Billions in the Ground: How American Cities Keep Betting on Rail Lines That Riders Never Come to Use

In the spring of 2016, officials in Honolulu celebrated the ongoing construction of what they described as a transformative rail system — a 20-mile elevated line that would, according to projections, carry roughly 119,000 daily boardings by 2030. By the time the first segment opened in 2023, years behind schedule and billions over budget, ridership figures were a fraction of that promise. The Honolulu Authority for Rapid Transportation had spent the better part of a decade and nearly $10 billion constructing a system that, at its early operational stages, struggled to justify its existence in ridership terms alone.

Honolulu is not an isolated case. It is, in many respects, the clearest illustration of a pattern that has repeated itself across American cities for decades — a cycle in which political ambition, federal grant structures, and optimistic modeling conspire to produce rail infrastructure that underdelivers on its core promise: moving more people, more efficiently.

The Anatomy of an Overestimate

Ridership forecasting is a technical discipline, but it is never a purely technical exercise. The models that planners use to project future transit demand are fed assumptions about land use growth, population density, transit-oriented development, and behavioral shifts in commuting patterns. Each of those assumptions is a judgment call, and each one creates an opportunity for optimism to crowd out rigor.

A 2020 analysis by the Eno Center for Transportation found that major transit capital projects in the United States have historically overestimated ridership by a median of roughly 50 percent in their first years of operation. That figure is not a rounding error — it represents a fundamental and recurring disconnect between what planners promise and what commuters actually do.

Part of the problem is structural. The Federal Transit Administration's Capital Investment Grant program, commonly known as New Starts, requires project sponsors to demonstrate projected ridership as part of their funding applications. Cities competing for those grants have an institutional incentive to produce favorable numbers. When the models are constructed by consultants hired by the sponsoring agency, and when the careers and political legacies of elected officials are tied to project approval, the conditions for inflated projections are baked into the process.

When Politics Drives the Route

The selection of rail alignments — where a line actually goes — is another domain in which data-driven planning frequently yields to political negotiation. In theory, alignments should follow corridors of demonstrated demand: the streets and neighborhoods where the highest concentrations of transit-dependent riders already travel. In practice, routes are often shaped by the competing priorities of city council districts, real estate interests, and the preferences of influential constituents.

The result is lines that serve political geography rather than commuter geography. A rail corridor that threads through a commercially stagnant district because a council member demanded inclusion, or one that terminates at a suburban park-and-ride rather than a dense employment center, is a line that will struggle to generate the organic ridership that justifies its cost.

San Jose's BART extension offers a relevant illustration. The project, extending heavy rail to Silicon Valley at a cost that has escalated past $9 billion for fewer than seven miles of new track, has faced persistent questions about whether the alignment and station placement reflect where workers actually need to go, or where the politics of multi-jurisdictional negotiations led. Critics have argued that the money could have funded a far more comprehensive network of bus rapid transit and first-last-mile connections serving a broader population of commuters.

The Integration Problem

Even when a rail project is competently planned and reasonably well-aligned with demand corridors, its real-world performance depends heavily on how well it connects to the surrounding transit network. In too many American cities, rail expansions are treated as standalone infrastructure investments rather than as components of an integrated mobility system.

When a new light rail station opens without reliable bus feeder service, without protected pedestrian access from nearby neighborhoods, and without real-time information systems that help riders understand how to complete their journey, the rail line is functionally isolated. Commuters who might otherwise leave their cars behind will not do so if the final quarter-mile of their trip is an unresolved obstacle.

This integration deficit is partly a funding problem. Federal capital grants flow generously toward construction and rolling stock but are far less accommodating of the operational expenses associated with feeder networks and service coordination. Cities end up with gleaming stations surrounded by mobility deserts — precisely the condition that transit technology platforms and urban planners have long identified as a primary suppressor of rail ridership.

What the Data Should Be Telling Us

The irony of the current moment is that cities have access to richer mobility data than at any prior point in transit history. Anonymized cell phone location data, automated passenger counters, fare transaction records, and third-party navigation app datasets can now paint a detailed picture of how people actually move through a metropolitan area — where they originate, where they are going, at what times, and by what modes.

That data, rigorously applied, would in many cases argue against expensive rail construction in favor of targeted improvements to existing bus networks, congestion pricing to shift road demand, and investment in the connective tissue of first-and-last-mile mobility. It would identify corridors of genuine unmet demand and distinguish them from corridors of political aspiration.

Some metropolitan planning organizations are beginning to incorporate these richer data sources into their modeling frameworks. But the pace of methodological reform has not kept up with the pace of capital project development, and the institutional incentives that reward ambitious rail proposals over pragmatic network improvements remain largely intact.

Rethinking the Measure of Success

There is a broader question embedded in this pattern, one that goes beyond any individual project or city. American transit planning has long treated rail construction as an intrinsic good — a signal of urban seriousness and a magnet for economic development. The ribbon-cutting ceremony has become its own form of performance, distinct from the question of whether the resulting infrastructure actually improves the daily lives of commuters.

A more honest accounting would evaluate transit investments not by the ambition of their engineering but by their measurable impact on mobility outcomes: reduced commute times, increased transit mode share, improved access to employment for transit-dependent populations, and cost-effectiveness relative to alternatives.

By those measures, many of the rail projects that American cities have celebrated in recent decades would be difficult to defend. That does not mean rail is the wrong answer everywhere — in dense, high-demand corridors with strong network integration, rail investment can genuinely transform urban mobility. But density of investment does not substitute for density of riders.

Until the planning processes, funding structures, and political incentives that govern major transit capital decisions are reformed to place commuter outcomes at the center of the evaluation framework, American cities will continue breaking ground on the next Honolulu — and presenting the bill to the riders they were meant to serve.