The Wrong Kind of Careful
A review of IFP’s Transit Abundance Playbook
Before we begin, a programming note. Tomorrow, i.e. on Wednesday, 24 June 2026 at 1100h ET, on the second Changing Lanes livestream, my guest will be the foremost American authority on regulating automated driving: Professor Bryant Walker Smith of the University of South Carolina. Smith helped develop the SAE’s ‘five levels’ of driving-automation taxonomy, and earlier this year he testified to Congress on how the United States should regulate self-driving. Subscribers will be notified when we go live; you may also sign up here.
The Transit Abundance Playbook, various authors, published by the Institute for Progress, June 2026
Italy is not an easy place to build a subway. Though not a poor country (it has the eighth-largest GDP in the world), it is poorer than the United States, and it suffers a massive debt-to-GDP ratio, severely constraining its fiscal flexibility. The density and antiquity of its cities means that tunnelling must contend with challenging work constraints, unhelpful soil composition, and archaeological protections. Notoriously, until a few decades ago its public-works contracts were utterly corrupt, with bids rigged and bribes paid throughout: the Mani Pulite (‘Clean Hands’) investigations of the early 1990s arrested some 4,800 people and placed more than half the national parliament under suspicion.
Given all of this, it’s no surprise that Milan and the smaller northern cities build subways (metros) for about $300 million per mile, and the ruins-riddled centre of Rome builds for about $800 million per mile.
The United States has every advantage Italy lacks. It is the richest country in the history of the world, with the deepest capital markets and the best engineering schools, a civil service largely free of corruption, and a century-old talent for building transit; before the war, it laid more streetcar track than any other nation on earth.
It is therefore surprising that it builds rail rapid transit for around $800 million per mile on average, i.e., it pays on average what Italy pays in its most-difficult environment. And in New York City—which builds the most expensive transit in the world—Americans paid roughly $2.6 billion per mile for the Second Avenue Subway.
To be specific: per mile, New York’s Second Avenue Subway cost more than ten times what one global peer city, Madrid, pays, and NYC’s East Side Access, at $11.2 billion, opened as the most expensive railway station ever built.1 As a consequence, despite its wealth, the country now builds very little transit indeed. Exhibit A is California’s high-speed rail line, whose projected cost has ballooned from $33 billion to more than $126 billion; with federal funding now clawed back and not a mile of usable track laid, whether the line will ever be built is an open question.
The richest country in the world, with all its technical capability, builds less transit and pays far more per mile than peer countries with far fewer advantages. Why?
That puzzle is the spine of the Transit Abundance Playbook, a collection published by the Institute for Progress on 17 June 2026. It consists of fifteen short essays by transit practitioners, researchers, and advocates, each aimed at naming and solving one driver of the USA’s exorbitant transit-construction costs. Most reach for a federal lever to fix it, because this is a ‘playbook’ for the US Congress to run to bring American transit up to a global standard.
My readers should be advised that I’m no neutral third-party reviewer; I wrote one of the Playbook’s fifteen essays. (More about it later.) That alone would make me a strong proponent of the book and its prescriptions. But I have more reason than that to recommend it.
Before I became the writer of Changing Lanes, I worked in the space the playbook describes. As a public servant, I spent four years planning a new bus rapid transit line in western Toronto (across Mississauga): I wrote the project scope, procured a consortium of planners and engineers, and led that team as we drafted, consulted the public upon, and ultimately obtained political approval of the project plan. Later, I consulted on several transit projects, both as part of a major firm and as a one-man shop.
And given that experience, I say that the Playbook is right. The problems it identifies are real; they are the most important ones to solve; and the solutions on offer should be implemented.
Beyond that, I want to say what the Playbook is too polite to make plain, and name the root of the problem. What is holding back American transit? The Introduction states that “there is no single cause”. On the contrary, I think that there is, and that we can see that cause at work behind almost every individual contribution. Knowing the enemy is vital to fixing the problems holding back American transit: not only its construction, but its operation too.
That enemy’s name is defensive proceduralism.
Rules as Tools, or as Weapons
The key word here is defensive. Let’s define proceduralism as the tendency to establish rules in advance for how to scope, design, and deliver a project, and for what kind of input the project will accept, from whom, and how. This is a good thing, allowing everyone involved to know the game being played and how to play it. Italy’s experience demonstrates the value of proceduralism. The country enjoys very low per-mile transit construction costs today because, to combat corruption, it relies heavily on open procedure: published price lists, an anti-corruption authority, bids scored on merit, and internal state capacity such that civil servants are able to run a project themselves. These measures aim to prevent corrupt practices, and have the helpful downstream effect of keeping costs low.
So Italy has proceduralism. What is the USA’s problem, then? Not the presence of rules, but their aim. In Italy, the rules aim to deliver projects on time and on budget. In the USA, the rules aim to stave off lawsuits or blame. That is the American disease, and once you look for it, you will find it behind nearly every contribution in the collection.
Start, for instance, with Transit Projects Need a Single Decision-Maker, written anonymously (!) by someone who “works in project delivery for a public transit system”. Perhaps because they are anonymous, the author of that piece states the case more plainly than the authors of the introduction dare to. American institutions are unusually open to litigation, everyone understands that judges rather than elected officials or civil servants are the final arbiters of what is permitted, and that those judges are not trained as engineers or planners, but rather to determine whether the rules were followed. Agencies, knowing this, lower their ambitions and spend their days documenting decisions rather than making them. In such an environment, procedure is less a tool for building and more a weapon to stave off lawsuits. Environmental review under NEPA or CEQA is one example of this; so are rigid sequencing gates, or lowest-bid mandates that exist mainly so no official can be accused of favouritism.
Consider also procurement, as Alon Levy (whom Changing Lanes admires deeply) writes in We Should Know How Much Transit Components Cost. American agencies prefer to construct their transit projects in lump-sum contracts. Why do they do this? So that when cost overruns appear, as they will, the contractor must absorb them, keeping the project on budget and hence the agency free of blame. That’s defensive procedure doing exactly what we would expect, and the result is exactly what we would expect as well: bids are inflated dramatically to cover the risk being transferred. On top of that, there is the change-order racket. Contractors win bids by submitting low, and then any change that comes up after award is a change order, and the opacity of their costs—lump-sum contracts mean that per-component and per-activity costs are kept secret as commercially sensitive—allows for maximized change-order values, to the tune of 20 to 50 percent. The fix is not less procedure but different procedure, namely the itemized, public pricing Italy and Spain use, which Levy estimates would cut 15 to 20 percent from a typical project.
The thread runs everywhere. It’s in Eliminate Redundant Subway Cross-Passages, where Brian Potter shows that American tunnels carry roughly twice as many cross-passages as European ones, because the National Fire Protection Association (NFPA) 130 fire standard demands one every 800 feet against Europe’s 1,640, even though the NFPA concedes it has no technical substantiation for the spacing. The requirement is merely defensive, a margin of safety nobody can be blamed for adding.
The same instinct shows up in Fast-Track Democratically Approved Transit Projects, in which Hayden Clarkin describes projects that voters have already approved at the ballot box being sent through years of NEPA review, not to inform a decision the public has made, but to hand opponents a second venue in which to undo it. The defensive party here is the agency: it dares not waive or even shorten the review for a project voters have already approved, because an under-reviewed approval is exactly what draws the lawsuit. So the agency runs the full multi-year process to protect itself, and that very thoroughness is the weapon its opponents pick up.
Rules First, Build Second
From my days as a transit-project lead on the public side, I remember the order of operations: my job (and that of everyone I worked with) was to follow the rules first, and bring the project in on time and budget second. Nobody wanted delay, but the facts of the matter were that breaking a rule, no matter how small, was a thing one could be punished for; breaking the schedule or the budget was not. In such environments, the rational move is to give up decision-making power, since that means also dropping the burden of blame.
In Put Transit Staff in Charge of Their Own Projects, Paul Lewis traces why American agencies hollowed out their own expertise at transit building: agencies lean on consultants as decision-making buffers, hiring outside firms to validate controversial choices and supply political cover, a habit that reinforces risk-aversion and diffuses accountability so that no one inside is answerable for anything. A staffer who uses judgment and is wrong owns the mistake; one who can point to a consultant’s report does not. I have worked as a consultant in this world, and out of respect for everyone involved, I will say only that Lewis is absolutely right.
Four of the collection’s pieces map this same instinct to seal off every angle of attack and to pay for the invulnerability in taxpayer money and additional years spent. In Get the Best Value in Transit Procurement, Anthony Potts shows that lowest-bid mandates—defended as a guard against favourtism and graft—instead shut out the best-value bids. In Focus Capital Investment Grants on Improved Project Delivery, Stephanie Pollack shows us a federal grant process grown so rigid that its own rules now obstruct the timely, affordable delivery they were written to secure.
The remaining two turn on environmental assessment. In Let Agencies Do Their Own Environmental Review, Jamey Tesler shows agencies compiling reams of paperwork merely to certify that a routine project has no significant environmental effect, a status that all but a handful of projects plainly qualify for. Tesler’s remedy is to let agencies self-certify these routine findings. As it stands, the documentation protects not the environment but the FTA’s decision from challenge, and gets done in full even when the answer was never in doubt. In Let Transit Agencies Buy Land, Aidan Mackenzie shows agencies unable to buy land or move utility lines until NEPA review is complete, to avoid prejudicing the findings. That process takes on average 5.7 years, so they end up paying more for the land, and sometimes routing around it entirely.
Artist’s conception, after original art from the Playbook
Public servants still have some capacity for discretion, and some use it. On my own project, as part of the engagement process, one party whom I had a duty to consult tried to bill the project for the time and capacity spent on its own due diligence.2 I figured the demand was a bluff, refused it, and the request went away. But I remember that stand vividly today because it put the project, and my leadership of it, at risk; the party might have come back with a lawsuit. The smart thing to do for my career would have been to pay, write a memo explaining the situation, and move on. I didn’t because I found the demand offensive, but I can’t blame other leaders who decide the other way, their prudence outweighing their outrage.
The better path, of course, would be to draw the rules about consultation narrowly enough that leaders are never put in that position.
Fifteen Problems but One Foe
The Transit Abundance Playbook offers an institutionalist case for building transit more cheaply. The modesty of its introduction insists “there is no single cause” and lists the drivers as equals: overdesign, poor planning, veto points, permitting, thin capacity. That is sound coalition-building, but in my view it undersells the book’s best insight, the one its anonymous single-decision-maker essay says out loud: there is a single cause, namely the fear or being blamed, exacerbated by the litigiousness of Americans and the unusual latitude for American courts to entertain it.
Seen in that light, defensive proceduralism is the problem. Anyone who funds, plans, builds, or votes on transit should read this book, and should read it as one argument, that transit agencies need the expertise to run their own projects, the power to get them built, and the freedom from interference by opponents and rent-seekers.
In closing, I’ll note that defensive proceduralism shapes not only how America builds its transit, but also how it runs it. My own contribution to the collection, Close America’s Transit Automation Gap, is therefore an outlier in that regard, because it’s less about how to build more cheaply than about how to build a system that will be cheaper to operate. I encourage readers of Changing Lanes to read the whole thing (it’s about the length of a typical issue).
My argument there is also about defensive proceduralism: the reason that American transit agencies have not retrofitted their systems for full automation is a sixty-year-old labour-certification rule. That rule was drafted in 1964 to solve a real problem, but one that was solved by the end of the decade. It wasn’t removed, and not only remained on the books, but became supercharged in the 1970s as the national government became a major funder of transit operations. The result is that while the rest of the world (including Vancouver but not, to my chagrin, Toronto) automated its rail transit, the USA largely has not. Rectifying that failure would save hundreds of millions in operating funding, each and every year.
The Institute for Progress has done Americans a great favour by showing how transit construction could be improved. But there is also need for another playbook that shows how transit operation could be improved as well. If the Institute is looking for a new project, it could do much worse; I can think of a few ideas legislators would benefit from hearing.
The East Side Access project carved a new Long Island Rail Road (LIRR) terminal out of bedrock beneath Grand Central Terminal, linking the LIRR to Manhattan’s East Side through a tunnel under the East River; it opened in 2023 as Grand Central Madison.
Name omitted to protect the guilty.