Federal transit formula grants pay 80 percent of the net cost of a capital project and zero percent of operating expense above 100 peak buses in an urbanized area of 200,000 or more, as read on 2026-08-08

What is happening?

Federal transit formula grants pay 80 percent of the net cost of a capital project and zero percent of operating expense for any transit system running more than 100 peak buses in an urbanized area of 200,000 or more. As of 2026-08-08 that bar reaches 85 agencies, and those agencies carried 6.488 billion unlinked passenger trips in National Transit Database report year 2024 — 84.6 percent of the national total of 7.669 billion — while spending USD 46.4 billion on operations.

The statute is 49 U.S.C. 5307, read on 2026-08-08 in two independent renderings that agree on every figure. Operating costs are eligible in an urbanized area of fewer than 200,000 people. In an area of not fewer than 200,000, operating costs are eligible only for a system running 75 or fewer peak buses, up to 75 percent of the share of the apportionment attributable to it, or a system running 76 to 100 buses, up to 50 percent. Above 100 buses the share is zero. Where an operating grant is possible at all, the federal share is capped at 50 percent of net project cost, against 80 percent for capital.

The bar is narrower than the phrase suggests, and the narrowing is the whole mechanism. TEA-21 in 1998 removed operating eligibility in areas above 200,000 and in the same act moved preventive maintenance into the capital column, where it draws the 80 percent match, and defined ADA complementary paratransit as a capital project. Federal money therefore still reaches a large share of what the industry books as operating cost. On 2020 data, vehicle maintenance is 16 percent of operating expense and non-vehicle maintenance is 11 percent, so 27 percent is reachable through the capital label. Vehicle operation is 42 percent by function and compensation is 62 percent by object class. Federal formula money can keep the bus maintained but cannot pay the person driving it, and driver hours are the input that decides how much service exists.

Whose problem is this?

The people who decide are not the people who pay, and the funder with the tightest restriction on how its money is spent is also the smallest of the four.

Where does this problem end?

The boundary here is a spending label rather than a service need, so the question is about relative prices and not about the size of the program. The money exists and the institution exists. What is fixed is which column the money is allowed to enter.

What is the state now, and what should it be?

Now

Needs a new measurementthe target state: no source opened here states a target federal share of operating expense, a target operating-eligible fraction of the formula, or a threshold that the 100-bus line is meant to approximate. The Department of Transportation conditions and performance report explicitly makes no recommendation about the federal share of capital investment, and the Congressional Research Service is non-recommending by design, so the absence is institutional rather than accidental.

How big is it?

Between 149.7 million and 155.6 million residents. Both ends count people living in urbanized areas whose principal transit operator cannot spend federal formula money on service hours, computed from National Transit Database report year 2024 records retrieved on 2026-08-08 against the 2020-census urban populations the database carries.

The high end, 155.6 million, is the combined population of the 57 urbanized areas containing at least one agency above 100 peak buses, the barred set under current law. The low end, 149.7 million, is the combined population of the 50 areas containing an agency above 125 peak buses, the set that stays at zero percent even if H.R. 8870 becomes law as reported out of committee on 2026-05-22.

The width of the band is the finding. It is 5.9 million people, about 4 percent, and that is the entire population effect of the only change on the legislative table. The 15 agencies in the proposed new band carried 118.1 million trips in 2024, which is 1.54 percent of the national total, and spent USD 1.36 billion on operations.

This is a residence count and not a rider count. The database reports unlinked passenger trips, which are boardings, so a two-vehicle commute counts twice and a daily rider counts roughly five hundred times a year. No federal dataset reports distinct persons served by a transit agency, and the American Community Survey interface refused the request without a registered key, so a rider headcount is not derivable from anything opened here.

Under what conditions does it arise?

1. The statute sets a relative price rather than a budget. For a barred agency a dollar of capital costs twenty cents of local money and a dollar of driver time costs one hundred cents. Nothing in the rule asks whether the community needs another mile of guideway or another hour of service. The discount attaches to the label on the invoice.

2. The escape hatch relocated the constraint instead of removing it. Because preventive maintenance became a capital item in 1998, federal money reaches maintenance freely. What it does not reach is service hours. So the visible symptom of a squeeze is never a rusting fleet. It is a cancelled route.

3. The founding rationale is still standing and still has evidence behind it. The 1962 study on transit needs by Researcher A recommended against federal operating support on the ground that it weakens the incentive to control cost. Researcher B measured bus cost productivity declining 1.4 percent a year before federal aid, 2.1 percent in the capital-only era, and 3.1 percent once operating aid arrived. Researchers C, D and E found federal operating assistance roughly twice as cost-inflationary as state subsidy. No source opened here retires that evidence, which is why the bar survives each reauthorization rather than being quietly dropped.

4. The bite is uneven and the sources say so. The Congressional Research Service records that the restriction matters less to large agencies with continuous capital programs, because they can shift their own funds between accounts, and more to smaller agencies whose capital spending is lumpy. Read strictly, that bounds the mechanism to agencies just past 200,000 population or just past the bus threshold.

5. The federal share is small and inflexible at the same time. In 2019 the four funders of transit operating expense were fares at 34.0 percent, local at 35.1 percent, state at 22.9 percent and federal at 7.9 percent. Only the federal share carries a statutory restriction on service hours, and only the federal share cannot be waived by a local decision.

What has been tried?

Two directions have been tried across sixty years and both stop at the same place. Every permanent change has moved the threshold, and none has removed it.

What was found?

Why is it still unsolved?

Incentive inversion. The rule attaches its discount to the accounting label rather than to the outcome, so the cheapest thing a barred agency can buy with federal help is an asset it may not be able to operate, and the most expensive thing is the service that asset exists to provide.

The first movement is the price itself. A barred agency faces a federal price of twenty cents on a capital dollar, one hundred cents on a service-hour dollar, and fifty cents on the rare operating dollar that is allowed at all. The Congressional Research Service states the consequence directly, warning that the inflexibility may lead agencies to overcapitalize by buying buses and building rail lines they cannot afford to operate. That is the inversion in one sentence, and it is stated by the analytic arm of the body that wrote the rule.

The second movement is that the rule does not stop operating money from being found. It only determines where the conversion happens. In 2019 the federal government supplied 7.9 percent of transit operating revenue against 35.1 percent local, 34.0 percent from fares and 22.9 percent from states. Every one of those other three can be moved by a decision taken locally. Pennsylvania proved it on 2025-09-08, when the governor directed the state transportation department to approve the use of up to USD 394 million in capital assistance for daily operations, drawn from an unobligated capital allocation, to preserve service for two years at an agency running more than a thousand peak buses. The Chicago regional authority did a smaller version of the same thing in its 2026 patch, moving USD 17 million from capital to operating while a commuter rail board proposed sending USD 60 million of farebox revenue the other way. Neither source mentions federal eligibility at all, and no source opened here draws the connection, so the link between the federal restriction and these particular conversions is not asserted here. What the sources do establish is the shape: the restriction binds the smallest funder, it is the only one that cannot be waived locally, and asset money is what gets spent down when service has to be preserved.

The third movement is why nothing dislodges it. The 1962 argument that operating aid weakens cost discipline has real measurements behind it and has never been retired, so the bar is a live policy position rather than an oversight. The industry itself asked only to move the line from 100 to 125 buses. The bill carrying that ask was reported out of committee on 2026-05-22 by 62 votes to 2 and has had no floor action, and the authorization it would replace expires on 2026-09-30. A threshold defended by evidence, challenged only by a request to move it slightly, and attached to a bill that has not reached a floor vote is a threshold that survives by default.

What observation would mean it is solved?

Candidates — (a) Congress amends 49 U.S.C. 5307(a)(2) so that the federal operating share no longer falls to zero above a fleet-size threshold (b) the federal share of national transit operating expense rises and stays risen outside an emergency (c) agencies stop converting capital funds into operating funds to preserve service.

(a) alone is weaker than it looks. Section 3005 of H.R. 8870 is an amendment to exactly that subsection and it changes the position of 1.54 percent of national ridership. An amendment that moves a threshold can be counted as reform while leaving 83 percent of riders where they were.

(b) alone can be produced by disaster. The federal share of operating expense went from 7.9 percent in 2019 to 15 percent in 2023 without a single word of the statute changing, because pandemic relief money was temporary and unrestricted. A rise that comes from emergency appropriations reverses when the emergency money is obligated, and more than 99 percent of it was obligated by the middle of 2023.

(c) alone measures a symptom that can be suppressed. Capital-to-operating conversion stops if there is no capital money left to convert, and it also stops if service is simply cut instead. Chicago published both paths in the same document, listing a small conversion alongside a potential 25 percent service reduction. The three have to be read together, and (b) has to be read against the year it came from.

What is it connected to?

Fills with researchhighway trust fund solvency and the same capital-versus-operating asymmetry in the highway program, state transit funding formulas and the fiscal cliff literature, the labor economics of transit operator recruitment and retention, and the treatment of operating subsidy in other national transit systems. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• Where the numbers came from. No source opened here explains the origin of 200,000 population, or of 100 and 75 buses. The 1998 fact sheet states the elimination flatly and contains no transition or phase-out. A search for the legislative rationale of the 2013 provision returns the statute, agency overviews and encyclopedia entries, and no document addressing why the line sits at 100. The thresholds appear as given.
• The condition is missing from the money notice. The fiscal year 2025 apportionment notice, published 2025-09-15, is the document that tells each recipient how much money it has. It contains no 200,000 threshold, no 100-bus rule and no 75 or 50 percent caps, and points readers to a circular instead. Read on its own, the notice is silent about what the money may not be spent on.
• Nobody counts who crossed the line when the census moved it. The 2023 redesignation notice records 15 urbanized areas newly designated above 200,000, two falling below, and two held harmless, and it does not mention operating-assistance eligibility. An agency can gain or lose the right to spend federal money on service hours because a boundary was redrawn, and no source counts those agencies.
• The central causal claim has no quantity attached. The overcapitalization warning is asserted without an instance, without a count of assets built and not operated, and without a dollar figure, in every source opened here.
• No published count of who is on either side. The Federal Transit Administration does not publish an operating-eligible against operating-ineligible split of the formula. The counts of 85 barred agencies, 15 in the proposed new band and 1.54 percent of ridership had to be computed from microdata for this document. Congress is being asked to move a threshold with no published tally of who it moves.
• The asset number is eight years stale. The USD 124 billion backlog rests on 2018 data published in 2024. Every argument that operating flexibility would slow backlog reduction is made against a figure whose as-of predates the pandemic that reshaped the demand it is sized for.
• Fiscal year 2026 has no apportionment record in the Federal Register. Of 52 agency documents published there since 2025-10-01 and checked directly on 2026-08-08, none is a fiscal year 2026 apportionments notice. The most recent is fiscal year 2025.
• No source describes the default. Neither research service report nor the bill states what happens to formula apportionments on 2026-10-01 if reauthorization has not passed.

See the evidence

Two renderings of the statute were read directly and they agree on every figure, which is the strongest link in this document because everything else rests on what the text actually says. Two Congressional Research Service reports were read in full from downloaded files and supply the cost structure, the funding shares, the history and the economics literature. The bill text was read from the primary source, and its committee action was confirmed from a third-party tracker because the congressional site refused automated access. Where sources disagree the disagreement is left visible rather than resolved: the count of large urbanized areas is 192 in the 2023 designation notice and 184 in the 2024 database extract with no published crosswalk between them; the program size for fiscal year 2026 is USD 22.3 billion authorized against a USD 20.937 billion actual appropriation for fiscal year 2025 and an unverified USD 20.6 billion circulating in secondary summaries; federal operating support is USD 5.0 billion a year in 2022 dollars and USD 5.3 billion in 2024 dollars from the same underlying series; and the reinvestment backlog is USD 126 billion on 2016 data and USD 124 billion on 2018 data depending on which edition is cited. The agency counts of 85, 70 and 15 are a proxy computed from database vehicle counts, not the statutory peak-bus roster, and no source publishes that roster. This is a Path A output, so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range 149,700,000–155,600,000 As of National Transit Database report year 2024, populations on the 2020 census urban delineation, retrieved 2026-08-08

Sensitivity The chain terminates at the high bound. The low bound is not a separate derivation but the same three steps rerun with the fleet filter set above 125 rather than above 100 peak buses, which yields 70 agencies in 50 urbanized areas holding 149,700,000 residents, so 155,600,000 minus 149,700,000 is the 5,900,000 the threshold move would release. The width is 5,900,000 people, about 4 percent, and that narrowness is the finding rather than a weakness of the estimate: moving the fleet threshold from 100 to 125 buses removes 15 agencies and 1.54 percent of national ridership from the barred set and leaves 83.1 percent of national ridership exactly where it was. What this number fails to count is people rather than places. It counts residents of affected urbanized areas, not riders, because the database reports unlinked passenger trips, which are boardings, and no federal dataset reports distinct persons served by a transit agency. The American Community Survey journey to work table would have given a commuter floor but the interface refused the request without a registered key, so a rider headcount is not derivable here. The count also excludes riders in large urbanized areas whose local operator runs 100 or fewer buses, being 389 agencies and 522.7 million trips in 2024, even though many of those riders also use a barred regional operator. In the opposite direction the estimate is generous in three ways: it credits every resident of an affected area whether or not that person uses transit, it is frozen at the 2020 delineation so fast growing areas are counted at their 2020 size, and it cannot count demand already destroyed, because riders lost to a service cut leave the ridership series and the measure shrinks as the problem worsens.

Regional breakdown No source publishes a state or regional split of operating-eligible against operating-ineligible transit agencies or of the section 5307 apportionment. The counts here were computed from national microdata for this document, and the underlying peak-bus figure is a proxy rather than the statutory count, so any regional decomposition would carry that proxy error into every cell without a published roster to check it against. Splitting the national figure by population share is not permitted and was not done.