Per-Mile Choice Pilot

The problem

The federal fuel tax is in its final decade as the foundation of surface-transportation finance. Under the Congressional Budget Office’s February 2026 baseline, the Highway Trust Fund’s Mass Transit Account is exhausted in fiscal year 2027 and its Highway Account in fiscal year 2028, with a cumulative shortfall of approximately $295 billion by 2036 — against a gasoline tax unchanged at 18.4 cents since 1993, which has lost roughly 39 percent of its purchasing power (Institute on Taxation and Economic Policy), levied on a fleet at a record 27.1 miles per gallon (Environmental Protection Agency) that burns less fuel for every mile it drives. A replacement mechanism will be chosen in the next decade. The only question is on what evidence.

At present, the American evidence base for the leading replacement — charging for road use by the mile — is a patchwork. Congress has funded state experiments through the Surface Transportation System Funding Alternatives program, and thirteen states and two multistate coalitions have piloted variants since fiscal year 2016 (Government Accountability Office).

Each program is small, state-specific, and differently designed; none operates at national scale; and public support for mileage fees in the abstract is low — a mean of 24 percent across national survey questions (NCHRP synthesis). A universal system legislated on extrapolation from that record would rest its largest fiscal decision in a generation on the thinnest evidence in the program. The problem is therefore twofold: the revenue instrument is expiring, and the knowledge required to replace it responsibly does not yet exist at the scale the decision demands.

24 percentMean support across national survey questions

Source: NCHRP synthesis of public opinion on mileage fees (Agrawal, Mineta Transportation Institute).

The proposed solution

Test the replacement before adopting it. For five years, volunteers anywhere in the country could pay for their roads by the mile instead of by the gallon, in a single national trial built to produce the evidence a permanent decision needs: whether the technology works at scale, what it costs to run, and whether the public accepts it when they have actually lived with it.

Every volunteer chooses how their miles are counted. The default counts miles and nothing more: a simple odometer method that records a vehicle’s total distance and no location information of any kind, charged at one flat national rate set to match what the average driver now pays in fuel tax. The alternative exists only for those who actively choose it: satellite-based reporting that can vary the price — less for miles driven on designated rural roads, more for miles driven in heavily congested city areas at peak times — so the trial can learn whether pricing by time and place is workable, and whether drivers accept it, before any legislature is asked to impose it.

Taking part must cost nothing. Participants are held harmless: every dollar of federal fuel tax they pay while in the trial is credited or refunded, and each receives a yearly payment of between $100 and $400 for the data their participation generates. The group must look like the country — city, suburban, and rural drivers, across income levels, driving gasoline, hybrid, and electric vehicles — because a trial whose volunteers cannot speak for the national fleet cannot support a national decision. Money raised by the peak-time surcharge is returned to the transport agency of the area where it was collected, to spend on alternatives to driving alone, so the congestion charge is tested as a local instrument from its first day.

Privacy is part of the design, not a promise in a leaflet. The trial may collect only the minimum data needed to run it. Nothing collected may be used for law enforcement, or for any purpose beyond the trial itself, without a warrant from a court. The standards for anonymizing, encrypting, and storing the data are set through public rulemaking, where they can be examined before they are relied on. And the trial cannot end in a drawer: when it concludes, the results must be converted into a report to Congress with a recommended rate structure that keeps the Highway Trust Fund solvent over the long term, and a phased plan for moving the whole country to per-mile charging and repealing the federal gasoline and diesel taxes.

Test the replacement before adopting it.

The legislative proposal

Subtitle B is the “American VMT Choice Pilot Program Act” (Sec. 707), and it is the long-term half of Title VII’s design. Where Subtitle A bridges — a dedication of existing tariff revenue, a capped transitional fee on the vehicles the fuel tax cannot reach — Subtitle B builds and tests the replacement. Against current law, the change is the creation of the federal instrument itself: Congress has funded state experiments and directed studies toward a national pilot; it has not enacted one. Title VII’s two halves are also sequenced against each other: the Subtitle A fee exempts pilot participants (Sec. 704(b)), so the bridge never charges a driver the pilot is already measuring.

Sec. 708 establishes the American VMT Choice Pilot Program as a five-year, nationwide, voluntary pilot, and states its purpose in the bill’s own words: “to test, at a national scale, the feasibility and public acceptance of a fee based on vehicle miles traveled as a sustainable, user-based replacement for the Federal gas tax, and to provide Congress with a final report and legislative recommendations for a permanent, nationwide implementation.” The pilot is an evidence-producing instrument, and its product is defined in advance.

Sec. 709(a)–(b) — Who may take part, and on what terms

The pilot is open to owners of light-duty vehicles registered in all 50 states. Participation is incentivized on two terms. Revenue neutrality: a participant receives a full credit or refund of the federal gasoline and diesel taxes paid during participation — the mechanism is set by the Secretary of the Treasury, but the hold-harmless result is statutory: “A participant shall be held harmless for the cost of participating in the pilot.” And an annual stipend is authorised for participants, fixed in amount by Sec. 709(d) below.

Sec. 709(c)(1) — The Simple Odometer System

The default model. A participant uses “a non-Global-Positioning-System-enabled device or other approved method that securely records only the total miles traveled by a vehicle”; the charge is “a single, flat national rate per mile, calculated to be revenue neutral with the average Federal gas tax paid”; and the system “does not collect or store any location-based data.” The design premise is that a national per-mile system must be joinable by a driver who will not be tracked — and that the system should be required to prove it can work for that driver.

Sec. 709(c)(2) — The Smart VMT System

The opt-in model. A participant may choose a GPS-enabled telematics device or a certified software application, and the rate becomes variable, including, at minimum, “a lower rate for miles traveled on designated rural roads” and “a higher rate for miles traveled in designated, highly congested urban areas during peak hours of travel.” Nothing about this model is a default: it exists only for drivers who affirmatively select it, and its purpose is to test whether location- and time-differentiated pricing is technically workable and publicly acceptable before any legislature is asked to adopt it.

Sec. 709(d)–(e) — Selection, stipend, and the destination of revenue

Enrollment is by a simple public online application. Selection must produce a geographically and demographically diverse cohort — urban, suburban, and rural drivers, across income levels, driving gasoline, hybrid, and electric vehicles — because a pilot whose sample cannot speak for the national fleet cannot support the recommendations Sec. 711 requires. The annual stipend is set at not less than $100 and not more than $400. Revenue handling is prescribed by Sec. 709(e): collections go to a separate Treasury account and are distributed “not later than 60 days after the end of each calendar quarter” — base per-mile revenue to the Highway Trust Fund, and urban congestion surcharge revenue to the metropolitan planning organization or public transportation agency for the area where it was collected, to be used for “projects that provide transportation alternatives to single-occupancy vehicle travel, including for the capital and operating costs of public transportation, and for projects to improve safety and accessibility for vulnerable road users.” The surcharge is thus tested as a local instrument, funded locally, from its first day.

Sec. 710 — Privacy as statute

Three rules, each legislative rather than administrative. Data minimization: the program may collect only the minimum data necessary to administer the fee. Purpose limitation with teeth: “The use of any data collected under this subtitle for law enforcement purposes or any purpose other than the administration of the pilot program, without a specific warrant from a court of competent jurisdiction, is expressly prohibited.” And security by public process: standards for the anonymization, encryption, and secure storage of program data are to be set through public rulemaking, where they can be examined before they are relied upon.

Sec. 711 — The report and the recommendations

Not later than one year after the pilot concludes, the Secretary reports to Congress on the technological feasibility, the public acceptance, and the administrative complexity of each model tested. The report must carry legislative recommendations for a phased, nationwide, mandatory transition to a universal per-mile fee, including a recommended rate structure ensuring the long-term solvency of the Highway Trust Fund, and a proposed timeline and legislative framework for the full repeal of the federal taxes on gasoline and diesel fuel. The pilot therefore cannot end in a drawer: its statutory output is a bill-ready conversion plan, argued from five years of national operating evidence.

Section 707Short title.

Section 708Establishment of a National VMT Pilot Program.

Section 709Pilot program design and incentives.

Section 710Privacy and data security protections.

Section 711Report to Congress and legislative recommendations.

Source: FRAME Act, Official Draft, December 15, 2025, Sections 707 to 711.

The data case

The institutional warrant: GAO asked for this pilot

The Government Accountability Office recommended the substance of Sec. 708 more than a decade before FRAME was drafted. In GAO-13-77, examining mileage fees for the vehicles that contribute least to the Trust Fund, GAO concluded that “without a federal pilot program” to evaluate charging options and their costs and benefits, “Congress lacks critical information to assess whether mileage fees for these vehicles could be a viable and cost-effective tool” — and found more than half of state transportation departments would support federally led field tests.

Its 2021 follow-up on the state-grant program that Congress funded instead (GAO-22-104299) records thirteen states and two multistate coalitions piloting since fiscal year 2016, and recommends that the Federal Highway Administration develop criteria for judging how pilot results scale to national policy. The Eno Center for Transportation’s Driving Change report is devoted entirely to advising the design of the national pilot. Section 708 is that pilot, legislated: national in scale, explicit about scalability, and required by Sec. 711 to report against exactly the questions GAO posed.

What the operating programs show

Oregon (OReGO, since 2015) runs the longest-standing American per-mile program, at 2 cents per mile, open to electric and fuel-efficient vehicles (National Conference of State Legislatures). Its Road User Fee Task Force found a per-mile charge worth approximately $340 million in additional gross revenue over ten years against the fuels tax, with rural drivers as a group paying less than under the fuel tax — the empirical basis for Sec. 709(c)(2)’s lower rural rate, which prices a difference the data already shows. Utah (since 2020) charges 1.5 cents per mile (Utah Foundation); its early economics were those of every small program — fixed costs against a small cohort — but by fiscal year 2025 collections of $384,000 exceeded program costs of $291,000 (Utah DOT reporting, compiled by the Reason Foundation), the crossover the national pilot’s scale is designed to reach faster. Virginia (Mileage Choice, since 2022) supplies the retention evidence a voluntary design lives or dies on: some 23,000 participants as of June 2025, 80 percent renewing after the first year and 71 percent still participating after two (Virginia DMV). Drivers who can leave, stay.

Acceptance: low in the abstract, earned in operation

The research record on public acceptance is consistent, and FRAME’s design reads as if written from it. In the abstract, support for mileage fees is low: the NCHRP synthesis of national survey research (Agrawal, Mineta Transportation Institute) found mean support of 24 percent across 33 survey questions. Exposure reverses it. In the University of Iowa’s national field study — more than 2,600 volunteers, 21 million miles, ten months — favorability rose from 42 percent before participation to 70 percent after, with over 70 percent of participants judging the system fair, reliable, and accurate. Colorado’s pilot reported the same pattern: 91 percent of participants would take part again, 81 percent agreed a road usage charge is a fair funding method, and 88 percent felt their personal information was secure (Colorado DOT). The lesson for design is precise: acceptance is produced by participation under fair terms — which is why Sec. 709 is voluntary, hold-harmless, stipended, and offers a no-location model as the default, and why Sec. 710 puts the privacy rules in the statute rather than in a brochure.

• Renewed after the first year
• Did not renew

Each box is 10 percent of participants.

Source: Virginia Department of Motor Vehicles, Mileage Choice presentation (November 2025).