$3 billion to fix 130,000 places where roads meet rails

Every highway-rail grade crossing is the same unsolved problem: a road and a railroad occupy the same few square feet, at the same height, and something has to tell a driver a train is coming.▸ The U.S. railroad network spans 140,000 route miles carrying about 204,000 grade crossings, of which roughly 130,000 are public crossings, maintained by a public authority rather than a farm or a factory. Two federal programs pay to make those crossings safer. One is old and automatic; one is new and competitive. Neither is close to finishing the job.

The Railway-Highway Crossings Program (RHCP) has existed in some form since the 1970s: a formula program under 23 U.S.C. § 130 that hands every state a fixed annual amount to upgrade existing signals and gates.▸ The Infrastructure Investment and Jobs Act (IIJA) locked in a floor for it — at least $245 million a year, every year, from fiscal 2022 through 2026.

Layered on top is something new: the Railroad Crossing Elimination (RCE) Program, a competitive grant program the IIJA created from scratch (49 U.S.C. § 22909), authorized at $3 billion over the same five years.▸ Where RHCP money buys incremental upgrades, RCE money buys the expensive fix — building an overpass, an underpass, or closing a crossing outright. It has gone out in three rounds: $573 million in FY2022 (63 projects, 32 states), $1.15 billion in FY2023–24 (123 projects, 41 states), and $1.15 billion made available for FY2025–26, for which applications closed in June 2026 and awards haven't yet been announced.

Add it up and the two programs will have put roughly $4.1 billion toward grade-crossing safety over five years — real money, and close to the full $3 billion RCE authorization already awarded or on offer.▸ It is also a rounding error against the size of the system it's meant to fix.

Grade crossings come in two kinds.▸ Active crossings have a warning system — bells, flashing lights, gates — that triggers automatically when a train approaches. Passive crossings have only a sign or a pavement marking: a driver has to notice the crossbuck, look both ways, and decide for themselves whether it's safe. Of the country's 130,209 public crossings, according to FHWA's 2022 Biennial Report to Congress, cited in the GAO's April 2025 review, 54% are active. 46% are passive — nothing but a sign, at tens of thousands of places where roads and railroads cross.

Most of the money doesn't go toward closing that gap.▸ GAO's analysis of RHCP project reports found that of roughly 6,050 projects states funded between 2019 and 2023, about 3,600 — nearly 60% — added or upgraded warning devices at crossings that already had active devices. State officials told GAO they prioritize already-active crossings because those tend to carry the most traffic and the highest risk. It's a defensible engineering choice, but it means the passive-crossing backlog moves mostly through the pricier RCE program, one grade separation at a time.

The cost of not closing it shows up every year in FRA's accident data: nearly 1,900 accidents at public crossings in 2023, and more than 800 people killed or injured.▸ Drivers or pedestrians failing to stop caused 31% of those accidents (584 of 1,882); stopping on the crossing itself caused another 26%. Pedestrians bore a rising share of the toll — 89 killed at public crossings in 2023, 41% of all grade-crossing deaths, up from 31% in 2014. GAO's own recommendation wasn't to spend more; it was that FHWA give states clearer guidance on which pedestrian-safety projects even qualify for RHCP funding, because the agency's current guidance mostly just repeats the statute.

Funding figures cover FY2022 through FY2026 as reported by GAO and FRA; the FY2025–26 RCE round was still pending award as of this writing. Accident and device figures are for calendar year 2023 and 2022, respectively, the most recent years GAO independently verified.