2006: Andrew Selden sees a new year to fix old problems and bad perceptions
Print This Post
Andrew Selden offers three examples of Amtrak problem areas that are not only hurting the company, but hurting America’s domestic transportation network as well. These examples highlight why Amtrak’s national long distance system is the most important part of Amtrak.
Andrew Selden
1. Facing the truth about the Northeast Corridor
By Andrew Selden
When a Northwest Airlines A-320 touches down at New York’s LaGuardia airport and docks at Gate D-8 to discharge its passengers, Northwest incurs charges from the Port Authority of New York and New Jersey, the owner of the airport. The charges include a hefty landing fee (for use of the runways and other airport facilities) charged on a per-landing basis, and gate rental (rent for use of the gate facility and Jetway at Terminal D). These fees are paid in cash on a regular basis by Northwest to the Port Authority. (Northwest pays its own employees to fly the plane and staff the “station.”) The Port Authority uses the cash to clean and maintain the terminals, plow snow from the runways, hire police, heat and cool the buildings, and make scheduled repayments due under the bonds that the Port Authority uses to borrow money periodically to build, repair and improve the buildings, runways, radar, parking ramps and other infrastructure of the airport. Uncle Sam provides traffic control for free, at taxpayer expense.
What does this have to do with Amtrak?
It is important on two levels. First, Amtrak has misrepresented for decades its corresponding costs of ownership and operation of its Northeast Corridor facilities: track, stations, dispatching, security, etc. When Amtrak and its echo chamber, the National Association of Railroad Passengers (NARP), tell Congress how “successful” the Metroliner and Acela trains are, when they glibly report that Acelas had a “profit” on “operations,” Amtrak doesn’t count its counterpart costs to Northwest’s landing fees and gate rentals at LaGuardia. Those costs are hidden elsewhere in Amtrak’s books, disguised and misrepresented as “capital” costs, somehow unrelated to the train operations they support. These trains are not charged with the costs of improvements and upkeep to the track, tunnels and bridges on the NEC, or the stations. It’s as if United Airlines, stuck in a Chapter 11 bankruptcy proceeding for the last three years, suddenly announced it was profitable again, but didn’t report that it was because they were no longer accounting for the costs of landing fees or gate rentals.
Amtrak has Congress to thank for making up the difference every year, to the tune of about a billion tax dollars a year. In addition to the billion dollar a year subsidy to the NEC, the Acela project cost the federal government about $3 billion in track, facility and new equipment costs, none of which is being charged against Acela revenues in Amtrak’s route performance reports.
United and Northwest don’t have free access to the Federal treasury, so when their costs (counting all of them) exceed their revenues for very long, it’s off to bankruptcy court. When Amtrak runs out of cash, they play “hide the bean” and run to Congress with the “lie of the year” (Glidepath to Self-Sufficiency, State of Good Repair, etc.) for more money, blaming the long distance trains for their $300 million “operating subsidy” and blowing a major smokescreen in front of the NEC and its ongoing billion-dollar-a-year “capital investment” (i.e., maintenance) needs.
Which brings us to the second major lesson to be learned from the Northwest A-320 landing at LaGuardia. Amtrak’s long-standing distinction between “operating costs” (or losses) and “capital costs” is phony to the point of being openly fraudulent, because it deceptively mischaracterizes these costs. NARP chirps “depreciation doesn’t count, it’s a ‘non-cash’ expense.” This is dead wrong.
Amtrak would have you believe, as it has fooled Congress, that its NEC trains are “successful” because they break even, more or less, on above-the-rail “operating costs,” while the long distance services “lose $300 million a year.” And the billion-a-year subsidy to the NEC? If, but only if, you force Amtrak to discuss it, they will blow fairy dust over that as “necessary for state-of-good-repair,” or “needed to correct years of under funding” or “modal fairness because aviation is subsidized.”
All this is designed to hide the fact that the NEC is a financial disaster, top to bottom, and in fact is the least-productive place in the whole country to invest capital dollars in rail service. Calling the NEC’s billion-dollar-a-year subsidy a “capital” cost (or “investment”) is a major part of that, and a fraud.
When Northwest lands its A-320 and pays the Port Authority its landing fee, is that a “capital” cost meant to “achieve a state of good repair”? Or its it an “operating cost” that impacts (a) whether that flight or that route as a whole is financially successful and worth pursuing, and (b) had better show up on Northwest’s route accounting reports?
If you are an executive of Northwest, landing fees are real costs, they consume cash (paid by voluntary customers), and if you don’t report them accurately, you get fired, or indicted (as happened at Enron and Worldcom).
At Amtrak, maintenance of way costs (their equivalent of landing fees) are reflected on the end-of-year company-wide audited financial statements, but you sure won’t find them on Acela or Metroliner performance reports.
At the Port Authority, the money borrowed to build or to improve or to maintain the runways is real cash with real costs associated with it. The landing fees might be a “capital item” to the extent they are used to repay the bonds, or they might fund an “operating expense” to the extent they are used to plow snow or pay the electric bill. But just as happens at Northwest, these costs are tightly associated with their related revenues. This is a First Principle of accounting: to match revenues to the costs incurred to produce the revenue.
Amtrak doesn’t do that, because, to make its high speed “Wondertrains” look good and to fool Congress, it keeps some of the major costs incurred to earn the revenues from Acela and other NEC trains isolated and does not charge these against NEC revenues on its route performance reports. Amtrak gets away with this only because it both owns the NEC and operates the trains there. When the Empire Builder runs on BNSFÂ’s railroad and incurs trackage fees (i.e., rent) to BNSF, Amtrak has to pay out cash to the host railroad, and those costs are charged 100% to the Empire Builder.
In the NEC, the “hidden” costs include the $800 million cost of acquiring the 20 Acela trainsets and their dedicated maintenance shops, the billion-dollar-plus cost of electrifying the 156 miles of NEC track from New Haven to Boston, and much of the billion-dollar-a-year free federal subsidy plowed into constant maintenance and further enhancement of the NEC track, structures and stations. Contrary to NARP, these are real costs and involve real cash dollars. The fact that Amtrak got its capital subsidy from Congress does not mean that Acela trains aren’t chewing up real (and cash) costs running up and down the NEC at 135 mph, or that they aren’t gradually wearing out (like every piece of equipment) and thereby consuming their cost of acquisition.
Calling these expenses “capital” costs (or any of Amtrak’s other euphemisms) doesn’t change anything.
That’s why, in the aggregate, if Amtrak were ever to charge Acela trains and the revenues they produce with their fair, but full, share of all of the expenses that Amtrak incurs to produce those revenues, including track costs, stations, and the amortization of the cost of acquiring the trains and their dedicated maintenance shops, as well as the “operating” costs of crew labor and electricity consumed, we would see a very different picture from what Amtrak tries to show the world about the NEC, and its other “high density corridors.”
And that, in turn, is why Amtrak’s senior staff managers, NARP and NEC politicians are so hysterically opposed to splitting off ownership of the track and structures of the NEC into a separate subsidiary, versus what Amtrak’s Board of Directors wants to accomplish.
Remember that Amtrak trains on the NEC ran just fine between 1971 and 1975 when the NEC was owned, maintained, and dispatched by Penn Central. The fear now is that a separate owner of the physical plant isn’t going to be fooled by Amtrak’s phony labels and propaganda – the NEC track owner is going to want to be paid rent for the use of its property, just as the Port Authority charges rent to Northwest (as landing fees and gate rentals) and BNSF charges Amtrak for the Builder. And not at Amtrak’s make-believe (or zero) rates, either, but at fair market value, or at least at the real cost to maintain the property.
Suddenly, Acelas, Metroliners, Regionals, and transient long hauls (like the Crescent, or Silver Meteor) are going to cost real money to run on the NEC, and the faster they go and the more redundant schedules there are, the more they will cost to cover higher track and infrastructure maintenance.
When that happens, the 30-year Amtrak-NARP fantasy about the “success” of the NEC will collapse in a New York minute. Economic rationality will drive economically-rational behavior on the NEC for the first time in 30 years.
Northwest Airlines doesn’t fly 747s with 38% load factors two or three times an hour to LaGuardia, in pursuit of some vague social benefit or 40-year-old marketing theory, and at the expense of the rest of Northwest’s system. And Amtrak will no longer be able to send half-empty Acelas and more than half-empty Regionals to New York two or three times an hour, either.
Rail service in the NEC will be forced to scale back to match actual consumer demand, because the costs of providing redundant, underutilized services are very real, and splitting off the NEC will make them visible, real and “operating” in nature. That is the heart of the [Amtrak Board of Directors Chairman David] Laney Plan to reform Amtrak, and is long overdue. It is also a good thing.
Related information about the cost of speed: Amtrak spends hundreds of millions of dollars a year on track maintenance in the NEC, much of it to allow Acelas to run at 135 mph for 39 miles in New Jersey and at 150 mph over all of 18 miles in Rhode Island. What does that save in trip time over a steady 125 mph? Not much. The 150 mph stunt in Rhode Island “saves” 1 minute 26 seconds. The 135 mph in New Jersey “saves” 1 minute 23 seconds. If these trains were slowed to 125 mph, passengers wouldn’t know the difference, but the taxpayers would save more than $100 million a year, plus reduced costs of excessive power consumption, and equipment maintenance.
2. There is no contest between corridors and long distance trains: long distance always wins.
By Andrew Selden
Amtrak’s year end results for its fiscal year ending September 30, 2005 are out, and paint an interesting picture of the performance of its various trains.
Empire Builder ridership was up 9% to 476,500. Revenues were up even more, to $42.1 million. Again, the Builder earned more ticket revenue than any other single train (or, pair of trains) Amtrak operates. To give just one quick comparison, look at the results of the Builder as compared to California’s Capitol Corridor (San Jose-Oakland-Sacramento), with 12 trains a day each way between Oakland and Sacramento (supplementing very busy I-80 in that segment).
The Capitol Corridor saw improved ridership (i.e., number of transactions), which totaled 1,260,000 passengers. (In the aggregate, all those riders represent less than one lane-hour per day of traffic, even by California standards, but that’s a different issue.) But, Capitol Corridor managers who (rightfully) bragged about ridership also failed to report on revenues, load factor or output, which would have offered a much more meaningful measure of performance of the services offered. The Capitols in 2005 generated about $12.7 million in revenue, and achieved a load factor of 28.4%, with about 95 passengers per train mile.
Comparison with the Empire Builder – admittedly Amtrak’s strongest performer – offers some interesting contrasts. The Builder “only” carried 476,500 riders (up about 9%), or about one-third as many as the Capitol Corridor. But it earned three and a half times more revenue ($42.1 million), had a load factor of 57.5% on a load of 189.9 passengers/train mile, and generated more than four times the output: 359.3 million revenue passenger miles versus 85.9 RPMs for the Capitol Corridor.
On one round trip a day, versus 12 on the Capitol Corridor, the Builder ran 1.8 million train miles; the Capitols about half as many.
This kind of comparison can be made all over the system. The Builder generated more transportation output than all of the Acelas combined (359.3 million RPM versus 312.2); even adding in Metroliner RPMs (because “Metroliners” pinch-hit for Acelas during the withdrawal of Acela trains last summer due to the brake system defect), their sum of 421 million RPMs only slightly exceeds the Builder. And, any two western routes combined dwarf the output of all of Amtrak’s “high speed” trains, and do so at a trivial fraction of the cost. The two New York-Florida trains, the Silver Star and Silver Meteor, combined also had higher output (427.3 million RPMs) than all of the Acelas and Metroliners combined.
AutoTrain had higher total revenue than the Empire Builder, but that includes its substantial fees for carrying passengers’ motor vehicles. For passenger ticket revenue alone, the Builder was much higher. The next closest train was the California Zephyr ($33.2 million) followed by the Southwest Chief ($32.5 million), and the Coast Starlight ($27.4 million).
All of the Surfliners combined (San Diego-Los Angeles-Santa Barbara) earned $36 million on 202 million RPMs. Two eastern corridors showed strong results, with the Empire Corridor (New York-Buffalo-Toronto) having $42.4 million and 181.6 million RPMs, and the Keystone (Harrisburg-Philadelphia-New York) earning $25.5 million on 899.6 million RPMs. The Chicago-Milwaukee Hiawathas earned just $8.5 million on 42.7 million RPMs.
These corridors can be compared to Amtrak’s weakest long distance route, the beleaguered Sunset Limited, which earned $9.4 million on 89 million RPMs. The corridors all had load factors of about 33% and 80 to 120 passengers/train mile, compared to the Sunset’s 59% load factor, and 103.5 passengers/train mile.
The only broad conclusion that we draw from these comparisons is that the long distance trains are performing very well on all the metrics that matter: production of transportation output, generation of revenue, and economic efficiency (the high load factors and utilization). And they do so on almost no capital investment (compared to the tens of billions of dollars that have gone into the NEC and other corridors). Why Amtrak reports only “ridership” and never output, efficiency or return on investment is a mystery. It is easy to infer that management is just too thick to understand its own business; or, perhaps, that they understand it all too well and realize that the corridor services would not survive hard economic scrutiny if all the real data were widely known.
3. It takes more – not less – passenger equipment to make money
By Andrew Selden
The only new equipment Amtrak has brought into the fleet in the last few years was a set of SUV-friendly auto racks for AutoTrain, and the 20 Acela trainsets. The Acela procurement allowed for redeployment or retirement of as many as 75 older Amfleet cars. But no new long distance cars have come into the fleet since the Superliner II cars in the early 1990s. The original Superliners are now much older (25 – 26 years) than much of the single level cars were that Amtrak inherited in 1971. In fact, the surviving original Superliners are within 15 years or so of the end of their useful lives for daily revenue service, but even now Amtrak has no plans to replace or supplement them.
Because of chronic equipment shortages, many trains are offering fewer seats and berths than they did 10 years ago, thus artificially limiting Amtrak’s potential revenues (except in all of the short distance corridors and NEC, where very low load factors of 30 to 45% show that way too many cars are being operated). The Empire Builder operates with one less sleeper and one less coach (in peak periods) than it used to.
Overall fleet availability numbers tell part of the story. As of October 1 (first day of the new fiscal year), Amtrak’s total fleet included 1,397 active cars (and more than 700 out of service). Of these, 441 were assigned to NEC trains, 207 to west coast corridors, and 749 to all other services.
The Midwest corridor fleet used 82 Horizon cars, and 12 Amfleet. The western corridors used 77 “California Car” bilevels, 48 “Surfliner” cars (mostly bilevels), 17 Horizon cars, and 64 Talgo cars (in Vancouver-Seattle-Portland-Eugene services).
Superliner trains used a fleet of 159 coaches of various configurations, just 111 Superliner sleepers (5 of which are unique all-bedroom cars assigned to AutoTrain), 49 Sightseer lounge cars, 57 diners, and 39 dormitory cars. Eight Santa Fe hi-level cars are still active. Seventy-one baggage cars are in use.
To add one sleeper to one western train on a regular basis would require as many as seven Superliner sleepers, to allow for cars in use, cars out of service for maintenance, and one spare to cover unexpected problems. Of the 106 Superliner sleepers configured for general use, 73 are required for minimum daily consist needs, with 86 cars available (as of October 1; the “available” number of cars fluctuates day-to-day as cars are taken out of service for maintenance, overhaul, or wreck repair). Coaches are even worse: 125 cars were available to cover a minimum daily need for 128 cars. So there really isn’t very much that can be done to add cars to existing trains, based on the existing fleet.
The reason this is a problem is something Amtrak never talks about: the load factors on western trains are nearly twice the load factor in any corridor, averaging nearly 57% versus 33% in the corridors. A long distance train is effectively “sold out” at about 65% load factor, because of its many stations along its route: a seat that is unoccupied at point A will be occupied at point B. On the western trains, every seat and berth turns over about three times on every trip. The peak loading point on the Empire Builder usually occurs around Wolf Point, Montana. With its western Superliner trains statistically nearly full, Amtrak cannot improve ridership or its own revenues significantly by selling more tickets. It can only raise prices, which it does aggressively, especially in peak travel periods. For example, the fare for two adults to travel from Portland to St. Paul on the Empire Builder in a Superliner bedroom jumps from about $650 in April to nearly $1,400 in May.
Amtrak’s failure of vision, its utter failure to understand the revenue potential of its own most effective, most productive product, the Superliner railcar and especially the Superliner sleeping car, prevents it from ordering new Superliners to expand the existing fleet, even just to replace aging cars or to add one or two cars to existing consists, much less restore or add routes. A minimum order would be for a total of about 300 new Superliners, delivered over about a three year period; a visionary order would be essentially open-ended, but anticipating delivery of more than 1,500 new Superliners over a 10 to 12 year period.
