Sound Transit recently released its 3rd Quarter ridership report, and it is good news all around for Link, including improved reliability, solid ridership growth, reduced operating costs, and reduced customer complaints.
Reliability: Comparison across years and across modes is a bit of a mess because each mode’s definition of reliability is different, and Central Link moved from a schedule-based reliability metric to a headway-based one this year. These metrics are defined every year in the budget.
ST Express defines “on-time” as no more than 10 minutes late. Sounder must be no more than 7 minutes late at the terminus. Central Link now defines on time as maintaining headways within three minutes of the expectation but the budget doesn’t specify a time standard.* In previous years, a train was defined as late in accordance with the internal Link schedule.
With all those warnings, here’s the chart:
Ridership: 3rd quarter ridership was up 13% over Q3 2012, breaking a record with over 2.5 million total boardings and nearly 29,000 average weekday boardings.
Operating Costs: While costs are still high, they are consistently moving downward. Link is now cheaper to operate per boarding than the overall average for ST Express, though ST Express operating costs vary widely (from $2.93 for the 550 to $11.34 on the 592).
Customer Complaints: On Link, Sound Transit receives only 1 complaint for every 143,000 boardings, down from 1 per 16,000 boardings in 2010.
The 4Q numbers should be especially interesting when they are published early next year, as that will be our first chance to quantify the effects of Ride Free Area elimination on Link’s ridership and reliability. But for now, it’s all good news.
*ST wasn’t able to get me a clarification on short notice. Spokesman Bruce Gray confirms the window.
The only way to compare operating costs between different modes is to use operating cost PER PASSENGER-MILE. Then you could compre Link to Sounder to ST Express.
So, what are ST’s figures for “operating cost per passenger-mile” for each service? That is the only meaningful operating cost number.
And, it would be interesting to get the operating cost per passenger-mile for each of the different ST Express routes.
Meh, that’s not the best metric. A useful one, taken in the context of others, but definitely not the “only meaningful operating cost number.” A bus that drives in very fast circles all day with an unchanging load of passengers would appear to be the perfect service, by that metric.
But I think the cost per passenger mile has been posted here before, frequently. Maybe you can find it in an old post? ST Express is always the predictable leader there, with it’s fleet of buses that get on the freeway and drive very far without stopping or changing passengers.
So do you then think that a bus moving 100 people in a trip from Lakewood to Seattle is more valuable than a bus that moves 1000 people per trip running between several neighborhoods in the same city?
THat’s not the question. The question is: is a bus route between Lakewood and Seattle more cost-effective than a commuter train between Lakewood and Seattle, or light rail between Lakewood and Seattle?
What good is “cost per boarding,” for anything, without adjusting for the length of the trips?
I have never seen any ST document giving cost per passenger-mile for any of their modes. If someone has seen that information, I would like to have a link to wherever it is online.
Maybe we should point out how ludicrous it is that the bus fare on route 592 (and pretty much all of the 590 series) is less than the Sounder fare when the cost of operating the bus is much much greater, and the bus contributes to congestion on I-5 as well
Do the Sounder and Link operating costs include payments to BNSF for use of the tracks and maintenance of Link’s tracks, respectively? If they do, then those costs include the costs of the right of way. The bus costs surely do not include the costs of building and maintaining the streets and highways.
“Do the Sounder and Link operating costs include payments to BNSF for use of the tracks and maintenance of Link’s tracks, respectively?”
Yes. The exact split between operating and capital costs on Sounder is a bit confusing to me, because some of the payments to BNSF were up-front and so are “capital” rather than operating, but “operating costs” definitely include the ongoing payments.
Link track maintenance is definitely an operating cost. Some transit agencies in other cities have been known to let the tracks rot until track replacement becomes a capital cost, but Sound Transit does not do that.
So the only places buses should run are freeways? Everyone needing to get up First Hill should walk? Those buses run absurdly full but perform very poorly per passenger-mile.
You have to look at platform hours because that includes deadheading. That is important. Some routes that don’t go very far but are used heavily rake in a high number of passengers and result in a cost per boarding of less than $1. Commuter routes which operate peak direction only can cost a lot more because they have to deadhead back to downtown.
If you’re going to compare buses with private cars, don’t forget that private cars often have deadheading also. Deadheading in private cars happens anytime you make a special trip to drive someone around. Every time you pick someone up at the airport, half the total miles traveled is deadheading. And every time you drive your kid to school, half the total miles traveled is also deadheading.
So, if you’re going to compare the costs of Link to the airport vs. someone else driving you, you have to include the cost of the driver deadheading back. If you also place a reasonable value on the time of the person driving you to the airport (say $10-$15 per hour), you’re well above Link’s average cost per boarding.
Long commuter routes and core all-day routes are often so different there’s little point comparing them with the same metrics. Among similar types of routes, if you’re trying to measure something about the route itself apart from locations of bases, through-routes, and other operational stuff.
For those of you who complained about my calculating the operating cost of a “small sedan” as about 10 cents per passenger-mile, becuase I did not include depreciation of the car, please note this at the bottom of the page in the ST “Third Quarter 2012 Service Delivery Quarterly Performance Report”:
“Cost per boarding is calculated as the total actual operating costs (including agency overhead) BEFORE DEPRECIATION divided by the total number of riders.”
Sound Transit does not include depreciation in the operating costs in their reports, so if you want to compare ST transit operating costs to the operating costs of a car, you should not include any depreciation in the operating cost of the car, either.
Your pedal car probably doesn’t depreciate much Norman. You can always drop it off at Salvation Army.
While I don’t have a horse in this particular race, one could argue that depreciation is a bigger concern when the vehicle’s expected lifespan is shorter. A passenger car has an average usable lifespan of a few hundred thousand miles; a transit bus can approach a million miles.
But ok, go ahead and eliminate depreciation. That only makes up about half of the AAA cost estimate. Even if you go ahead and use the “small car” numbers (which you shouldn’t, because the number of small cars actually bought in America is trivially small), it’s still well over double what you cite. Take the actual average american car as your example instead of the ultra-rare “small car”, even without depreciation, and cost-per-mile jumps to triple or quadruple your ten-cent number. So what was the source on that again?
Here is the source, often quoted by other posters here: AAA “Your Driving Costs, 2012 Edition”
http://newsroom.aaa.com//srv/htdocs/wp-content/uploads/2012/04/YourDrivingCosts2012.pdf
Small sedans in the U.S. are “ultra-rare”? lol Here are some of what the AAA use as “small sedans”: Chevrolet Cruze, Ford Focus,
Honda Civic, Nissan Sentra and Toyota Corolla. Those cars are ultra-rare? Cars like that make up a significant segment of the U.S. fleet.
Page 6 of that report gives “operating costs” for “small sedans”: 16.28 cents per VEHICLE mile. At an average of 1.6 passengers per car, that comes to about 10 cents per PASSENGER-mile for operating costs.
ST has put out documents where they show what the depreciation for Cental Link is, and the depreciation is greater than the operating costs. So, if you added in depreciation to Link’s operating costs, the total costs for Link would more than double.
They are ultra-rare. You can look at the sales chart I linked, or go dig through Car & Driver’s quarterly sales analysis on their blog. There are more Toyota Camrys (2-ton midsize sedan) sold in a quarter than Corollas, Focuses, Fits and all their competitors combined. Ditto for the 4-Runner SUV. America doesn’t buy B-segment cars.
Norman–that’s an absurdly low number. Here are my own numbers for the complete ownership of a car—these numbers certainly cheaper than what most people spend. Bottom line was $9 per day and $0.40 per mile. You’ll notice from those numbers in the link that the actual capital costs of the car was a pretty small piece of the pie.
My current car won’t ever come close to being that cheap (and I’m tracking it just as closely). Even at an average of 40 mpg, that already represents $0.10 per mile, and like the numbers in my link show, gas represents a small fraction of the total cost.
See my comment above. I am talking only about operating costs — not any ownership costs, like depreciation or interest payments on your loan.
Secondly, on average, in the U.S. there are about 1.6 passengers per car. So, you need to divide the operating costs by the number of passengers in the car, on average.
Thirdly, I don’t have to pay to park.
If you are not paying to park, then someone is subsidizing your parking.
Parking costs money to build and maintain. The land has value and should generate property taxes too. If you aren’t paying for it, someone else is. It’s a subsidy that causes you to understate the cost.
Parking costs stupendous amounts of money to maintain, unless you’re parking on mud lots. This very blog has documented the costs of parking. So who’s subsidizing that parking?
Another interesting metric to use in this case is the cost of ownership per day. When my wife and I were able to bike to and from work, this directly translated into owning one less car. So, in a very real sense, we were saving about $10 per day—that was the value of a car for us. That gives another metric that’s actually probably easier to compare in some ways.
Again, the costs in the ST document are only OPERATING costs. They do not include any “ownership” costs, like depreciation, or interest on the bonds ST sold to build Central Link. They are ONLY OPERATING costs.
If you can find the total cost for ST’s services, including all depreciation, servicing the bonds they sold, et. al., then we can compare that to total (ownership + operating) costs for autos.
I might enter into this debate if tou didn’t continue to dream up such wildly inaccurate numbers. You said buses are .90 per passenger mile, right? Where are you getting/calculating that?
Only my numbers make sense for my comparisons, and my use of ALL CAPS FOR EMPHASIS should be enough for you to see that I am right and you are wrong.
So you’re not going to explain your previous assertion that buses cost $.90 per passenger mile to own and operate?
The funny thing about your comment, Velo, is that Jeffrey Early’s operating costs are just about exactly what AAA came up with for “small sedans.”
From Early’s own records:
He put 63,000 miles on his car
He spent $5,124 on gas, and $5,287 on maintenance and tires, for a total of $10,411 in operating costs. This comes to about 16.5 cents per vehicle-mile for operating costs. If he had the average of 1.6 passengers in his car over the years, that would come to slightly over 10 cents per passenger-mile in operating costs for his car.
From the “KIng County Metro Transit 2009 Year End Report
April 2010”
http://metro.kingcounty.gov/am/reports/2009/2009YE-RTC.pdf
Scroll a little more than half-way down and you will find: “Transit Cost
Effectiveness
Operating Cost per
Passenger Mile
Motorbus & Trolley Bus, 2008
NTD”
Gives the operating cost per passenger-mile of Metro buses, including trolley buses, at $.80 per passenger-mile in 2008.
http://www.ntdprogram.gov/ntdprogram/pubs/profiles/2011/agency_profiles/0001.pdf
This document shows the “Operating Expense per Passenger Mile” of King County Metro buses to be $0.91 in 2011.
It also shows the operating expense per passenger-mile of the S.L.U.T. to be $3.79.
Velo, don’t confuse me with “Noman”, although whoever posted that was pretty clever.
And the $0.91 per passenger-mile for King County buses is only OPERATING costs. It does not include any depreciation or other capital costs.
You might want to take a closer look at the name of the guy who responded to your original comment…
It’s not fair to compare the per-mile cost of transit with the average per-mile cost of driving, as the actual person’s per-mile cost of driving can very tremendously depending on his/her situation.
For instance, the cost of auto insurance can very tremendously depending on your age and driving record. And the per-mile cost of insurance and the car itself varies greatly depending on how much you drive. If you don’t drive much, the cost of buying and insuring the car divided over the few miles you do drive can be gargantuan, especially if it involves the exorbitant insurance rates charged to people under 25.
And if you’re visiting from out of town, the per-mile costs of driving are also usually way higher than average. By the time you add taxes, fees, and gas, rental cars typically go at about $50-70 per day, more if don’t already have auto insurance from owning your own car back home. If you are paying $55 for a rental car, plus $5 for gas, that you drive a total of 30 miles in a day, you are experiencing a driving cost of $2.00 per mile – much higher than the AAA average and considerably higher than the per-person-per-mile operating cost of public transit as well.
So, the bottom line is, if you cut transit and tell everyone who rides it that it’s cheaper for society for them to just get in their car and drive, the people who are riding transit in the first place are often not the ones who experience average per-mile driving costs. These are people who would have to buy a whole car, insure it, and park it, to be only driven a couple times a month. Or people who would be forced into rental cars or taxis. Or people who, if the transit system didn’t exist, would be forced to cajole friends and relatives to give them rides everywhere. Anytime you have to make a special trip to drive someone somewhere, the cost per useful mile is at least twice the cost per driven mile because you drive out to where they’re being picked up, then drive back home from where they’re being dropped off.
Please do not fall for the fallacy of assuming the average driving costs are everybody’s driving costs, or even the average transit user’s driving costs.
Apologies for missing the spoof. I’ve seen you twist so many numbers in the past that I’ve pretty much discounted all of your numbers. This one looks accurate. Thanks for pointing to the source.
As noted above, Norman, you’re relying on roads and parking being paid for by “someone else”, since the train operating costs *include* the cost of track maintenance, while the car operating costs *don’t* include the cost of road maintenance.
If you’re trying to make some sort of point, you need to make apples-to-apples comparisons. There are several different possible ways to compare, depending on what question you’re asking — the way to answer the question “Is it more socially beneficial for the government to provide train service or not?” is different from the way to answer the question “Should *I* buy a car or not?” — but your numbers fail to provide the appropriate comparsion for any of the questions I can think of.
Thanks! I love statistics. It’s no surprise that the 592 has the highest per-boarding cost. Thank you Sound Transit for picking up where Pierce Transit left.
Instead of running all those 592 runs that compete with South Sounder, why not just have a shuttle route that connects DuPont to each of the train runs?
Or better yet, make a deal with Intercity Transit to fund the DuPont-to-Lakewood portion of the Olympia Express, at least for the train connections. Start building the ridership for the eventual passenger rail extension to Olympia.
So, am I the only one who recalls the blissful 10 months or so when no one indulged Norman with a response, he predictably flitted away, and we still somehow managed to have robust comment threads that didn’t require tethering ourselves to trees to ensure we stayed connected to reality? Because that was really nice.
I’ve tried to make the same point a number of times, but I can’t seem to get any traction with it. The whole thing seems to be about as productive a use of time as having a debate with my toaster.
Except you can actually WIN AN ARGUMENT with a toaster.
Oh you’re a silly goose
There’s a reason dishonest trolls are normally banned from blogs.
There’s a reason not to feed trolls. It only gets worse when every comment a troll makes is accompanied by 20 angry replies.
You’d be surprised how many car owners, while trying not work the numbers as aggressively as Norman does, have the same basic viewpoint.
Norman is good for exercising the synapses.
A lot of people might feel the emotion of “Geeze, owning this car is killing me!”, but don’t carry out the math with enough honesty to come to a logical conclusion.
People become emotionally attached.
The auto industry relies on this.
We are complicit in the process because we go along with the “gas tax is a user fee” charade.
Plus, many folks don’t look at their municipality’s budgets for road maintenance and ‘improvements’, so
the cycle continues.
Just out of curiosity, what exactly constitutes a customer complaint? A call about a late bus? An angry letter to ST about “wasted tax money”? An email notifying ST of a missing sign?
I would like to note, from the 2013 draft SIP, that the 590 is going up in cost per boarding ($7.46 -> $8.07) while South Sounder is going down in cost per boarding ($12.23 -> $10.56).
Add more trains to South Sounder and make them faster. This may require more triple-tracking. Eventually you’ll want to get rid of the 590 entirely.
When ST next raises bus fares, I hope the cash fare can be raised to $4 inter-county and $3 intra-county. I’m afraid ST will have to take the lead on differentiating between ORCA fare rates and cash fare rates.
The best way to do this, and not run into Title VI issues, is to simultaneously lower the ORCA fare by a penny. Oh, and stop charging so much for the card, since no other transit agency in the country charges more than $2 for a bus smart card. If the issue is getting agreement among the agencies, at lease get ST on record in favor of eliminating or reducing the ORCA card fee.
.
Rather than lowering the Sounder fare, let me suggest something a little more bold: the peak-direction 592 should go the way of the Metro 194. I mean, really, the only thing the peak-direction 592 does is bribe riders to leave seats empty on Sounder with the $3.50 round-trip savings. Yeah, it also serves DuPont, but the 574 or 594 could do that, or a Du Pont shuttle could be timed to meet the trains. Some shoulder 592 service could remain, if ridership warrants it, but I doubt the 594 doesn’t serve that purpose adequately.
ST already differentiates between ORCA and cash by not providing transfers. I sell a lot of ORCA cards at BTC to people by explaining that ORCA users get 2 hour transfers and pointing them to the nearby TVM. I also get a lot of folks with reduced fare permits which are actually ORCA cards. When I explain that they can load money on the card and get free transfers, most are genuinely surprised and happy to hear it.
The percentage of ORCA boardings has been steadily climbing and is up to 60% as of January 2012. (See graph and discussion on page 6)
“ST already differentiates between ORCA and cash by not providing transfers.”
That’s only true for people who transfer. People who don’t have zero financial incentive to use Orca over cash. And routes like the 592 competing with Sounder are all about getting to the north end of downtown without needing to make that transfer which is free only with Orca.
“ORCA … 60%”
I think part of people’s frustration is that this varies widely between routes. On some routes like the 71/72/73X most people use ORCA, while on other routes like the 169 almost everybody pays cash. (When I counted a couple times, only 3/30 people used ORCA.)
To tell you the truth, looking at those rather low operating costs for ST Bus and LINK, I would almost push for making the fare reflect those costs.
LINK and Express buses are premium services…as long as the slow, local services on Metro are kept low, there is no reason you couldn’t start reducing subsidies.
As far as Sounder, while I love it, that high cost is hard to digest for something that essentially does the same job as a LINK. Twice the cost, really? Why?
First, there’s what’s known as the “FRA overhead”; running trains on tracks shared with freight subjects them to FRA rules, which are antiquated and wasteful. You can google “FRA Reform” to find some information about this, but the biggest problems are the “tank” or “concrete block” requirements which require totally unnecessary weight to be added to the trains.
Second, BNSF owns the tracks and charges an access fee which is large enough to make them a profit. In contrast, the Link tracks are owned directly by Sound Transit.
Although there are other reasons, you really don’t have to go much further than that; those two effects are quite substantial.
Here’s something interesting. The 522 has similar total rider counts to the 511.
Obviously the 511 turning to the 512 on weekends affects this and the 511 probably has more parallel express service from CT. Still, with numbers like that, I think it’s time to start turning the 522 into a core regional service with strong network connections in northeast Seattle. It won’t have the reverse-peak demand of the 550 or 545, but of course neither will the 511, and the 522 really may be held back off-peak and reverse-peak by lack of network connections.