Metro VanShare at Tukwila Sounder Station

Bob Pishue, of the Washington Policy Center, has a recent piece highlighting the growth and improved economics of Metro’s vanpool program. He notes that Metro’s vanpool program is now running a farebox recovery rate of 107%. Other programs in the region are also doing well. Pierce Transit’s vanpool program is at 73% and Community Transit is at 70%. All of these numbers have been trending up.

Ridership is up too, outpacing other transit modes. In the ten years to 2013, Metro’s vanpool ridership grew 96%, and the program is now the largest in the nation. All other Metro modes grew 25% in aggregate. Operating costs per rider have fallen from $3.10 in 1996 to $3.02 in 2013, a 34% decrease in real terms. Pishue credits reduced operating costs to increased ridership. While scale efficiencies may play a part by spreading the burden of administrative overhead, shorter average trip lengths (down from 27 miles to 21 miles) have also reduced mileage-related costs.

These are great numbers, although Pishue goes too far in arguing they make the case for reducing investments in bus and light rail. Martin expanded on that point in response to a previous Washington Policy Center advocacy piece in 2010. It’s interesting, however, to look at why vanpools are successful and to understand their limitations.

Vanpools provide small-scale commuter ridership in scenarios where operator costs would otherwise be prohibitively high. Operating costs are very low (because the passengers drive themselves). Ridership per platform hour is healthy (the vanpool doesn’t run at all without a minimum of five regular riders). They are very demand-responsive (once ridership falls below a threshold, the service goes away, and new routes can be added with a minimum of overhead). They can access office parking areas and other locations where scheduled Metro service can’t reach, making for more convenient passenger drop-offs.

But the capital costs per passenger are fairly high. Vans are cheaper than larger vehicles, but they make precisely one round-trip per day. Metro’s capital expenditures per vanpool rider are twice that for bus riders, offsetting some of the operating savings. Some of the higher capital costs may just reflect fleet expansion, however. The van pool fleet is relatively young because of the expanded ridership.

Vanpools are still a niche product. While Metro vanpool ridership has grown four times faster than bus ridership in the last ten years, it still accounts for less than 3% of Metro’s ridership.

Vanpools fail at short distance trips, where the time cost of getting to the vanpool is too large. The model does not extend to any sort of spontaneous trip, so they also fail at non-commuter service generally. Ridership is overwhelmingly made up of people who make the same, fairly long, journey each day at predictable times. That’s because the coordination costs of vanpool service are too high for any other sort of journey. One hint that coordination costs are high: average ridership per trip is 5.2, just above the minimum membership for a vanpool. Evidently, it’s too much effort to find extra riders. (For some years around 2005, average ridership was below 5. Not every vanpool member rides every day, so ridership can be less than the membership limit).

The model does not scale well to higher ridership environments. Although vanpool capacity comfortably outruns single occupancy vehicles, they are limited in urban areas where yet higher capacity is needed. Even in the suburbs, some of the cost efficiencies are not scalable. For instance, many vanpools have pickup points at shopping malls. A grocery store is unlikely to object to five cars parked in their lot on a weekday, but they will surely object at fifty and insist the transit agency build its own lot.

Simple projection of current trends suggest that the vanpool share of transit trips will continue to grow. Regional development patterns will support growth too. As the region densifies unevenly, vanpools will be an increasingly effective means to serve commuter trips from low-density places to high-density centers. The bus and rail network will focus more on connecting dense places to each other. Bus and rail services will do so through frequent all-day service and core commuter services, and will pull back from low-volume coverage service.

44 Replies to “Vanpools are a Success Story”

  1. “Ridership is overwhelmingly made up of people who make the same, fairly long, journey each day at predictable times. That’s because the coordination costs of vanpool service are too high for any other sort of journey.”

    Traditionally, that’s always been the case, but in the modern digital age, it doesn’t necessarily need to be that way. With a properly designed smartphone app, one could browse through available vanpools and reserve a spot (either long-term or for a single trip) with the touch of a button. The app could also automate the payment process, similar to Lyft and Uber. One can even imagine a feature where someone headed to where the vanpool is going could make a dynamic request to be picked up along the way (in exchange for a slightly-higher-than normal fare).

    One could even imagine a world where Metro rents out is otherwise-idle vanpool vans on weekends to whomever is willing to pay for it.

    1. Lyft and Uber already offer exactly this with Lyft Line and Uber Pool. I don’t know how successful or widespread they are.

      1. There was an interesting piece on StreetsBlog just yesterday about some new entrants to this market.

        Key quotes:

        “Bridj, based in Boston, bills itself as “pop up” transit, and is using on-demand pick-up services to scout underserved routes. Initial fares are higher than transit ($6 a ride) but expected to fall as the service “learns” the most efficient routes and as more contenders enter the market, including Uber…”


        “Auto-oriented land patterns are not going away any time soon and there is great cost associated with the mandated on-demand transit that serves low intensity land uses. By being able to reduce that cost with help from the private sector, resources can be reallocated to the high intensity transit serving high intensity land patterns. However, there is always the fear that shiny, techy transit might greenwash a new generation of sprawl.”

      2. Currently, none of these services operate in Seattle (they prefer to start in their home city of San Francisco to iron out the bugs before expanding elsewhere). If they did, I think they could be successful. At current UberX rates, a trip between major nodes within Seattle (e.g. Ballard->U-district, Wallingford->Capitol Hill), split before 4 passengers, costs just pennies more than Metro fare (albeit without passes or transfers). With cheap enough prices, and the user of passenger cars, rather than buses, they could probably run some of Metro’s peak-express routes (e.g. 15X downtown->Market St., 76 downtown->Green Lake P&R) all-day, in both directions, and actually turn a profit.

        Of course, without the existence of regular Metro service to prime the demand for non-private-car travel in these corridors, such a service would never stand a chance.

    2. Metro has a program like this called VanShare. You’ll notice that’s the name on the vehicle in the photo.
      These will provide a fixed route service, maybe from one P&R to another P&R where this is a wider selection of routes to transfer to. It doesn’t have fixed ridership, but it just takes people who show up. I think you might have to be enrolled in the program to take it.

      For Snoqualmie Valley alternative service, Metro proposed running one or more VanShares between Duvall P&R and Woodinville P&R.

      1. Vanshare is similar to vanpools but helps people make a first/last mile(s) type connection to/from transit hubs. The roundtrips must be less than 20 miles. For example there are a ton of Vanshares that are based at King Street and Tukwila (Sounder) stations and connect folks to Sounder. For example a friend of mine that worked on Harbor Island took Sounder from Auburn and then a vanshare from King Street to work. You do have to sign up similar to a vanpool. The biggest difference is the length of the trip.

  2. Comparing van pools without including capital costs to transit service that does include capital costs is a very misleading premise. Does anyone know the amount of Metro’s per-hour service cost that is allocated to capital expenses? Reduce Metro’s operating cost by that amount and I’m sure that some, perhaps many, transit routes will be profitable.

    1. I’m fairly certain that the answer to your question is 0. The traditional way to compute farebox recovery is divide operating expenses by farebox revenue. Looking at the 2012 annual report, which, disturbingly, appears to be the most recent available, I don’t see evidence of capital expenses being charged as operating expenses.

      1. William is correct that farebox recovery statistics generally don’t include capital costs. The numbers I cited do not include capital.

        But Guy is correct in a larger sense. Vanpools are more capital intensive, and I get at this in the 5th paragraph of the article. Farebox recovery is an incomplete metric.

        It’s hard to tell how much this distorts the statistics. It certainly narrows the gap somewhat. A fair and more complete accounting would include operating costs + capital costs depreciated over time, but the numbers I’ve seen all assign capital costs to the year of expenditure.

      2. But what are the capital costs and maintenance costs included here? Eg, if tires need to be replaced that would be a maintenance cost for the bus fleet. For the vanpool fleet would that be included? Or is it capital cost? Transmission replacement? Engine overhaul?

  3. I think you have this just about right, Dan. Vanpools are a huge success story, but will remain a niche product that isn’t broadly scalable. Their best application is in dense but far-flung employment centers unable to be efficiently served by transit, such as DuPont, JBLM, REI-Kent, any of a dozen Boeing sites, etc. For urban centers, fixed-route transit is usually more efficient. In Downtown Seattle, carpool/vanpool have held a mostly steady 9% for the last 3 survey cycles, declining very slowly in relative share while growing slowly in absolute numbers.

      1. I don’t know I’d go so far as “Seattle” versus “the rest of the region”. There’s lots of successful transit service to suburban employers in this region.

    1. Zach, I speculated that the growth in vanpool/vanshare might be coming out of the declining coverage-oriented bus services.

      Looking at your numbers, it occurs to me that maybe some of it is coming out of carpools. I wonder if some carpools are scaling up to vanpools?

  4. In East Africa and similar places, vans have been the most common form of public transit for decades. Rural life doesn’t include much commuting, so service is an all-day matter.

    Usually what we used to call “jitney”. The van runs a more or less fixed route, generally without a schedule. None is really needed, since there are so many vans.

    At terminals- usually in a market area, driver waits until the van is full, then leaves- and picks up and leaves off passengers along the way.

    Most of the world is much more tolerant of close quarters than we are. I think that a very large percentage of the world’s population tend to feel uneasy the emptier things get.

    Turkish van and bus service us called “Dolmush”- from word for “Stuffed Like A Cooked Grape Leaf.”

    In this country for the fifty years of my working life, there has always been a somewhat angry academic cult- possibly a branch of the Lyndon LaRouche campaign- declaring that line haul transit is just plain wrong.

    Over recent decades, computer dispatch has improved along with computers. Uber and Lyft are only recently possible. But places that don’t like fumes and crowding wouldn’t permit Lyft- only transit.

    Also, capacity of a single light rail car do same thing to I-5 between Everett and Seattle every morning with vans instead of cars.


  5. Do van pools ever park in church parking lots? They might be empty during the week when commuters are commuting. They might scale better than a mall or supermarket lot.

    1. There are church parking lots that are advertised as park n rides. Some of them may have vanpools.

    2. Some church parking lots double as small park & rides. The issue would then be where to store the vans on Sundays.

      1. I would assume the driver would simply take it home with him and store it in his driveway.

  6. “Bus and rail services will do so through frequent all-day service and core commuter services, and will pull back from low-volume coverage service.”

    Only with dial-a-ride type vans, not with fixed-membership vanpools. “Coverage” is usually defined as all-day service in low-ridership areas, while vanpools are more suited to replacing peak expresses. But the peak expresses are full almost everywhere, so it would take five or more vans to replace one bus (assuming they fit 10). That may or may not be worth it if you can coax certain routes to it, but it would probably be the most residential routes (114) rather than the high-volume ones (Federal Way TC).

    Metro has been implementing “alternative service delivery” in the peripheral areas for a few years now. Snoqualmie Valley Transit is run by a senior center with support from the tribe and Metro. It has two routes, a shuttle from North Bend to Duvall, and a circulator in North Bend and Snoqualmie. The shuttle runs every 1 1/2 hours weekdays with a timed transfer at the Duvall end. Both end have 10 minutes reserved for deviations, while the middle part is fixed. It averages around seven people a trip according to a driver. The circulator runs hourly but without a schedule, so you either go to the business where it stops and wait until it comes, or call to find out when the next one will be there. These services are much cheaper for Metro because the vehicles are smaller, the drivers are non-union, and other organizations operate it and may pick up some of the expenses.

    I have thought that a similar service might work in the Issaquah Highlands and central Issaquah. Find somebody willing to operate a half-hourly or hourly van, and let them drive the non-peak commuters around. (The peak commuters might overwhelm the van, so they may still have to drive to the P&R.) To be an effective alternative to cars, it must run at least hourly and for a long span. 1 1/2 hours is only suitable for rural areas.

    Taxis might work as long as they’re subsidized. Limiting them to a small area like Issaquah might keep the cost down. Regular taxis and Uber and Lyft are OK for occasional trips, but at ten or fifteen trips a week they’re unaffordable.

    1. I would argue that for a community like Issaquah Highlands, simply getting a service like Uber/Lyft to operate there at all, with their regular rates (which are already 50% of what a taxi would cost) would be a huge improvement. Currently, getting any ride out there on any service (including plain-old taxis) is unreliable. Essentially, you need a driver to either have just happened to drop somebody off in the area a few minutes before, or a driver willing to deadhead all the way from downtown Bellevue to pick you up.

      A small subsidy for an Uber/Lyft driver to wait around for a fare when no other drivers are in the area would make an order of magnitude of difference, and would cost a fraction of what a fixed-route hourly bus would cost.

      Even if the service is still to expensive to use for daily commutes, the standard Uber/Lyft rates are low enough to function as an alternative for people whose car is in the shop, or have other special situations.

  7. If you have been down to TIBS lately, you probably noticed the signage advertising VanPool, and the offer of a few months’ free trial for newcomers. ST/Metro are pushing VanPool hard as a way to get more efficient use of the parking lots.

  8. I don’t know much about Vanpools, and who takes them, and why, so I thought I would take a look at the Vanpool Riders Wanted page. It looks like most vanpool riders could take regular transit to and from work if they wanted to. Most work regular hours, and their to and from locations aren’t obscure and out of the way. Example: Origin, Bellevue. Destination, Seattle, SLU. Work hours, 9 am to 6 pm.

    1. While a lot of vanpool riders could take regular transit to work, in many cases, the regular transit option would take considerably longer. Even in the case of Bellevue->SLU, the point of origin in Bellevue might be somewhere not directly served by the 550. Or, even if it is, the vanpool offers the advantage of a guaranteed seat for a ride that, on the 550, would probably be standing. Some people prefer the guaranteed seat and the one-seat-ride to the front door of the office over the flexibility that regular transit offers to show up whenever you want, without needing to adhere to a fixed schedule. I am personally not one of those people, but, nevertheless, such people do exist.

      1. In many cases, regular transit would take longer for the vanpool rider. So vanpool is a considerable convenience to them. Then it’s a premium service. Is that reflected in the rider’s cost? I don’t think it is. I just did a quick calculation, and a round-trip vanpool rider from Bellevue to Seattle pays LESS than the same person taking the bus. Shouldn’t they being paying more than bus riders?

      2. … and if the 550 doesn’t go directly to the exact point of origin in Bellevue, it’s a three-seat ride.

        But the fact that there’s regular transit service covering the trip makes the vanpool a lot more usable. Have to work later than the rest of the pool one day? Going out after work? You can still get back home without the expense of a cross-lake cab ride. Regular transit in SLU is a must; vanpools are nice to have. Doesn’t have to be mutually exclusive.

        For a look at a different point on the spectrum, when I worked in Canyon Park I saw a decent number of vanpool vans around (I took lots of walks on my lunch breaks — as office parks go, it’s pretty scenic, not a destination if you’re not already in the area, but worth wandering around if you are… I often packed a lunch, but there’s a little cafe every few buildings, so you can get more variety if you go on walks). There’s at least some fixed-route transit to most parts of Canyon Park. It’s not totally deserted but it wouldn’t surprise me if more people took vanpools there. If CT couldn’t sustain regular transit in Canyon Park vanpools would make this something less than a tragedy, but it would be tough to work even a little late without driving alone.

    2. Al, let’s say I live in Surrey Downs in Bellevue and work in SLU, and I work M-F, 9-5. Do you believe it’s fair that a month of Metro Vanpooling to and from work would cost me $65, but if I took the route 550, it would cost me $90/month? Why are vanpoolers paying less for a more convenient, pleasurable and premium subsidized public transit service?

      1. Good work finding out the fair discrepancy, Sam! In some ways, vanpools are more convenient, but in others not – you’re limited to one round trip a day at a fixed time.

        But the real question is about subsidies, and it’s questionable how much vanpools are subsidized. Any subsidies would have to be going to capital costs. I’d love to see numbers for those; perhaps they’re higher than the average overall subsidy for bus routes, but maybe not.

      2. Sam, maybe this sort of troll works for people that are interested in defending the logic of Metro and ST fare policies, but that ain’t me. There’s a lot that’s illogical about fares around here, and I mostly don’t give a hoot. This might not even be all that illogical — operating costs seem to be pretty low because Metro doesn’t have to hire drivers. Maybe if the fare is low and it’s convenient and it’s not expensive that means it’s just a good project.

        But I don’t really have a big stake in the argument. If vanpools were the steal of the century we’d see an armada of vans on I-90 instead of packed 550 buses.

  9. Thank you Dan, for highlighting vanpools and their contributions to mobility around the region. By the way, this is Bob Pishue.

    Although service levels of light rail and buses are a conversation for another day, I didn’t talk about reducing spending on bus and fixed rail transit.

    What I recommend is saturating the vanpool market before expanding other modes that require more subsidy, and even allowing private vanpool operators the chance to compete by phasing in full cost recovery.

    Not a big deal, but just wanted to clarify.

    Good piece.

    1. You wouldn’t advise someone in business to pursue a single idea at a time, to “saturate the market” in one area before investing in another. If you gave a business that advice you’d be fired. The only reason you aren’t fired for giving that advice in policy is that your job is not to provide serious advice in this area.

      1. For Al Dimond to be genuinely serious is absurd.

        Bob does raise some important points of efficiency. Nothing wrong with listening.

      2. The point of contact between Bob’s and Al’s views is the question of whether the high-volume core routes need more service in their own corridors. Investing in vanpools first implies the core routes are doing just fine. Most of us here strongly disagree with that. There are also corridors that should have core routes but don’t yet, mainly crosstown corridors. It may be possible to replace some peripheral coverage routes and peak expresses with vanpools, but that doesn’t address off-peak trips, nor workers going to small businesses.

    2. Thanks for the kind words, Bob.

      I’ll agree that low-cost transit is a generally good thing. The more we embrace opportunities to provide transit at low cost, the more transit we can afford to provide. So I see vanpools/vanshare as a highly valuable mode at the low-volume and commuter-oriented end of the market.

      Where we perhaps disagree is about how to address the higher-volume and/or less commuter-oriented segments. Vanpools have contributed disproportionately to improving transit’s share of trips. But they run up against natural limitations that I discussed in the article.

      Vanpools’ limitations are greatest in exactly the places where we would seek to invest in better high-capacity transit. So while they deserve more attention, they don’t really look like substitutes to me.

  10. Who would have guessed that things are efficient when you font have Metro driver wage bloat running the service.

  11. Pishue will use anything to argue for less investment in transit. Vanpools have a place, but what is the mode share? And what could you realistically raise it to? That is about all you need to know.

    1. No, it’s a more fundamental issue. You start with people’s mobility, or better yet Jarrett Walker’s “abundent access” to places. What are everybody’s desired trips, all aggregated? Which transit routes and vanpool services can serve the largest number of people, the widest cross-section of the population, and allow the most car-free households, for a given amount of money? What level should that amount be to reach a sweet spot, a robust and adequate network? Then add a factor that the most isolated houses have chosen to put themselves outside transit’s reach, so we can’t cry unduly for them or make extraordinary investments for them. Just as we can’t go without peripheral coverage service and vanpools, we can’t go overboard with them either. “80% of people drive” does not mean we should focus 80% on highways and vanpools.

  12. Vanpools and smaller buses, like King County Metro uses, make a lot of sense for all of the transit providers, other than Sound Transit (whose routes are long distance, well used routes), to be incorporating in their fleet. To continue to run near-empty buses on short distances is a waste of taxpayer’s money when a van or smaller bus can do the job far more cost-effectively, which might open up new service hours elsewhere.

    1. The thing is for Metro the primary cost of running a bus is the driver. In the cases where Metro’s labor contract allows outsourcing running a van is indeed cheaper but the marginal cost difference to Metro of a cutaway van vs. a 35′ coach, 40′ coach, or even 60′ coach is minimal when full time drivers are used.

      Past that Metro does indeed need frequent 60′ coaches just to attempt to meet demand on some routes.

Comments are closed.