thoughts sparked by Boardman presentation

[Nathanael]

I'm not sure what to call the new thread.

Sorry, Nathanael, but I missed this post back in December.

It sparked some thoughts on mine, tho, when I did read it.

But to comment here would go completely off-Acela-topic.

Would you please post your comment to start a new thread?

(I could start one quoting you a lot, but that seems clumsy.)

(Former comment follows

Eh, just speed up the LDs until they're profitable.

OK, OK, I kid, I kid. But I've been looking pretty deeply into these for a while. The Boardman presentation with direct costs was very illuminating. A lot of the "costs" we see are actually overhead which can only be covered by expanding operations. The more state-subsidized corridors there are to spread the overhead out across, the better that will get. Meanwhile, most of the eastern long-distance trains are within striking distance of profitability on a direct-costs basis; a few targeted speed improvements and better on-time performance would do the trick.

Only some of the trains require large operating subsidies (CZ, SL, SWC), and even there it's only parts of their routes (Denver-Chicago is pretty good financially, Denver-San Francisco is horrible). Amtrak rejected breaking up routes into connected corridors in the PIPs, but frankly I think it's not crazy to break the CZ at Denver -- except that it would reveal how bad the western half does and how well the eastern half does.

There's a very real sense in which the Amtrak federal operational subsidy, and many of the state operational subsidies (as opposed to the capital subsidies) are for a very specific list of line segments -- ones with particularly slow running. I mentioned the terrible speed of the Crescent south of Atlanta in another thread, and the terrible Indianapolis-Chicago speed is well known. The mountain crossing of the Coast Starlight and both mountain crossings of the California Zephyr have bad running times compared to alternatives, the SWC has awful running times over Raton, and the Sunset Limited has awful running times for most of its distance.

If I were master-planning for Amtrak, I would be tempted to sit down and find out what routes have geometric alignment suitable for continuous 80mph+ running between major cities, and then move heaven and earth to get those lines under passenger-operator control -- and focus all service on those lines. But I suppose in some sense they're doing exactly that in the attempt to get a "South of the Lake" route out of Chicago. Hasn't really succeeded yet.

Anyway, sparked a lot of interesting thoughts though.

 

[Tokkyu40]

But then we have the problem of Amtrak not keeping track of what it costs to run any particular route, so we don't actually know which routes are making a profit and which are losing money. Amtrak simply allocates costs based on route miles and, apparently, distance from the main office in Washington.
What are the load factors on the eastern long distance? In the west they're turning away passengers. We need more cars, and preferably a second daily train at 12 hour intervals. It would be nice to have a train stop in Cleveland at 2:00 pm as well as 2:00 am.

 

[Paulus]

The direct costs given out do not actually all direct costs much less the total costs of running the train. If we go back to FY09, the last year with reported "FRA defined costs," "Total remaining direct costs," and "total non-direct costs," we find the following definitions and results:

-Total FRA Defined Costs represents Host Railroad MofW and Performance Incentives, Fuel and Power, T&E Crew, OBS and Commissary costs, Car and Locomotive maint. and Turnaround Costs, Commissions, Reservations, Call Centers, Psgr Inconvenience, and Route Stations.

-Total Remaining Direct Costs include Shared Stations, MoE Supervision and Training, Maintenance of Way, Yard Ops, Marketing and Distribution, Insurance, Terminal Payments, Procurement/Purchasing, Police/Environmental and Safety, and T&E Overhead.

-Total Non-Direct Costs includes Amtrak Infrastructure Maintenance and System costs.

Note that I've highlighted the shared stations. The eastern routes, especially the Silver Service, will appear remarkably better than the Western routes because it is entirely possible to quote "direct costs" in such a manner that the Silver Service has hardly any station costs. The Coast Starlight is the only Western train which sees significant overlap on its stations and even then it has many unique stations.

Furthermore, speaking solely of FRA defined costs, every single long distance train except for the Auto Train lost money (the Auto Train lost money when including the "total remaining direct costs").

As for the idea of expanding operations reducing the losses: Nope. You need at least $40 per train mile in revenue for one of those trains to break even on the marginal costs (a two man crewed commuter train is just above $30). But the actual results were (approximately, going off scheduled train miles with no adjustments for cancellations, FY2013)

Auto Train: $121.13 (mostly freight revenue)

Coast Starlight: $47.58

Sunset Limited: $22.17

Silver Meteor: $41.93

Silver Star: $37.38

Crescent: $34.42

Capitol Limited: $41.38

Lake Shore Limited $41.75

Cardinal: $24.03

Empire Builder: $38.43

Southwest Chief: $29.90

California Zephyr: $31.38

City of New Orleans: $34.41

Texas Eagle: $31.58

Palmetto: $31.64

The California Zephyr PIP says 1,775,800 train miles and gives a list of costs. Discounting management, stations, sales and marketing, commissions, insurance, passenger inconvenience, police, and G&A gives what should be an approximation of the marginal cost of running, $74.1 million, or $41.73 per train mile. Unless it results in significantly more than 100% gain in passenger revenue, additional frequencies on the LD routes will not reduce their financial losses.

 

[Woody]

I'm not sure what to call the new thread.

Sorry, Nathanael, but I missed this post back in December.

It sparked some thoughts on mine, tho, when I did read it.

(Former comment follows

Eh, just speed up the LDs until they're profitable.

OK, OK, I kid, I kid. But . . .

The Boardman presentation with direct costs was very illuminating.

. . .

Not to seem insatiable, but could you link to the Boardman

presentation you refer to?

Meanwhile, I'm very skeptical that simply speeding up

(or flattening out) the LDs would cure their problems.

In fact, when I look at your examples, I'm surely

unconvinced. You mention that the big subsidy hogs

include the California Zephyr, the Southwest Chief,

and the Sunset Limited. Then you say (not sure of

your source, is this nugget in Boardman's testimony?)

that the CZ does O.K. Chicago-Denver, but terrible

Denver-S.F. Bay Area. You sure?

I've heard that the CZ often sells out crossing the Rockies,

where it operates as "a land cruise" with many tourists

among its passengers. But lots of those tourists jump ship

in Salt Lake City (or at Denver on the other side).

So isn't the "terrible" section the passage thru the desert

of the Great Basin, mostly in the dark? The problem

there isn't the speed of the train, or lack thereof, but

the total lack of population centers to provide or to

receive passengers. Don't care how much faster you'd

make the CZ go here -- 110-mph running won't change

the fact that there's no city-pairs generating traffic for

the 600 miles between SLC and Reno.

Same problem, I'd say, killed the Desert Wind, running

SLC-Las Vegas-L.A. thru the Big Empty. Much the same

for the Pioneer, SLC-Boise-Portland-Seattle. Granted

SLC, Boise, and Vegas have seen solid growth since

the plug was pulled on the Pioneer and Desert Wind,

but in between them it's still empty. Restoring those

trains is very, very low on my wish list -- even if we got

them to go much faster.

The Southwest Chief suffers from a lot of empty as

well -- not just mountains around Raton, but a huge

swatch of Great Plains in western Kansas and the

eastern edge of Colorado, and then Arizona, too, is

almost empty for it, except for stops near the

Grand Canyon. Basically, it connects Chicago-

Kansas City-Albuquerque-L.A. with very little

population between those cities. The suggested

re-routing to add Wichita and Amarillo could help

a bit. Running it at 80 mph might not help at all.

The Sunset Ltd is a special case in several ways. It

remains crippled by a three-day-a-week schedule.

Normalizing to a daily train would change all its numbers,

doubling its passengers, slashing its loss per passenger,

and so forth. (But it would still lose lots of money.) It

is surely the most population-base-improved train

in the national system: New Orleans has stagnated

or lost population, but Lafayette, Houston, San Antonio,

El Paso, Tucson, and Phoenix have as a group (now

I'd wild guessing) likely doubled in population since

Amtrak was formed over 40 years ago. It needs to go

daily, of course, then it will need time for a train culture

to grow. And overlapping corridor services would

change its cost structure. Now, I'll grant that 110-mph

trains Tucson-Phoenix-L.A. would outperform the

Sunset Ltd at its current speeds. But building such

a corridor would restore a route into Phoenix and

let the Sunset Limited serve the center city instead of

remote Maricopa.

I'm looking for more of the "targeted" speed improvements

you mention. Cincinatti-Indianapolis-Chicago for sure.

Cleveland-Toledo-Chicago. And NYC-Albany-Syracuse-

Rochester-Buffalo. Taking a couple hours out of each

of those corridors would transform the Cardinal, the

Capitol Limited, the Lake Shore Limited, the Maple Leaf

and the Empire Service. Then do the Potomac Bridge

and D.C.-Richmond-Raleigh-Charlotte to improve the

Palmetto and the Silvers, especially the Silver Star thru

Raleigh, help the Carolinian, and give a little help to the

Crescent.

I'm not looking for 80 mph between Chicago and NYC

or DC. I'm looking to get the Lake Shore Ltd out of

NYC after full day, like, depart 5:40 p.m. instead of

3:40 as now. Depart Toledo at 5:55 a.m. as now, but

arrive Cicago at 7:59 a.m. to allow a full morning of

work. I want to do more than tweak the schedules,

but I don't need to rebuild the whole damn route to

go 110 pmh, or to average 80 mph. Just upgrade

a couple hundred miles here and there to take an

hour or two out of the times between significant

city pairs. Of course, in most cases, the way to get

an average of 80 mph over a coule hundred miles

is to build as Higher Speed corridor, the way Illinois

is doing St Louis-Chicago.

 

[Woody]

. . .

What are the load factors on the eastern long distance?

In the west they're turning away passengers. We need more cars. . . .

They're turning away would-be passengers in the East, for sure.

Famously, the high-priced sleepers from the Northeast to

Florida are sold out months ahead of time thru the winter

peak season.

So everybody here thinks Amtrak needs more cars. The belief

is probably not shared by a majority of Congress Assembled,

at least, not this Congress.

Lacking a few billion to throw into orders, Amtrak has been

forced to take baby steps. We are all looking forward to

the order for 130 new Viewliners, that is, 25 each of diners,

sleepers, and crew dorms, and 55 new baggage cars. That

won't mean much for capacity. But 25 sleepers, and in effect

another dozen sleeper-equivalents as the crew dorms open

up compartments for paying customers that have been

occupied by crew members, this will mean something like

a 40% increase in capacity at the high revenue end of things.

This order is NOT being paid for thru borrowing, so any

operating surplus that may appear can be invested by

Amtrak in more cars, or anything it needs.

Another 130 cars, bi-levels, are being built for Midwestern

corridors, as well as for California, and Washington State,

(to haul Talgos faster). These new cars will replace Horizons,

for the most part, best suited for corridor trains, but they

will also free up some Superliner cars so badly needed

on the Western LDs. The new cars are being paid for by

Stimulus grants to the several states. Stimulus funds

will also pay for a batch of NextGen diesel locomotives

to haul them. The new diesels will free up a number

of the older locomotives for use on the LD trains.

Amtrak is at a critical point in ordering up to 28 new

Acelas for the NEC. Depending on when the new trains

get delivered and other imponderable factors, a few

Regional trains might be replaced by older Acelas,

freeing up a good number of Amfleet cars for the

Eastern single-level LD trains.

If the Acela bids turn out to be affordable, Amtrak

might be left with loose change to order a few more

Viewliners. It has an option for another 70, after

backing down from plans to order 200 cars at once.

Even a few more Viewliners could have a major impact

on the ability to upgrade service, or expand it, like

taking the Cardinal daily.

But the biggest need is for a complete replacement

of the Amfleets and Superliners. Talking about many

hundreds of coaches and other cars, and orders

totaling billions of dollars. (The Viewliners, a mix

of complicated diners and sleepers and simpler

baggage cars, are costing on average $2.3 million

each. Figure 1,200 new coaches . . . billions.)

So if you want to see Amtrak modernized and

expanded to meet demand, write your Congress-

critter . . . or try to get him replaced. LOL.

 

[Anderson]

One issue with the Western LD trains is that they tend to have some very busy segments that cause Amtrak all sorts of fits. Going by route:
-Empire Builder: Minneapolis-Chicago
-California Zephyr: Reno-Emeryville, Grand Junction/Glenwood Springs-Denver, and Denver-Chicago
-Coast Starlight: Seattle-Portland and Sacramento-Los Angeles
-Southwest Chief: Chicago-Kansas City
-Texas Eagle: Chicago-St. Louis

In several cases, additional capacity can be added with cut-off cars, but in a lot of cases Amtrak gets stuck locking out sales to preserve through capacity. On some routes, Amtrak could probably sell a huge number of additional seats given the lack of another train, and indeed some of those pairs could probably fully support another train.

 

[Aaron]

The Sunset Ltd is a special case in several ways. It

remains crippled by a three-day-a-week schedule.

Normalizing to a daily train would change all its numbers,

doubling its passengers, slashing its loss per passenger,

and so forth. (But it would still lose lots of money.) It

is surely the most population-base-improved train

in the national system: New Orleans has stagnated

or lost population, but Lafayette, Houston, San Antonio,

El Paso, Tucson, and Phoenix have as a group (now

I'd wild guessing) likely doubled in population since

Amtrak was formed over 40 years ago.

From quick googling:

1970 2011 percent change
Lafayette 276,569 277,307 0%
Houston 1,903,191 6,087,000 220%
San Antonio 901,220 2,195,000 144%
El Paso 359,291 820,790 128%
Tucson 351,667 989,569 181%
Phoenix 1,039,807 4,263,000 310%
Total 4,833,715 14,634,677 203%Lafayette's an obvious outlier, but all of those metro areas have at least doubled in population. The group summed as a whole has actually tripled in population.
Here's another thing I noticed from that set of NARP statistics that was released recently. The Sunset Limited as a whole is the most improved LD train in terms of ridership numbers for the period 2007-2013. The Sunset went from 62,200 riders in 2007 to 100,761 in 2013, an increase of 62%. The Texas Eagle had a similar increase of 55.3% That compares to a 24.5% increase for the LD trains as a whole.

 

[Anderson]

On the Sunset's ridership: While that is true (my numbers, sourced off the Monthly Reports, are a bit different), part of that is down to the Sunset still being in "rebound mode". I'm not certain of it, but I want to say that the Sunset's OTP was awful until after Katrina. Through FY10, there's a spike of nearly 50%, which would be consistent with such a recovery. The numbers are more "normal" after FY10.

As to the population along the route, you're right. Not only that, but the feeder routes on the LA end have improved (look at the number of San Diegans that were left on A-Day vs. the number of Surfliners now) and you have meaningful transit systems that have gone in or expanded dramatically in most of those cities. In particular, if the train could get back into Phoenix, the timing PHX-LAX is very good now, and you've got good transit in both cities.

 

[Aaron]

When I was looking at the Sunset numbers, I was initially going to guess that at least some of it was due to the recent schedule change. But, you're right, growth has been flatter the last year or two. The real spike in growth came the few years before that so it's obviously attributable to something. The great improvement in OTP is the likely contributor.

Where the growth is attributable to the schedule change, though, is in specific cities. I'm looking at selected cities that have unstaffed stations and used to stop in the middle of the night but now stop in the day, like Benson or Lordsburg, and I'm seeing 50% jumps from '12 to '13. Palm Springs is in the opposite situation, and had a big drop.

Staffed stations don't seem to show the same impact. In Tucson, for example, the train used to stop in the middle of the night. Now, it's morning and evening. Yet, there is only a modest spike in ridership in '12 or '13, not far beyond what could be tracked on the existing curve. My guess is that it's because you've got a waiting room, so a bad time isn't as much of a deterrent as climbing off onto the ballast in the dark. Houston which got better time only shows a similar modest increase in '13, although it is bigger than the increase in '12. San Antonio, which got sort of shafted time wise in the rescheduling shows only a similarly modest drop in ridership.

My very uneducated conclusion from all this data about the Sunset is that it's just got to go daily. The gains could be potentially huge. I would never expect it to be the top financial performer in the system or anything, but I've got to imagine that a daily Sunset would move up from bottom 3 loss-wise to more like middle of the pack at least. But yeah, Phoenix... Considering that TUS-LAX is the top city pair by ridership, and that Maricopa-LAX only comes in 5th, it's clear that Maricopa is just not capturing the Phoenix market. Get the train going through PHX and you'd have at least double the ridership of that TUS-LAX number.

 

[Anderson]

Another few thoughts...

(1) Based on present ridership of around 100,000, a daily run of the present configuration would probably add 150-200k riders, assuming available space (bear in mind you increase frequencies by 133% by going daily, so an increase of 150% in ridership is a minimal increase in daily ridership). Supplementing this with an Ambus between Maricopa and downtown Phoenix would probably supplement this with another 10-25k if planned properly (since if it did nothing else, it would mean that the "legal" Phoenix-Los Angeles route didn't run through Flagstaff).

(1a) Connecting the Sunset to Phoenix, assuming minimal other changes to the timetable (assume the train must be at Tucson at the same time it is now and go from there, since an earlier LAX departure or a later arrival, within reason, wouldn't be the end of the world), would probably boost another 25k riders into the mix. Total ridership of 100k or so at Phoenix is quite believable, and might well be on the low end. This would be in line with Denver, Atlanta, and Minneapolis (of cities with reasonably similar profiles in terms of train service and train service), and a bit above Spokane and Reno. Notably, the times into/out of LAX would enable the Sunset to operate as a solid "business train" (it is one of the few major city pairs so timed both ways; the others are WAS-ATL, WAS-JAX, and to a lesser extent BUF-CHI and BOS-WAS via 66/67).

(2) I can't guess how the "stub train plan" would or wouldn't work. On the one hand, you'd lose at least some through traffic towards NOL at SAS, but this would be offset by the effect of the train going daily.

(2a) One question I'm not clear on: If a passenger travels LAX-SAS or some subset thereof on the 421 coach or sleeper, which train gets the ridership? Likewise, what about someone going CHI-LAX in the same sleeper? This leads to the question of how ridership would be accounted for in the "stub train plan" (i.e. Would the "Sunset" be just the SAS-NOL train, with the Eagle being an LAX-SAS-CHI train?).

(3) As to an LAX-TUS corridor train, this does seem like it would work well. Running that during the day and the Sunset at night would seem to be a good combination.

(3a) If you could substantially improve speeds LAX-TUS, there's likely a ton of ridership to be had. Phoenix-Tucson is likely a workable market for a fast train (hence the commuter studies), LAX-PSN is another plausible market, and a 7-8 hour run LAX-TUS (6-ish hours LAX-Phoenix) could probably support a pair of state-sponsored trains in addition to the Sunset.

Edit: Good point on unstaffed stations. As to SAS, the numbers there are muddled by (a) the presence of the Eagle and (b) at least some passengers who get stuck transferring between 1/2 and 21/22 because they did a bad booking or 421/422 was sold out.

 

[Pragmatist]

...

Furthermore, speaking solely of FRA defined costs, every single long distance train except for the Auto Train lost money (the Auto Train lost money when including the "total remaining direct costs").

As for the idea of expanding operations reducing the losses: Nope. You need at least $40 per train mile in revenue for one of those trains to break even on the marginal costs (a two man crewed commuter train is just above $30). But the actual results were (approximately, going off scheduled train miles with no adjustments for cancellations, FY2013)

...

Cardinal: $24.03

...

...

Even a few more Viewliners could have a major impact

on the ability to upgrade service, or expand it, like

taking the Cardinal daily.

....

Cardinal doesn't seem to be able to pull in enough business for even its current schedule. How would increasing it to daily, help in getting it to break $40 per train-mile?

 

[Trogdor]

But then we have the problem of Amtrak not keeping track of what it costs to run any particular route, so we don't actually know which routes are making a profit and which are losing money. Amtrak simply allocates costs based on route miles and, apparently, distance from the main office in Washington.

Do you have a reliable citation for this claim?

 

[jis]

But then we have the problem of Amtrak not keeping track of what it costs to run any particular route, so we don't actually know which routes are making a profit and which are losing money. Amtrak simply allocates costs based on route miles and, apparently, distance from the main office in Washington.

Would that be your personal opinion based on nothing? Or is there something substantive that you base it one, specially the last part? Or perhaps it is all just tongue firmly planted in the cheek and not to be taken seriously? I could go along with last one since I have myself made such statements in such spirit from time to time

On a more substantive matter, if one is trying to build an integrated system, trying to decide which trains to run and which not based on some per train or per route cost and cost recovery figures, is exactly the wrong way to go about it and playing straight into the hands of the cost cutters. Instead one should be looking at an overall service plan which requires some set of trains to be run and then cost out the whole plan as an unit, compare multiple such and choose the best. Based on that analysis then work out necessary cross subsidies and fund sourcing. Per train costing is used only by railroads that are working hard towards getting rid of trains.

 

[afigg]

I'm looking for more of the "targeted" speed improvements

you mention. Cincinatti-Indianapolis-Chicago for sure.

Cleveland-Toledo-Chicago. And NYC-Albany-Syracuse-

Rochester-Buffalo. Taking a couple hours out of each

of those corridors would transform the Cardinal, the

Capitol Limited, the Lake Shore Limited, the Maple Leaf

and the Empire Service. Then do the Potomac Bridge

and D.C.-Richmond-Raleigh-Charlotte to improve the

Palmetto and the Silvers, especially the Silver Star thru

Raleigh, help the Carolinian, and give a little help to the

Crescent.

The Chicago hub corridors such as Chicago-Indianapolis-Cincinnati and Chicago-Cleveland would see significant speed increases under the Midwest Regional Rail Initiative proposals. (Michigan DOT webpage with MWRRI documents). If the MWRRI plan were to be fully built out and implemented, it would be a major boost for all the Chicago LD trains. Of course, with the project cancellations and dropping out of the program by several of the Midwest Governors, the only 2 corridors that are getting the higher speed upgrades are CHI-STL and CHI-DET/MI service while CHI-Quad Cities is slated to start service as a 79 mph route in late 2015. But if the CHI-STL and CHI-DET corridor projects are seen as economically successful, that should help to revive the plans for other 90/110 mph corridors in the Midwest. Although changes in leadership in Ohio, WI, Iowa will also be necessary for any of those states to embrace passenger rail projects. Have to take a long term view with the MWRRI plans.

 

[afigg]

With regards to a daily Sunset Limited, the agreement that Amtrak signed with UP on the schedule change had a clause that Amtrak would not make a request to UP for SL daily service for 2 years after the implementation of the schedule change. That agreement was signed in February, 2012 and the SL schedule change took place in May, 2012. So the 2 years restriction will end this spring. Amtrak could resume efforts for a daily SL, although they may want to wait until an opportune time to reopen discussions with UP.

 

[Anderson]

...

Furthermore, speaking solely of FRA defined costs, every single long distance train except for the Auto Train lost money (the Auto Train lost money when including the "total remaining direct costs").

As for the idea of expanding operations reducing the losses: Nope. You need at least $40 per train mile in revenue for one of those trains to break even on the marginal costs (a two man crewed commuter train is just above $30). But the actual results were (approximately, going off scheduled train miles with no adjustments for cancellations, FY2013)

...

Cardinal: $24.03

...

...

Even a few more Viewliners could have a major impact

on the ability to upgrade service, or expand it, like

taking the Cardinal daily.

....

Cardinal doesn't seem to be able to pull in enough business for even its current schedule. How would increasing it to daily, help in getting it to break $40 per train-mile?

You have two issues:

-One is a chicken-and-egg problem: Less-than-daily trains seem to have per-train ridership cut significantly, based on the experiences of Amtrak and VIA (since less-than-daily operation limits travel options). This keeps something of a lid on ridership, etc. Daily operation should be good for boosting per-train ridership by 10-20% at a minimum.

--This is, in turn, likely suppressing PPR due to reduced demand. All the yield management in the world can't make up for a big enough mess elsewhere.

-The other issue is an equipment/capacity issue: Notwithstanding the trip I took on it, the train doesn't even have a full sleeper available. Adding a crew-dorm should be good for $2-4/mile due to added sleeper space. If you can fill another coach worth of seats, that should be good for another few dollars.

Basically, it's really hard to make a train generate a desirable amount of revenue when you don't run it daily and stick it with a nasty shortage of equipment. Daily operation with some additional equipment should get you to around $30-35/train.

 

[andersone]

how nice the Dessert Wind and Pioneer are remembered. If you read many of the TA boards, or even the ones here the number of folks who want to go to Vegas and have no options I think are significant. Personally i wan to go to Yellowstone next year, and have to rent wheels from SLC rather than getting closer on the Pioneer.

But you know if wishes were horses, then beggars would ride, or so me sainted mother used to say.

I think we are seeing progress not perfection and moving forward is not standing still.

 

[jis]

Personally i wan to go to Yellowstone next year, and have to rent wheels from SLC rather than getting closer on the Pioneer.

But you know if wishes were horses, then beggars would ride, or so me sainted mother used to say.

I think we are seeing progress not perfection and moving forward is not standing still.

If Yellowstone is where you want to go you ought to be remembering the North Coast Hiawatha rather than the Pioneer. The NCH would drop you close to the north entrance of Yellowstone in a manner of speaking, right at Livingston MT, or if you prefer the Northwest entrance instead of the North entrance then Bozeman MT, both of which were stops on the NCH.

 

[andersone]

Ah yes,,,,, I had forgotten about the Hiawatha , which replaced the train I took as a kid to Yellowstone, the North Coast. Hanging out with you all is bringing back so much of my youth,,,,

 

[Nathanael]

Not to seem insatiable, but could you link to the Boardman

presentation you refer to?

http://www.amtrak.com/ccurl/778/373/Amtrak-Covers-88-Percent-of-Operating-Costs-ATK-13-022.pdf
Page 10. It's not THAT informative, but the bar graph showing operating loss on a *direct costs* basis really does show what a huge percentage of the "costs" in the usual reports are actually allocated overhead of one sort or another. Railroads thrive on economies of scale.

 

[Nathanael]

The direct costs given out do not actually all direct costs much less the total costs of running the train. If we go back to FY09, the last year with reported "FRA defined costs," "Total remaining direct costs," and "total non-direct costs," we find the following definitions and results:

-Total FRA Defined Costs represents Host Railroad MofW and Performance Incentives, Fuel and Power, T&E Crew, OBS and Commissary costs, Car and Locomotive maint. and Turnaround Costs, Commissions, Reservations, Call Centers, Psgr Inconvenience, and Route Stations.

-Total Remaining Direct Costs include Shared Stations, MoE Supervision and Training, Maintenance of Way, Yard Ops, Marketing and Distribution, Insurance, Terminal Payments, Procurement/Purchasing, Police/Environmental and Safety, and T&E Overhead.

-Total Non-Direct Costs includes Amtrak Infrastructure Maintenance and System costs.

Thanks for digging this information up, Paulus. You don't seem to realize what this means. Amtrak's "direct costs" bargraph is presumably based on this definition of "direct costs", including "total remaining direct costs" as well as "FRA defined costs".
This means even the numbers from the bargraph include a bunch of unavoidable overhead which would still be there if you cancelled the train. (T&E overhead, MoE Supervision and Training, Shared Stations, Police/Environmental and Safety).

Even some of the "FRA defined costs" are actually unavoidable overhead (Reservations, Call Centers).

*The cost of the system is practically all in big overhead*. Running a second frequency of the LSL is cheap. (Getting CSX and NS to *agree* to it is another matter, but supposing they agreed, running a second frequency is cheap. Buying more equipment is expensive, but running it is cheap.)

Furthermore, speaking solely of FRA defined costs, every single long distance train except for the Auto Train lost money (the Auto Train lost money when including the "total remaining direct costs").

That's way back in FY09 (and as I noted, even "FRA" costs are full of unavoidable overhead). Prices and ridership have gone up. The Palmetto and Silver Star are now profitable including "total remaining direct costs", and the Auto Train is breakeven. I'm not sure which other trains are profitable on FRA-defined-costs, but presumably the LSL is at this point given the miniscule loss based on "direct costs". We know this from Boardman.

As for the idea of expanding operations reducing the losses: Nope.

Or, here in reality where we know how to analyze numbers, Yep.
Expanding operations *while maintaining the same overhead* will reduce the losses as the overhead is spread out over more trains. Of course, this requires routes which are profitable on "FRA defined costs", but several of the routes clearly already are. (For the others, they need to be more attractive so that Amtrak can charge high prices and still fill the trains -- this means faster schedules and better on-time performance, among other things.)

You're very good at finding data, but I suggest you learn how to analyze the data you find; the very data you found proves my point and disproves yours. The marginal cost of running another train along already-used track stopping at already-staffed stations is *low*. The marginal revenue from it, however, may be high.

It turns out to be very expensive to run a *small* railroad. But the economies of scale are enormous.

 

[Aaron]

With regards to a daily Sunset Limited, the agreement that Amtrak signed with UP on the schedule change had a clause that Amtrak would not make a request to UP for SL daily service for 2 years after the implementation of the schedule change. That agreement was signed in February, 2012 and the SL schedule change took place in May, 2012. So the 2 years restriction will end this spring. Amtrak could resume efforts for a daily SL, although they may want to wait until an opportune time to reopen discussions with UP.

Since UP has been doing a lot of double tracking on the Sunset route with the intent of making the whole LA-El Paso section double tracked, I assumed that the "opportune time" that Amtrak would wait for would be when the double tracking was finished. That might be why they so readily agreed to a two year moratorium on discussions, since they knew that project would take more than two more years anyway.

Problem I'm seeing is that UP isn't working consistently on this project. They keep ramping up work and then scaling it back based on their own traffic and whatever else they need to spend money on at that time. Last thing I read was that they were 70% done with that whole section of line, but that they were nearly halting the double track project to sink all of their money into trying to meet the PTC deadline. With the final completion date for this project such a moving target, Amtrak may just want to jump into negotiations whenever they feel like they've got all their own ducks in a row. I'm crossing my fingers that it's sooner rather than later, of course.

 

[Guest]

Of course, my proposal to expand frequencies presumes that you can still charge high prices and still fill the trains. Increased frequencies do allow for higher prices to be charged (this is empirically verified) and often result in a more-than-proportional increase in business, but if the demand just isn't there, it just isn't there. So you can do this on some route segments and not others.

You can do it on all of the Lake Shore Limited certainly. On the California Zephyr from Chicago to Denver. On the Empire Builder from Chicago to Minneapolis. People have listed some of the other "chokepoints" in other comments. There's enough demand on both the Sunset Limited and the Cardinal that daily service would improve the bottom line.

But you can't do it on the Crescent from Atlanta to New Orleans unless it's made much faster. You probably can't do it from Grand Junction to Salt Lake City at any speed because the demand is simply not there. The Capitol Limited seems to show persistent weakness of ridership even though the speed is OK. And I doubt you can do it on the Southwest Chief from Kansas City to Albuquerque, at least without a reroute.

There are weak segments of most routes, and I wouldn't suggest expanding service on those segments unless a decision is made to substantially upgrade them.

But the strong segments where the trains are full and pricey -- which includes pretty much the whole LSL route -- these are another matter.

 

[Nathanael]

Unless it results in significantly more than 100% gain in passenger revenue, additional frequencies on the LD routes will not reduce their financial losses.

In cases which I've looked at of "frequency doubling" from 1 to 2 per day, it usually does result in significantly more than 100% gain in passenger revenue. Unless the train is just really slow, or doesn't go to the right places, and the demand isn't there because of that.
Even the third frequency often results in more than a 50% increase.

 

[Nathanael]

In fact, when I look at your examples, I'm surely

unconvinced. You mention that the big subsidy hogs

include the California Zephyr, the Southwest Chief,

and the Sunset Limited. Then you say (not sure of

your source, is this nugget in Boardman's testimony?)

that the CZ does O.K. Chicago-Denver, but terrible

Denver-S.F. Bay Area. You sure?

(1) it usually costs about the same to go Chicago-Denver as it does to go Denver-Bay Area, when I've checked(2) in peak season, Amtrak tries to add an extra car Chicago-Denver to meet demand

(3) Chicago-Denver has a runtime of 12:15; Denver-Bay Area has a runtime of 32:05.

I've heard that the CZ often sells out crossing the Rockies,

where it operates as "a land cruise" with many tourists

among its passengers.

The PIP explains the ridership patterns.The CZ is popular

- Oakland to Reno;

- Grand Junction or Glenwood Springs to Denver;

- Denver to Chicago;

I wouldn't call the successful section "crossing the Rockies"; the bulk of the ridership is tourists going up into the Rockies from Denver, and then staying somewhere in the Rockies. And it's much less popular to go up into the Rockies from the Salt Lake City side, for some reason.

So isn't the "terrible" section the passage thru the desert

of the Great Basin, mostly in the dark?

Yes, but also the western side of the Rockies. All the way from Reno to Grand Junction. All 16:26 of it.

The problem

there isn't the speed of the train, or lack thereof, but

the total lack of population centers to provide or to

receive passengers. Don't care how much faster you'd

make the CZ go here -- 110-mph running won't change

the fact that there's no city-pairs generating traffic for

the 600 miles between SLC and Reno.

In a sense, the problem is the speed of the train. If the train's runtime were fast enough from SLC to Denver or from Denver to Reno, people would take it.
I don't think that this section can be sped up in any reasonable fashion; it's competing with airplanes, not cars, and that would require 300-mph running to be even plausible, and there just isn't enough demand to support that sort of monumental construction.

This is why I said "I kid, I kid" in my original comment. When I said "speed up the LD trains until they're profitable", I was *not* thinking of Denver-Reno (where I don't think it's feasible), I was thinking of Chicago-Denver, Chicago-New Orleans, Chicago-Minneapolis, Atlanta-New Orleans, Chicago-Pittsburgh, etc., where I think it might be possible.

I think our views are actually similar.

I do think that speeding up trains has a positive effect no matter what. Sure, there may not be much between Albuquerque and Kansas City, but if the runtime is reduced by an hour or two, then more people will take the train, *and* they'll pay more to do so, *and* it will cost less to operate -- a virtuous cycle. (The only thing more important for generating ridership and revenue is on-time performance, a constant Amtrak weak point.)

(Also, I know most people here know this, but top speed is usually not the important number; it's the low-speed sections which burn the most runtime.)