Last month, alert readers noticed that ST’s quarterly ridership report showed a steep jump in cost per boarding.

With boardings up a bit and the number of trips essentially flat, this implied an alarming increase in ST’s unit costs. The good news is that a big chunk of this is a mere accounting illusion. The medium news is that some of it is one-time expenses. The bad news is that some of it is an increase in unit costs going forward.

First, the accounting change. Q3 expenses did jump $9.7m, or 53% from 2017 to 2018. However, $2.4m of that is simply due to a change in accounting. In September 2017, ST decided to move some agency overhead costs for the entire year 2017 from operations to the new capital projects. Therefore, that month operations got a massive credit for the entire year’s overhead charge. The chart below shows that 3Q 2017 was abnormally low, which inflates year-on-year growth.

More importantly, ST is spending an additional $1.7m on spare parts to overhaul the Link fleet, which is entering its second decade. A further $2.6m is going to the elevator/escalator problem, spending most riders would agree is long overdue.

About $2m is going to more staffing. ST’s Kimberly Reason shared some of the reasons:

  • Additional maintenance staff for the fleet refresh and future growth.
  • Operations taking responsibility for bits of Northgate Link and East Link that are already done.
  • Testing for new tunnels and prepping for the DSTT staffing to be 100% Link.
  • “Many of our craft positions entail extensive training and certification requirements, which requires early hiring.”

Part of the issue is that ST is taking on operations costs for segments that don’t have any riders yet, which raises cost per rider. The remainder is inflation, insurance, and “other expenses.”

29 Replies to “About Link Cost Per Boarding”

  1. Probably technically [OT] here, but as a citizen I wonder how much “inflation” falls in the “trade war” category.

    But for our own accounting, would like a lot of details about those elevators and escalators. Including, in plain English, all of them.

    Mark Dublin

    1. Maintaining the existing escalators in UW Station has not been cheap. That’s part of why the replacement ($) with “transit-grade escalators” (er, commercial grade escalators) and stairs is happening, to save operational cost over the long run.

      I’m pretty sure long-term operational costs were a consideration in deciding to add a stairwell between the platform and the street for U-District Station.

      Sadly, adding new elevators at SeaTac Airport Station, Tukwila International Boulevard Station, and Mt. Baker Station to have redundancy is not a cost saver, even if it is the right thing to do for ADA purposes. If those aren’t already in the works, it may take someone suing under ADA to get those needed retrofits to pencil out.

      1. @Ness A,

        Do you mean the up escalator to the mezzanine, the one to the northbound platform or the one to the southbound platform?

      2. On the TIBS topic, I wonder if it would be cheaper to just build elevators directly between the ground and the two platforms, and keep the ones where users have to transfer between elevators on the mezzanine as a back-up approach.

      3. The one from the bus bays to the mezzanine. I can’t climb stairs like those remotely fast enough for other people so if the escalator’s out, I’m detouring to the elevator.

      4. Thanks, Brent. But since I wouldn’t trust The Seattle Times even if I do exceed my article-number, here’s the very detail I’m most curious about:

        In order to save long-term maintenance and repair costs, let alone enough ill-passenger-will to possibly lose us an ST-vote….Why didn’t Sound Transit buy the correct grade of equipment?

        And, as would’ve happened to a transit driver making an equivalent mistake, fire whoever was responsible for the decision? If that’s more than one detail, put it on my bill next month.

        Mark Dublin

      5. Officially, they thought the standard-grade would be fine.

        Unofficially, they wanted to save a few million. Which they’re not spending and more putting the right ones in.

      6. “Why didn’t Sound Transit buy the correct grade of equipment?”

        That’s a good question. it’s probably in the auditor’s report or ST’s investigation. Maybe the staff were overzealous on cheapness, or didn’t realize the discount option couldn’t handle the load, or it fell through the cracks, or the vendor smooth-talked them into an inappropriate but more lucrative (for the vendor) option, or they didn’t communicate the operational environment clearly to the vendor. (Although the latter is unlikely, since the spec must have said “subway station”, and the vendor certainly would have noticed it when they installed the escalator, and the vendor must have some experience in the kinds of 20/7 loads that subways have.

      7. Poking around escalator vendor web sites demonstrates that transit-grade escalators is a well-recognized thing in the industry. They get heavy use and don’t operate in weather-safe (both in temperature and in indirect rainfall impacts) environments.

        If ST ordered the wrong escalators, it’s because of the stupidity of ST staff – pure and simple. The only saving grace is that it’s been such a debacle at the first few stations that the dozens more stations coming on line soon hopefully won’t suffer from this mistake.

      8. Sound Transit learned their lesson on escalator contracting after U-district station opened.

        Based on what I heard on a board-meeting livestream around that time, the previous escalator installations were design-build contracts, with a 1-year warranty (that started ticking at the date of installation, NOT opening day). So the contractors could under-spec everything, and a couple months after opening day wash their hands of it.
        Going forward, the escalator contracts are going to also include long-term maintenance & repair, so the contractors will actually have skin in the game regarding how the escalators perform decades into the future.

  2. I wonder how much more it would cost to move to all 3-car trains in the tunnel at the March service change. I lost track of whether there are 6 or 7 peak 2-car trains. At any rate, speeding up the total path by roughly 3-minutes each way should induce an instant bump in ridership, and enable removal of one train from the loop, freeing up a couple LRVs to be adding to two other 2-car trains.

    I’m also not sure whether “seat slides” (operators switching which train they are operating at the termini) are being employed at one or both ends already during peak. If just one, it seems like doing so at both would be an option to free up one more train (while making sure operators get break time), and get to all 3-car trains to handle the ridership bump.

    And then there will be another ridership bump next September with the roll-out of the SR 520 service restructure and removal of the Montlake flyer stop. Having a few Siemens trains, with their fewer seats and more spacious standing room, IIRC, will come in quite handy. Having a few 4-car Kinkysharyo trains could also be an option at that point, if necessary.

    At any rate, failing to have all 3-car trains at the March service change could come with a cost to reliability and loop time.

    What are people’s experiences lately of whether the peak 2-car trains are still able to handle peak traffic without extra dwell time compared to 3-car trains?

    1. Earlier this fall I saw people get left behind at SODO when a full 2-car came during morning rush! No special events, just a standard morning commute!!

      The evening peak 2-car that leaves UW around 5 is pretty bad. It’s already 3/4 full when it leaves UW. Since many people get off at Capitol Hill/Westlake there is enough room, barely.

      It’s clear that the ST strategy should be to get some peak 4-car trains in service before 2021 if possible.

      1. 4-car trains ought to be an option. But the capacity on a 3-car Siemens train might be about as much as that on a 4-car Kinkysharyo train. Standing room matters.

        Over the long run, I’d like to see the Siemenses run on the blue (south) line and the Kinkysharyos on the red (east) line to make maneuvering luggage easier.

        In the short run, running as many Siemenses as possible on weekends shouldn’t get choice commuters angry about not getting a seat, while accommodating SeaTac’s busiest travel days.

        Long-long-term, the next maintenance base ought to be built to accommodate open gangway supertrams.

      2. Possible to switch all seating to benches in the door-level section of the cars we’re using now? Also still wondering about leaving lower floor section on come cars completely seat-less for baggage and bikes.

        Though think that’d require platform staff to help load. Or worse, ride along as attendants. So more sensibly, could we “partner” with hotels, or encourage anybody else to carry baggage between trains and lodging?


    2. They can also save some time by cutting the excessive dwell time at SeaTac Station. What is this for? Luggage loading? It had never been a problem when I have been there.

      1. If both tracks at Angle Lake are occupied they’d have to wait for the next train to leave before going further south.

  3. The National Transit Database (NTD) publishes a summary of data about the 50 largest US transit agencies, which includes lots of financial data. Average cost per boarding for light rail is $4.31, so Link is right there near the typical. Heavy rail is more cost-effective, at an average $2.25 per boarding, due to longer trains, and considering the dataset, probably lots of deferred maintenance.

    If Northgate Link brings the ridership Sound Transit projects, the cost per boarding could decline near $2.25, but the opening of ST2 lines will dramatically increase operating cost and the cost per boarding will probably jump back up.

    1. If you dig around a bit, you should find NTD data for all of the public agencies, including those with just a single bus route.

    2. @Chad — Do you mean ST2 or ST3? OK, now that I think about it, probably both. East Link should be reasonably cost effective (at least initially — I do have to wonder about the long term issues with the tracks on top of the floating bridge). But I don’t think Lynnwood Link is going to be very cost effective. It is a lot of extra track, without that many new riders (most of the suburban folks will just switch from buses that end at Northgate to those that end somewhere else). I just don’t see anything substantial in terms of switching from buses to trains. Nor do I see new combinations (like Northgate to Capitol Hill) that are extremely time consuming by bus right now, but will be extremely fast once Link gets to Lynnwood. Lynnwood Link has its benefits, but better cost per rider isn’t one of them.

  4. In a logical world, the cost per rider to operate Link would be an informative way to evaluate the corridors currently under study. I would love to see how this number is projected to change as new extensions open — and if we are doing the right station and operations planning to optimize those.

    Heaven forbid that this rational measure of productivity gets any analysis and disclosure to inform leaders beyond hysterical neighbors and non-transit interests! It’s not the Seattle process!

    1. Al, that would require ST’s ridership projections to be reliable. Historically, they underestimate inner-city demand and overestimate demand in the ‘burbs.

      1. I think it is pretty obvious which prospective stations will have a factor over 2,3,4 and etc over others. Hell, I would settle for ST trusting their own numbers and using them. However this doesn’t seem to be the case for ST’s comprehensive planning.

    2. I agree with all three of you. It is a worthy thing to measure, and if you make a major extension and the cost per rider actually goes up, you have to wonder if you are doing the right thing. Particulars mentioned in this post not withstanding, ridership per mile went down with UW Link. They added a lot more riders, and didn’t have to add much in the way of extra time spent serving them. As Chad said up above, the same thing will happen with Northgate Link. They will get a lot of new riders, with a relatively small addition (an extra 7 minutes of service for three high performing stations).

      But my guess is the estimates (inflated as they are) for Lynnwood Link would still result in a decrease in cost per rider. Lynnwood to Northgate is going to take 13 minutes, and I doubt you will have that much of an increase. Even if you just look at it compared to today it doesn’t look good. We have about 80,000 riders, and a trip (one direction) takes 46 minutes. So that works out to about 1700 riders per minute*. For Lynnwood Link to compete with that, it would need to get 22,100 *new* riders. These aren’t riders that switch from Northgate to Lynnwood (which seems likely) but brand new riders to the system. I just don’t see it.

      I’m not saying it should be the only metric, or even a major one, but it should give you pause. If ridership per minute is going down, are we compensating by adding a lot of value for those riders? That could be the case (it would certainly be the case with riders at 145th and 130th). But how many riders will save how much time north of there? Of course rider time per dollar spent is another (more important) metric that ST never studied.

      * I know we don’t pick up 1,700 riders a minute. It is just the ratio that is important.

      1. This is a basic problem with embracing corridors based on new riders. It’s easy to say that a segment will add 30k or 40k riders a day and it looks wonderful and enticing to the public. However a cost per rider operating metric or subsidy per rider metric would establish if the investment is actually worth it.

        FTA looks at various rider metrics when determining grants. It’s but another reason why a metric should be disclosed.

  5. While this blog post answers some questions about the year-over-year variance for the Q3 Link “cost per boarding” metric, it raises other questions as well.

    The additional $2 million relating to staffing…..check.

    The additional $1 million or so (the “remainder”) relating to higher insurance costs, other expenses and inflation…..check.

    The “additional” $2.4 million of the total variance that was due to the accounting adjustment made in Q3 2017…..check. (Referring to this as an “accounting illusion” is kind of silly though as these types of adjusting entries happen all the time in the accounting realm. They typically are handled by a note accompanying the report, something ST failed to do here by only making some vague reference to a reconciliation in the report’s narrative.)

    The questions come into play when reference is made to the somewhat baffling claims about the $1.7 million expense for Link fleet spare parts and the rather nebulous $2.6 allocation for the “elevator/escalator problem”. This would imply that both components are being expensed instead of being capitalized, which may or may not be an appropriate accounting treatment without knowing additional details, but on its face seems to contradict the agency’s own stated policies regarding capital assets:

    “Capital Assets— Capital assets are stated at cost, except for capital assets contributed to Sound Transit, which are stated at the fair value on the date of contribution. Expenditures and contributions for additions and improvements with a value in excess of $5,000 and a useful life of more than one year are capitalized.
    Expenditures for maintenance, repairs and minor improvements are charged to operations as incurred….”*

    *Central Puget Sound Regional Transit Authority, Financial Statements and Independent Auditors’ Report for the Years Ended December 31, 2017 and 2016

    1. Spare parts should not be capitalized. The escalator costs would depend upon their nature.

      Thanks so much for following up on this. It makes a lot more sense now.

      1. Well the devil really is in the details, as they say. It’s up to the agency’s capitalization threshold policy ultimately.

        A couple notes from ST’s 2017 annual financial statements:

        “Inventory— Inventory includes spare parts and is recorded at the lower of average purchased cost, or net
        realizable value. Allowances for excess and obsolete parts are provided for spare parts currently identified as excess and obsolete. Allowances are reflected as a charge to operations and are based on
        management’s estimate that is subject to change. As of December 31, 2017 and 2016, inventory reflects
        an allowance of $902 thousand and $115 thousand, respectively.”

        Thus spare sparts even if they’re not treated as capitalized assets are still handled as a balance sheet item (other than the inventory obsolescence allowances which do hit the operating statement).

        There’s also this tidbit from the financials:

        “STATEMENTS OF CASH FLOWS, continued

        Supplemental disclosures of non-cash operating, investing and financing activities –

        Spare parts previously capitalized ($2.217M) December 31, 2017”

Comments are closed.