Low cost per rider, because it only takes them a mile. (Visitor7/wikimedia)

There is no transit performance metric quite like ridership. However, when any metric becomes the single point of evaluation, it can lead to bizarre conclusions.

Last month’s High Speed Rail study counted riders, and some readers couldn’t resist comparing the numbers to middling local transit routes. But it’s one thing to take riders between adjacent neighborhoods and another to deliver them 180 miles away.

A project’s cost, at a first approximation, is proportional to its length. Single-minded attention to cost per rider means never building a long-distance project at all, which is nonsensical. That’s not to say that passenger-miles is a perfect metric, either. It favors long-haul routes that are irrelevant to the spontaneous trips that enable urban living. But a comprehensive evaluation must consider both when the difference in spatial scales is this large.

Moreover, Seattle-area transit and HSR aren’t fiscally in tension with each other. A state subsidy to Puget Sound transit that actually moves the needle is inconceivable. Meanwhile, HSR could connect cities all around the state with the wealth centered in Seattle, at a time when it is increasingly hard to drive there. This aligns with Olympia’s traditional responsibility for intercity rail and bus, and offers a good value proposition to voters across the state.

47 Replies to “Passenger-miles matter, too”

  1. Your point about Olympia’s responsibility to connect the parts of the State outside the core Puget Sound counties is well-made, but impractical except largely with buses.

    Yes, it makes sense to provide capital investment and some operating subsidies in the Cascades corridor, but intercity rail elsewhere in the State is preposterous. While it’s reasonably inexpensive to link Auburn to Ellensburg, Yakima and the Tri-Cities, there aren’t enough travelers in the corridor and are unlikely ever to be.

    Further, BNSF is almost certain to refuse use of its Tri-Cities-Spokane trackage which is very busy. And of course, it’s almost ridiculously out of the way for Spokane-bound travelers — roughly speaking it’s “two legs of a right triangle” with a 300 mile long hypotenuse.

    Should the State subsidize inter-county buses? Absolutely. There should be service from Tacoma to Aberdeen and Hoquiam via Olympia, Olympia to Vancouver via the I-5 corridor, Seattle to Walla-Walla via Ellensburg, Yakima and the Tri-Cities, Seattle to Spokane via Ellensburg and Wenatchee, Wenatchee all the way to the Canadian border, and direct service all the way on I-90 between Seattle and Spokane. The State should fund expanded Strait Shot and other Clallam/Jefferson inter-county services as well.

    But NOT trains anywhere other than the current Cascades.

    1. Where does the author suggest building HSR to Spokane? No one is suggesting we put the cart before the horse, obviously VAN-SEA-POR come fist and we’ve got decades to go before that becomes a reality.

      1. A HSR line to Spokane has been part of the recent conversation and has been included in a lot of the graphics going around.

        While I agree that there aren’t enough riders in Yakima, etc., if the connection is fast, they will come, so I would be wary about saying there never will be enough (barring some other constraint).

  2. Questions:
    Do passengers prefer trains over bus? daaa
    Will more travelers bypass car travel if a decent train ride exist? daaa
    Will car travelers forsake a bus because of low cost tweaker experience? daaa
    Will travels prefer train over plain if train travel time is reasonable? daaa

    As below illustrates, the time is right for the investment in better state east-west train travel.

    “Amtrak took over intercity passenger rail operations from the private railroads on May 1, 1971. Initial service on the Seattle–Portland portion of the corridor consisted of three daily round trips–one long-distance train. ”

    Seattle to Portland distance: 174 miles

    Portland Population in the 70’s
    1970 382,619
    1980 366,383

    Seattle to Spokane 278 miles
    Seattle to Tri-Cities 202 miles

    Spokane and Valley 2017 315,000
    Tri Cities 2019 296,000
    Yakima 2019 94,000
    (Also, Seattle’s population is significantly greater than the 70’s)

    Ridership and passenger miles are optimal for more frequent and higher speed train travel than what exist today.

  3. The average person still travels far more miles within their region that between major cities, so HSR still performs poorly compared to urban rail, even on passenger-mile metric.

    Even for intercity travel, the local transportation to and from the stations can make up a lot of the travel time.

    Even if Seattle to Portland were cut to two hours, the fares would be too high to spur much induced demand, barring a massive increase in operating subsidies.

    Just fund urban rail and intercity bus service. Intercity rail is a distraction.

    1. Seeing how there are 38 one way flights between Seattle and Vancouver alone shows that independent of regional demand there is also strong intercity demand. This is why they make specialized trains for intercity rail which are different than commuter trains and light rail trains.

      1. Al has the right idea. There is no perfect metric, but passenger time saved per dollar spent is probably the best we have. It incorporates runs that are very popular, as well as improvements that are dramatic. Make that subway run a bit faster and it is worth the cost. Shave a lot of time off that commuter run, and that becomes a good value.

        The problem is that super high speed rail is usually not worth it. There simply aren’t enough riders (as asdf2 wrote). There are exceptions (between mega-cities) but that is certainly not the case here. Here is some napkin math to show what I’m talking about:

        The high end estimates for ridership are less than 10,000 people a day (for the entire corridor — Portland to Vancouver BC). I’ll round up, to make it simpler. Now assume that we have achieved what is relatively affordable (https://seattletransitblog.wpcomstaging.com/2019/06/14/which-way-for-washingtons-intercity-rail-program/). That means two hours and thirty minutes to Portland. Now assume that we spend a bundle, and change that to one hour. Now assume everyone is going to Portland (again, to make the math simpler). That is a huge time savings (90 minutes) for 10,000 people a day. That works out to almost a million minutes saved per day.

        That is obviously a lot. But consider that Metro carries about 400,000 riders a day. Now imagine that we do something that somehow allows half those riders to save ten minutes a day. That is over 2 million minutes a day, or twice the time savings as the rosiest projection for super high speed rail.

        It begs the question, though: what is cheaper? Again, improving urban transport is much cheaper. It isn’t even close. We could have the buses running on tunnels (or trains if you prefer) and it wouldn’t cost as much as what some are proposing for high speed rail. It makes more sense to do what is affordable, and focus on incremental improvements to our intercity rail system.

    2. Your argument is like, why should we invest in 787s when we have 737s. There is more domestic travel so lets bag the longer distance travel.

      1. Right, which is why the 737 significantly outsells the 787. Investing in everyday service over a premium service makes sense for both the government and airlines!

    3. This “intercity” rail line can double as partial commuter rail under some of the scenarios (hourly+ frequency, stations in Everett and Tacoma, stations in Lynnwood and Kent-or-Auburn). And it’s not really intercity at all; it’s under 200 miles so it’s clearly regional rail.

      1. In other parts of the world, the fares for HSR are way too expensive to be viable as an everyday commuter route. For instance, in theory, one could commute from Providence, RI to Boston on the Acela in 30 minutes or so. But, not many people are willing to spend $70/day, round trip, to do it. Instead, the commuters will ride the slower trains, leaving the fast trains for people that don’t travel as often.

        In our case, a hoarde of commuters getting on the train in Tacoma would require hoarde of empty seats as the train leaves Portland. Serving commuters on long haul trains is not efficient, and cannot be done without a huge operational subsidy.

      2. I’ve seen 1-2 stop fares in the $10-15 range. That’s not that much more than Sounder, and the faster travel time and more runs per day will entice some people. People don’t just commute five days a week; they also travel for business once or twice a week or a few times a month.

    4. Eh, Seattle to Portland in about two hours probably isn’t too crazy. That wouldn’t even be HSR necessarily. At those speeds, you can probably get a good mix of local and intercity service. True HSR doesn’t really mix well with local service patterns anyway.

      But I agree, in that the problem comes when people start talking about Seattle to Portland in under 90 minutes. There are just soooo many better investments, particularly when we start talking about federal dollars for HSR development again (when he-who-shall-not-be-named is gone).

  4. All trip distances are important. Britain has the London underground, regional trains, and high-speed national trains. Germany has local subways (some like Link’s Westlake-Rainier Beach segment), S-Bahn, regional trains, and multiple levels of intercity trains. Areas like Bellingham and Spokane would be part of Seattle’s regional train system, and S-Bahns are conceivable on Marysville-Olympia and Woodinville-Renton-Seattle. So that’s what a world-class transit system would look like.

    In Washington we have the problem of a car-centric legislature and a rural-centric national government and stockholder-driven freight-centric railroads, so we have to compromise. But we should at aim for at least similar frequency and speed via some modes on all short-, medium-, and long-distance trip pairs, and not force people into cars, taxis, and airplanes more than they really want to, or make them forego trips because none of those modes are feasible or attractive.

    High-speed rail at its best could function as a hybrid regional train/S-Bahn between Bellingham and Portland (and with caveats Vancouver), and could provide people in Everett and Tacoma with what they wish Link/Sounder provided. I have my doubts whether Washington is willing to pay for it (both state and private), but we can see. As long as it doesn’t shortchange needed investments in Pugetopolis transit. As a fallback if we can’t get HSR, we should have buses with similar frequency and fares on Bellingham-Seattle, Portland-Seattle, and Spokane-Seattle. That means running at least every 1-2 hours, not a couple times a day.

    1. Even with the current speeds, we just need more service. I am looking at the Alaska Marine Highway for a trip next year. It leaves from Bellingham, right at the Amtrak stop. Perfect. I’ll just hop on the morning Cascades train to make the 6pm sailing… oh wait, there are no trains that go all the way from Portland to Bellingham in time for that boat. You have to switch to a bus in Seattle. Transfer all of your luggage, kids, etc…

      Oh, and if you want to go from Portland to Vancouver, there is one train a day that gets you in at 11pm. And the only one that will take you all the way back to Portland leaves at 6am…

    2. Mike, we ALSO have “the problem” of 300 miles of just about nothing if you follow I-5. There’s Ellensburg which is the ONLY thing that’s not nothing, and then nothing.

      GIVE. IT. UP. Americans won’t ride intercity trains until the cars are consistently stuck in gridlock and no possible option is on the horizon.

    3. Ellensburg is 180 miles away. Spokane is less than 300. If you go 300 miles south you’ll be somewhere in central Oregon.

      It’s the lack of transit options that make people drive so much. Each additional run generates more riders as people make decisions based on how the available transit fits into their schedule. Eventually you reach a tipping point and the number of people accelerate. Then we’ll get, maybe not as many people as Europe, but something approaching it. The common assumption is Americans drive because they want to and they won’t take transit, but there are a lot of people who would take transit if it’s robust, and some of them don’t realize it until after the transit is there and they try it. I’m neutral on whether we have statewide HSR, incremental medium-speed rail (110 mph), or all-day bus service. But the current Amtrak, Greyhound, Bolt, Grape Line, Skagit/Whatcom connectors, etc are really skeletal — no wonder so many people drive or fly. We should aim for some kind of hourly service north-south, and 1-2 hours east-west, including Sundays.

      1. It’s the lack of transit options that make people drive so much.

        No, it’s not if you’re talking about east of the Cascades.

        There’s no “there” there to produce the sort of trips at which intercity trains excel: the urban portion of City A to the urban portion of City B or perhaps the airport at either end if served directly. Everyone has to get in their cars to get to any rail station, and that station is within twenty miles of 3% of the population of Eastern Washington if that.

        Do I think we should improve the Cascades Corridor? Absolutely. But let’s not go all foamer and expect people to ride to Spokane via Yakima and Tri-Cities through the 1888 Stampede Tunnel.

        And just to be clear, there will not be a new tunnel built through the Cascades for HSR unless climate change entirely depopulates the Southern Tier states. If things are that bad, people probably won’t be traveling anyway; they’ll be surviving.

  5. I think passenger-miles is the wrong primary metric. It should be passenger-minutes. We can travel between Seattle and Portland via rail today. The challenge is how to do it in faster time!

    It’s also true that most transit agency expenses are measured in hours and not miles. Labor costs predominate.

    1. Passenger-miles is really misleading. People don’t use transportation to experience rolling in a box the longest distance; they use it to get somewhere they want to go, and they’d prefer the fastest reasonable time. Transit networks should be scored based on that. And you can’t compare city, metro, and regional trips identically; they’re apples and oranges. City trips will always have the most riders per corridor or per cost, but that doesn’t mean we can just eliminate the others. That would be like saying metropolitan work commutes are all that matter; that would mean no more food deliveries from Eastern Washington and we would starve.

      1. [T]hat would mean no more food deliveries from Eastern Washington and we would starve.

        Of course it doesn’t. Nobody is advocating that straw man.

      2. Ridiculous comment. You know that very little of the food consumed in Seattle comes from eastern Washington, right?

      3. Apples, potatoes, wheat, beef and wine grapes. All of which we can live without; but hops, no hops no micro brew and Seattle definitely could not survive without this lifeline.

    2. It depends. My understanding is that part of the motivation of Cascades HSR is to reduce GHG emissions, and if you’re concerned about minimizing GHG emissions, then total passenger miles is a decent starting point.

      At least in first order effects, Martin’s argument is plausible- there will be many fewer trips on HSR compared to local transit, but the trips will be longer, so the aggregate GHG emission savings could be larger than the aggregate GHG emissions savings that would result from instead spending the same amount of money on local transit.

      I’m less sure this is true in the long run. Building more local transit could not only induce more current residents to switch their trips to transit, but also incentivize new residents to live in more urban (instead of suburban areas)- the GHG savings from local transit may grow faster with time than the GHG savings from HSR.

      1. If GHG reduction is the goal, I would that mere electrification of existing Sounder, Amtrak and freight trains would be the fastest and cheapest way to do that.

      2. The percentage of GHS in the Puget sound region arising from intercity trips is tiny, and if you take the subset of those trips, it’s even tinier.

        For instance, let’s suppose your work commute is 9 miles each way, a very typical distance. That’s 18 miles round-trip. 20 work commutes adds up to the same distance as one round trip to Portland and back. Each month has about 20 weekdays. So, unless the average Seattle resident goes to Portland at least once a month, that’s not where their GHG’s are coming from. Obviously, some people do go to Portland every month. But, the vast majority of the Seattle population does not.

      3. Here’s my 3 step plan to reduce GHG emissions:
        #1 Don’t rack up a million frequent flyer miles running as a nothing candidate for president.
        #2 Don’t drive to the gym to get exercise
        #3 Don’t fund buses to drive around empty
        On the theme of “just don’t”, what are your top ways to reduce GHG emissions?

    3. If there were no long-distance transportation it wouldn’t come from California or Asia either. That’s for trucks and ships but the same principle applies to transit. Most of people’s trips are within 10 miles of home, but that doesn’t mean longer trips aren’t valuable too. People need to visit their parents and grandparents, go to/from college, attend events, move to a job in another city, etc. Seattle-Portland and the cities in between have several trips per day, Seattle-Vancouver has a few, but Seattle-Spokane and the cities in between have only two or three, and one of them arrives in Spokane at 12:30am both ways so few people can use it. The Pasco-Ellensburg bus runs three times a day and if you miss your connection in Ellensburg you’ll be waiting 12 hours for the next bus. The bus schedules to Vancouver WA make it on-and-off impossible to get there and back without an overnight stay. If you’re sticking to cheap local buses, the Everett-Mt Vernon connector is peak-only with no weekend runs. I like to attend MMA tournaments but most of them are in the suburbs or exurbs or casinos on a Saturday evening with no transit back afterward; the last run back is often in the mididle of the event. No wonder people don’t take transit when it’s not reasonably there. All of these should be improved so that people can get around both the city and the state and smaller towns without a car.

      1. All of these should be improved so that people can get around both the city and the state and smaller towns without a car.

        OF COURSE! The State should step up and fund inter-county transit service throughout the state. Seattle-Walla Walla via Ellensburg, Yakima and Tri-Cities. Maybe some turn back at Yakima. Seattle (or Tacoma) to Aberdeen/Hoquiam via Olympia. Olympia-Vancouver (WA) to fill in around Cascades. Seattle (or Everett) to Blaine and subsidies for the existing Skagit County service between Anacortes and Sedro-Wooley. Ellensburg-Wenatchee and on to Oroville. If Canada will connect, on to the border. Pullman (Moscow?)-Spokane-Colville/Kettle Falls. Support for the Strait Shot and other Clallam/Jefferson joint services.

        These are all cheap, cheap, cheap for the benefits they would bring to people who today must have a car, even if they live right in the middle of one of the smaller cities.

        But trains on these corridors (except the Cascades of course)? No. Not yet, and probably not ever. Well, maybe Seattle-Auburn-Ellensburg “exurban commuter style service. Ellensburg is well situated to be a hub and trains are more reliable across the Cascades in the winter, it’s true.

  6. Trying to justify the merits of any inter-city mass transit project as a stand-alone project is generally a terrible idea. We should be planning inter-city “systems” first before jumping into a project like it’s a new highway segment. Our Metro area has a culture that often looks at one project at a time (streetcars, for example) that creates suboptimal results.

    Consider this: A decent part of the SEA-PDX air travel is to connect to a flight headed elsewhere. How will that market be served if neither end point (including walking time at the ends) is within 20 or 30 minutes from a HSR stop?

    So let’s begin by pursing a statewide inter-city travel plan. That plan should include closer-in inter-city travel like Olympia, Ellensburg and Bellingham rather than just Portland, Vancouver and Spokane. That plan should make increasing frequency as significant as making connections faster.

    I could see some strategic systems projects coming out of this effort. Just a few reveries:

    – A year-round high-frequency rail connection to get through the Cascades , especially in the winter. The current Amtrak service is way too infrequent, circuitous and slow and I would think that an east-of-Cascades terminal would be popular in Central Washington.

    – A realistic high-frequency, faster proposal connecting Downtown Olympia with the system (even a single-track rail shuttle alongside highways to at least a DuPont station if not further).

    – An implémentation strategy on the best way to expand Sounder — possibly structuring a financial and organizational model to increase frequencies, and extend current routes beyond the district boundaries.

    It may be that buses have to be used to provide linkages in the near-term (day the first 10-30 years) until the rail systems can be implemented. That’s not unlike how ST Express will have provided 30+ years of service before the Link spine opens.

    1. The point about SEA/PDX flights as connecting to points elsewhere is well-taken and applies to SEA/YVR as well. It’s technically possible to get from the train to the airport in all three cities, but it’s nothing anyone is going to do; long rides and requisite transfers.

    2. People do though. You meet people on Greyhound and Amtrak arriving in Seattle and taking Link to the airport to fly somewhere, and vice-versa.

    3. “Trying to justify the merits of any inter-city mass transit project as a stand-alone project is generally a terrible idea. We should be planning inter-city “systems””

      Right, but that’s not what excites the funders or businesses. They want a high-speed regional train and don’t think about how they’ll get to the station.

    4. I haven’t flown to Europe out of Seattle in about 6 years. I ALWAYS take the train to Vancouver and then the skytrain to YVR. I wouldn’t even consider paying $300 more to fly to Vancouver just because it puts me in the airport as opposed to the $25 train ride and $5 for skytrain.

      I’ve taken the train to Portland too for the same reason but only once as the flights are rarely cheaper or the timing isn’t right. I can’t imagine someone flying to either city on purpose although I have a friend who does it. She takes two hours getting from her house to boarding the plane in Seatac, flies for 40 minutes, then takes two more hours getting her bag and getting to the city center and thinks it’s a good idea. It does happen.

  7. It’s a good point that it doesn’t necessarily make sense to compare HSR ridership to local transit ridership. Actually, when the previous study came out I think I was one of the first people to point out in the comments that HSR ridership is low compared to regular transit ridership. Now I kind of regret this, as we’ve become fixated on an apples to oranges comparison.

    The question is: how does the ridership on Cascadia Rail compare to other HSR routes of comparable length? That’s an apples to apples comparison.

    For cascadia rail, we are talking around 2 million annual passengers and a 310 mile route (according to the business case study).

    Madrid to Barcelona in spain is 386 miles long and and today has about 4.17 million passengers a year according to this link:

    Using that comparison, Cascadia would have roughly half the ridership as this other route of similar length which is generally considered very successful. That’s not that bad, considering that Madrid-Barcelona is part of a huge European rail network that benefits from a lot of transfer traffic. Many people continue on from Barcelona to France.

    Now, if we were to make this somewhat silly comparison of Madrid-Barcelona to a local bus route what happens? It’s only about 11,000 daily riders. This is less ridership than the 40 bus (12,000 on a weekday).

    So, using this metric of comparing HSR ridership to bus ridership, have we proved that Madrid to Barcelona is a bad route? That seems ridiculous.

    Let’s dig in more: What about the Japan’s Tokaido Shinkansen? Probably the highest ridership route in the world. According to wikipedia, this route has 452,000 passengers per day in 2016. That sounds good, but how does it compare to transit ridership? Tokyo metro has a ridership of 6.8 million and Toei Subway has a ridership of 2.85 million (remember, Tokyo has 2 subway systems). The Marunouchi Line alone has a ridership of 1.1 million.

    So using this metric, have we proved that the Tokaido Shinkansen is a worthless HSR route that should never have been built because it has less than half the ridership of a single tokyo subway line?

    No. That’s ridiculous. This Apples to Oranges comparison is ridiculous. Let’s move on.

  8. How much time, money, aggravation, and car-repair could be saved local drivers by getting intercity traffic out of their, meaning also my, way?

    Especially infuriating radio listening are the swarming instances where thousands of car-drivers and passengers lose entire mornings and evenings because of a mistake by a single motorist, private or paid.

    Who needs a drop of C4 for terror? All the enemy has to do is listen to KING radio, and throw in a stray “Die Infidel Dog!” while claiming credit for half a blown work- day for a whole regionload of people.

    Meantime, one very low cost measure could start saving whole months of blown public time from inception: intensify driver training by same factor as in other rail-rich places where nobody is forced into the expense and stress of our average driving day.

    Every year, based on road-test results, not book, give the highway patrol the choice of issuing either a driver’s license or a pass lawful in all fifty states, Canada, and at least Mexico. For infuriating times like these….what’s to lose by a year or two trial?

    Mark Dublin

  9. For an E. Washington run there seems to be two possible routes. Either a straight Seattle to Spokane run, or a more full featured Seattle to Yakima to Tri-City to Spokane run.

    Straight Seattle to Spokane is 279 miles by car using I-90 around 5 1/2 hours

    The Eastern dip route is around 364 miles, again by car. So an extra 85 miles. Or about 6 1/2 hours

    Seattle to Spokane = 279 miles
    Seattle to Ellensburg = 107 <<using this point as its the natural point to either go south to Yakima or east to spokane.
    Ellensburg to Yakima = 36
    Yakima to Pasco = 85
    Pasco to Spokane = 136

    Tying Eastern Washington more fully to the Western Washington is always a good idea and will certainly help clear legislative road blocks. Because lets be honest even if Olympia found a pile of free cash to build the whole BC to Portland run we know that every representative in E.Washington would still vote against it. So for HSR to be truly viable there has to be some sort of major new spending happening in E.Washington too.

      1. Interesting read. Though by my read not ambitious enough. The improvements seem to only bring train travel time in line with car travel time. If you give people the choice to drive with the car they already have or pay + deal with departure times they’ll choose their car every time or worse say screw it and get on a plane. To win people over the train must have a noticeably faster travel time than by car, at least an hour. That build out reads more like what should have been done in the 80s/90s. I’m sure its cheaper/simpler to do but nobody cares about minor improvements that still cost hundreds of millions of dollars (this is why its like pulling teeth just do maintenance work).

    1. I agree that some arrangement to serve east of the Cascades is needed for statewide support (at least as far as Ellensburg). I would also suggest developing a direct rail track with service to be within walking distance of the state Capitol. Add to that augmented and/or faster service to Bellingham and commuter service between Kelso and Portland (Columbia River Rail Crossing) and the legislature may finally be willing to dance!

      1. Yeah I figured Ellensburg would be a good stopping point if they’re cheap about it (which they will be) but enough of a build out that it creates a good foundation for a more complete line in the future that can’t be easily ignored. In comparison to a build out that never actually crosses the mountains (a token spur that ends in the foot hills) but is vaguely promised will.

    2. It would have to serve those other cities. That would give all the significant cities in Washington good rail access (every one except Wenatchee), and then their residents wouldn’t be so isolated and the cities could be part of the statewide housing and livelihood solution.

  10. There is already service from Seattle to Lakewood, on to Olympia, on to Aberdeen, into Pacific County and on to Long Beach

    1. IT 612 from Lakewood is hourly and the span of service goes up and down. Currently the last run is at 8pm but in recent years it’s been an hour earlier or weekdays only or peak hours only. A trip from downtown to Olympia takes two hours each way, as compared to one hour driving.

Comments are closed.