High-speed rail's real constraint is ownership, and a bank doesn't fix it

Jonathan McDonald's case against American high-speed rail opens with the number that ought to end most of the argument: private companies own about 136,000 miles of US rail line, against roughly 530 miles of public track held by Amtrak. McDonald, an executive consultant at Raul V. Bravo + Associates and a board member at the Commuter Rail Coalition, proposes in Construction Dive that Congress charter a National Railroad Infrastructure Bank to own, manage and maximize the economic benefit of national rail infrastructure, drawing federal, state and private capital together without new taxes and structured so that no stakeholder, from the freight railroads to Amtrak to organized labor, finishes worse off.

The proposal is a serious one from an author who is not a tourist on the beat, and it is worth taking at full strength. It is also, on the op-ed's own evidence, a capital-markets answer to a title problem, and the title is the harder half. This publication has argued that consent has become the scarcest input in American infrastructure, and that the debt market will fund nearly anything with a counterparty and a schedule—rail is the purest case yet, because here the consent is held in fee simple by businesses whose core trade is not passengers.

The ownership arithmetic explains why the two standard delivery models keep failing: either a private developer spots a profitable route, or a government decides the line is needed and asks voters for the money, and both presume the ground is available. In the US it largely is not. McDonald is precise about why the freight railroads resist: high-speed track must be grade-separated from existing freight lines, and freight trains cannot run on it in any case, because heavy axle loads damage the structure. A corridor is therefore not a route on a map but a chain of easements, purchases and crossings, each held by a company whose revenue comes from moving freight over the same dirt.

Cost and duration follow from that chain, and the op-ed puts a single high-speed line at 15 to 20 years and more than $100 million per mile—figures that are the whole case for a bank, and the reason a bank alone does not close it. The hundred-million-dollar mile is a financing number, and money compresses financing numbers efficiently; the 15 to 20 years is a consent and construction number, and nothing in the proposal as published compresses that one.

The fiscal claim deserves close reading too, because a bank capitalized without new taxes has to be filled from somewhere—whether existing federal rail programs, loan guarantees, or the private capital it can induce—and the published argument does not specify the stack. That omission is the proposal's softest joint, because a bank whose capital structure is left to the appropriations process is a bank capitalized by whoever holds the most leverage in that process.

Then there is the condition that no stakeholder be worse off: freight railroads, Amtrak and organized labor each sit on something a corridor needs, whether the ground, the terminal capacity, or the crews, and a bank that must satisfy all three before it writes a loan would move at the pace of the slowest of them—a description of what a unanimity rule does to a construction schedule, not an allegation about any of the parties involved. McDonald anticipates the objection by insisting each one comes out ahead, and he may well be right on the merits, but the mechanism as described hands each party a veto without handing any of them a reason to spend it early.

Washington is already running a smaller version of the same experiment, since a Senate financing bill and USDOT's Union Station 'mirror' have put the next US P3 test in federal hands, and the question there is identical to the one a railroad bank would face: whether a federal balance sheet can hold the capital and the consents in the same hand. A National Railroad Infrastructure Bank would be that test at national scale, with two differences that matter: the asset lives for decades rather than for one station, and the counterparties own the ground underneath it.

Where a bank would genuinely open a door is greenfield, because an alignment that touches no private track sidesteps the 136,000 miles entirely, and a financing vehicle that can underwrite a long-dated greenfield build is a real instrument. The cost is the urban connection that makes a corridor worth riding, since the stations people want are on the ground the railroads already hold, and that trade, not the interest rate, is what a corridor sponsor would be pricing.

The op-ed's contribution is larger than its remedy: it names the ownership structure that has blocked American high-speed rail for a generation, and it does so in numbers a legislator can hold in one hand, while the remedy addresses the part of the problem that was never binding. If US high-speed rail gets built, the evidence will not arrive as a charter. The number to watch is the mileage of private track a corridor manager has signed for.