipped with a transponder to ID mileages/gallon for VMT charges (including foreign vehicles) would continue to pay existing fuel taxes.
(2) Donath: speed data from the On Board Diagnostics II (OBD2) port on all cars manufactured after 1996 would be used to derive estimates of VMT in an on board unit (OBU) and cellular wireless signals would be used to derive location allowing revenues to be allocated by geographic area, and to remit data for billing of VMT. GPS units could be installed in pre-1996 vehicles.
(3) GPS: a "coarse resolution" GPS in every vehicle to measure VMT supplemented with RFID gantries at strategic points and a connection to the OBD2 data port as backup to poor GPS data.
COMMENT: This report is a sad commentary on the incompetence of TRB and its subservience to DC lobbyists. AASHTO which has a vested interest in federal tax and grant policies was allowed to dictate a policy objective to TRB - raising money from motorists according to vehicle miles traveled. The report offers no justification for this crude money raising exercise except that AASHTO considers it important because legislators won't raise the gas tax. TRB has no business being a research arm for a lobby group.
The report is incompetent because it ignores the experience and expertise of the toll industry which collects about $6 billion a year electronically from on board units (transponders) in some 25 million vehicles and which has decades of experience in all facets of direct road use charges.
The people on the panel don't know tolling. And this ignorance shows throughout the report.
The various approaches favored in the report especially the Donath and GPS are worth more research. But they need a lot of development, and they need serious benefit/cost analysis.
The RAND report is unprofessional in its complete neglect to mention video tolling and automatic license plate recognition, though London's pricing is based on this and it plays an important secondary role in scores of toll and pricing projects.
Unforgivable is the RAND report's cavalier dismissal of the practicality of electronic tolling technology - RFID. The whole array of RFID from sticker tags through E-ZPass, FasTrak, and 5.9GHz OmniAir are airily dismissed in one page (p70) out of 150 pages by claiming that (1) RFID could not cover the entire road network and (2) cost of readers and gantries "would be considerable."
The naivety and emptiness of the RAND report in this respect is staggering.
Why must the entire road network be covered? Why collect charges on local streets and collectors? Developers built those roads and city/county property taxes maintain them. The VMT traveled on them is small anyway.
It was apparently beyond the capability of the RAND team to produce any estimates of costs of their schemes beyond declaring RFID costs "considerable", so here are some back of the envelop numbers. Most overhead costs including accounts maintenance, vehicle classification, and enforcement are common to all, so we omit those.
RAND propose expensive OBUs. At $25 to $100 each in 200m vehicles their cost of OBUs is a huge $5b to $20b. The gas station based approach using the cheaper OBUs involves a large infrastructure of readers at gas stations, the other with more expensive OBUs involves a major use of the cellular wireless network. Either way all three RAND proposals look like many billions of dollars of outlay to set up.
The toll industry's RFID that is so blithely dismissed by RAND has "OBUs" now going for below $2 each - Georgia DOT's price on ISO 18000 6C sticker tags (see http://www.tollroadsnews.com/node/4358) for I-85 HOT lanes. So toll industry 6C OBUs for 200m vehicles are probably $300m to $400m.
Readers for 6Cs are $3,000 each.
The US has 260k miles of rural interstate, rural principal arterials and rural minor arterials. One reader per 10 miles is 26k readers @$3k each or about $80m. In urban areas you probably need a reader every mile so with 178k of interstates, and principal and minor arterials your cost of readers is 178k x $3k or about $540m.
see road mileages http://www.fhwa.dot.gov/policyinformation/statistics/hm20_summary.cfm
Total reader cost to cover the country's interstates and arterials is $600m or so and the OBUs $300m to $400m.
Add in readers on 650k miles of rural collectors at one per ten miles and you add 65k x $3k and $195m. Urban collectors total 110k so the cost of a reader per mile is $330m.
RFID is about a billion without coverage of collectors and $1.5b covering the collectors as well. And that's with proven technology, that is in use and has high accuracy.
Versus the RAND's three half baked, ill-defined schemes with lightly tested technologies that run into a cost range $6b to maybe $20b+.
RFID has the further advantage that the gantries provide the infrastructure for a dense coverage of cameras for backup and enforcement, features the RAND schemes lack. And RFID readily allows different charges to be levied on different roads at different times of day. Such congestion pricing can only be accomplished with great difficulty with the RANDian schemes.
Here's the clunker of a report: