Seattle wants better transit. So does everyone who signed the coalition letter calling on the City Council to fund the Seattle Transit Measure. So do we. But wanting better transit and building better transit comes down to a question the coalition letter doesn’t ask: better how, and in what order?
The proposal before the Council would nearly double bus service in Seattle. That’s worth doing. But ten minute headways on a bus stuck in traffic aren’t ten minute headways. They’re a schedule fiction. The bus leaves the terminal on time. It arrives whenever Denny Way allows. The rider learns not to trust it. The rider drives. And the ridership numbers that would justify the next investment never materialize.
Frequency without reliability isn’t a transit system. It’s an aspiration with a bus number.
When you don’t need to memorize a schedule, you are on a network that works for people. That’s the test. Not headways on a spreadsheet. Not service hours logged. Whether a person can walk to a stop without checking OneBusAway first. Seattle’s current network fails that test on almost every corridor that isn’t Link. The STM is an opportunity to start fixing that. But only if we’re honest about what fixing it actually requires.
A LADDER, NOT A RANKING
Transit modes aren’t a prestige hierarchy where rail is the real thing and buses are what you get when the money runs out. They’re a ladder. Each rung is appropriate to a specific level of proven demand, and you climb by earning the next rung — not by skipping to the top because it looks better from the ground.
The rungs, in order of cost per mile:
Local buses are the foundation. Ubiquitous, flexible, the thing most Seattle residents actually use every day. Undermined immediately by mixed traffic. The baseline from which everything else is built.
Real BRT — not RapidRide as currently constituted — means dedicated lanes, signal priority, level boarding, and off-board payment. This is what makes ten minute headways mean ten minute headways. Not a bus with a better logo. A bus that is functionally a surface train, running on a schedule you can trust because traffic cannot touch it. Aurora. Rainier. The 44. These corridors are ready for this rung now.
Surface rail and streetcar is what a corridor with real BRT earns when ridership justifies fixed infrastructure. Permanent, legible, faster, higher capacity. Only justified when the corridor has proven demand through BRT performance. Not before.
Subway and heavy rail is what surface rail earns when the corridor is saturated and grade separation is the only way to add capacity. The most expensive rung by an order of magnitude. Justified where every rung below it has been maxed out — not because it was promised, and not because it would be wonderful.
The metric that climbs the ladder is riders. Actual riders making actual trips on a system they actually trust because it actually runs on time. Not political will. Not aesthetic preference. Not the age of the promise.
WHAT RELIABLE ACTUALLY COSTS
According to the USDOT ITS Deployment database and the Federal Transit Administration’s Transit Signal Priority Planning and Implementation Handbook, signal priority hardware and software runs $5,000 per intersection where existing controllers can be upgraded, rising to $20,000–$40,000 per intersection where equipment needs full replacement. [https://www.itskrs.its.dot.gov/its/benecost.nsf/ID/2706a0bd21f048f585257b65005f20b0 | https://www.itskrs.its.dot.gov/2008-sc00155]
Even at the low end, the math is instructive. The STM’s $3.5 million annual capital budget, spent entirely on signal priority, buys roughly 87 intersections at $40,000 each — enough, in theory, to cover Aurora Ave N from the city limit to Denny Way (approximately 45 signalized intersections) in a single year. But that exhausts the entire capital budget for one corridor, leaving nothing for dedicated lanes, stop improvements, or any of the other twenty-nine corridors that need the same treatment. Signal priority without a dedicated lane recovers time at intersections and loses it in the blocks between them. The lane without signal priority does better but hemorrhages time at signals. Both together are what make a bus functionally a surface train. One year’s capital budget buys one corridor. Thirty corridors is a thirty-year project.
The $3.5 million capital line is a start. It is not a system.
THE VIRTUOUS CIRCLE
Here is what reliable transit actually does, in sequence.
Reliable service builds ridership. Ridership makes the development case — investors and residents locate near transit they can trust, and not near transit they can’t. Development increases density and tax base along the corridor. Increased density and ridership make the case for the next rung on the ladder: from bus to BRT, from BRT to surface rail, from surface rail to grade separation where the numbers finally justify it.
This is why “rail attracts development but buses don’t” is true in practice but wrong in principle. Rail attracts development because it signals permanence and reliability. Real BRT with permanent infrastructure — dedicated lanes, signal priority hardware in the ground, stations that feel like stations — sends exactly the same signal. Developers read permanence. A painted lane that might be removed in the next road diet is not permanent. Dedicated infrastructure with signal priority hardware in the ground is.
Build the ladder properly and the development follows. The development funds the next rung.
The virtuous circle runs in reverse too. Unreliable service erodes ridership. Eroded ridership undermines the political case for the next investment. The corridor stagnates at the rung it’s on, not because the demand isn’t there but because the infrastructure never proved it. This is the history of every RapidRide line that got the branding without the lane.
SOUND TRANSIT ISN’T WHAT PEOPLE THINK IT IS
Ballard rail is not coming from Sound Transit. It was never in their remit to solve a neighborhood transit problem. And the longer the region waits for ST to deliver a tunnel Seattle can’t afford, the longer Ballard goes without the reliable transit it could have had a decade ago.
Sound Transit is a regional agency. Its job is to connect regional centers — Seattle to Bellevue, Tacoma to Everett — at a scale and speed that buses cannot achieve. Link does this. Sounder does this. When ST is doing its job, it is building and operating the top rungs of the ladder between Metro networks in different counties.
ST’s job is not to decide what Ballard needs. That’s Metro’s job.
The confusion between these two mandates has produced an agency making neighborhood transit decisions with regional capital, while the neighborhood network — the rungs most riders actually use — goes underfunded and underbuilt. The result is a spine without ribs: Link is excellent where it exists and nearly irrelevant where it doesn’t.
ST was designed as a megaproject agency, not a transit operator. Its incentives run toward large capital commitments — which generate political credit and cut ribbons — rather than toward network performance, which is diffuse, slow to show, and doesn’t photograph well. Reform that accountability structure or stop asking ST to do things it is institutionally incapable of doing well.
THE SHIP CANAL IS NOT THE PROBLEM. THE TUNNEL IS.
The Ship Canal is a real geographic constraint. But it is a constraint on a tunnel. It is not a constraint on a bridge.
A tunnel under the Ship Canal through maritime industrial properties, at a cost that has grown from $5.2 billion to $10.8 billion and counting [https://seattletransitblog.com/2025/02/28/ballard-light-rail-at-grade/], designed to deliver a one-seat ride to downtown for Ballard residents while the corridors that would serve everyone go unbuilt.
The Fremont Bridge already exists. The corridor is not a hypothesis. Streetcars carried riders from downtown Seattle to Ballard via the Fremont Bridge as early as 1914, documented on Kroll’s Standard Map of Seattle [https://transitmap.net/kroll-seattle-1914/] and in the Seattle Municipal Archives photograph collection [https://archiveswest.orbiscascade.org/ark:80444/xv91323]. The Westlake streetcar line ran until April 13, 1941. [https://en.wikipedia.org/wiki/South_Lake_Union_Streetcar] We are not proposing something new. We are proposing something that worked.
A dedicated transit lane on that crossing — or a new fixed-span bridge designed for transit priority rather than for accommodating every tall-mast vessel on Puget Sound — running via the Route 40 corridor through Fremont and Leary Way to Ballard, is BRT that functions as a surface train. Seattle Transit Blog’s 2025 analysis of the at-grade Leary Way/Westlake alternative priced it at roughly $2 billion in current dollars, compared to the tunnel’s $10.8 billion — with door-to-door travel times to Westlake that are comparable once the four-to-six minutes needed to reach street level from a 140-foot-deep tunnel station are factored in. [https://seattletransitblog.com/2025/02/28/ballard-light-rail-at-grade/]
Does that produce a slower platform-to-platform trip than a tunnel would? Yes. Is it faster than what exists now? Substantially. Does it move real numbers of real people on a timeline that isn’t 2042? Yes. And if those numbers justify the next rung — if the corridor saturates, if the development follows the reliability, if the ridership makes the case — you build the tram. You build the rail crossing. You build it with a proven corridor beneath it and a population of riders who have already demonstrated they’ll use it.
The Ship Canal stops being an obstacle the moment you stop insisting on a tunnel. You cannot fight geography. But you are not required to fight it with the most expensive tool available.
METRO AS SEATTLE’S MULTIMODAL PROVIDER
Metro should own the full ladder within Seattle and King County. Buses at the base. Real BRT on proven corridors. Surface rail where BRT has earned it. Water taxi where water crossings are shorter and cheaper than any bridge or tunnel — West Seattle being the obvious case that keeps getting ignored in favor of an elevated rail solution on a corridor whose ridership has never justified it. Each mode chosen by a decision matrix: cheapest and fastest to build for a minimum throughput of riders. Not by prestige. Not by promise.
This is not a consolation prize framework. A corridor on BRT is not a corridor that got cheated out of rail. It is a corridor on the rung it has earned, with a clear path upward when the ridership justifies the next investment. The BRT that works is the argument for the rail that follows. Success on one rung is the receipt for the next.
Fund the STM. Then have the harder conversation about what it takes to make the buses worth funding — and what it means to build a ladder instead of a monument.
STOP FIGHTING HISTORY
Every missed opportunity compounds. The Interurban ROW that became a trail — a forty-two mile corridor connecting Seattle to Tacoma that would have made the Ballard conversation look simple, now a pleasant bike path and a lesson in institutional non-decision. The Eastside Rail Corridor following the same path. The 1970s cycling moment that coincided with oil shocks and pollution mandates and produced painted lines on arterials instead of infrastructure. The promise of Ballard rail made before anyone seriously priced the Ship Canal crossing, now load-bearing in ways that distort every subsequent decision.
None of these were decided by a single bad vote. They were lost to sequences of non-decisions, each locally reasonable, collectively catastrophic. Nobody voted to kill Eastside rail. They just kept choosing not to save it.
We are not arguing against ambition. We are arguing that ambition aimed at the wrong target, at the wrong cost, on the wrong timeline, is not ambition. The perfect is the enemy of good enough. Asking people to ride a BRT line from Ballard to Westlake that they will never wait 10 minutes for, that will never take more than 18 minutes to make the journey is not settling or accepting defeat. That’s a clear win that could lead to better service but will never get worse.
The best transit system is the one that actually gets used.
—
The author is a researcher and data journalist. His book Why the Rent Is Too Damn High was published in May 2026.