October Monthly Performance Report

[neroden]

(And actually, how much overhead is there to allocate?)

From memory, roughly $1.5 billion total, with roughly 1/3 to the NEC, 1/3 to the states, 1/3 to the LDs. (To get a more accurate number I would have to add it up again. The source is the most recent budget.)

This does imply that the federal operating subsidy is pretty much entirely for overhead.

 

[Anderson]

Definitely agreed, though I think we could probably produce some synthetic models. Let's assume we go with something train-mile related like you found with the LD trains...where would that put us, assuming we could separate the NEC and LD trains?

Fully allocated costs per train-mile for all of Amtrak.

(And actually, how much overhead is there to allocate?)

How many angels can dance on the head of a pin?

I'd rather not compare Amtrak accounting with theology...if you do that, theology comes off looking awfully straightforward.

Looking over those numbers, there are two things that stand out:

(1) Those trains that have heavy operation on the NEC proper (the Vermonter, VA Regionals, and Keystones) get heavily discounted. I believe this has a lot to do with the fact that while passengers (ergo train-miles) are being applied on the NEC proper, costs are being divided between those trains and the NEC. This is, for the record, my understanding of what VA negotiated.

(2) The Downeaster and Piedmont stand out as low because I believe that Amtrak does not staff food service there. In the case of the Piedmont, IIRC it's basically an automat service (i.e. vending machines), while the Downeaster's service is contracted out separately (which is why they have Coke and not Pepsi...different vendor).

That handles all of the super-cheap trains. On the expensive side, my best guess is:

-The Hoosier State is just a basket case, and it is likely getting slapped with a bunch of costs tied in with the Cardinal. Based on this number, it is no wonder IN thought someone other than Amtrak could save them money.

-The Surfliner was, I believe, loaded down with costs at some point in the 90s (from what I recall reading in a report from the passenger rail folks out in CA, the trains reported an operating profit until the 90s). It and the Cascades may also have some messy track access costs.

-The Cascades and Maple Leaf may be dealing with border complications as well. Particularly with the Maple Leaf, if Amtrak is stuck putting staff on the clock for a bunch of hours while the Border folks do their thing there's room for costing issues. On the other hand, it is also possible that the Maple Leaf line is getting hit with costs for the cafes between NYP-ALB but not getting the miles south of ALB credited to it. This would be a bug in the system...basically the inverse of the VA/VT situation.

--What stands out here is that I've seen the Maple Lead turn seasonal profits at times.

 

[Paulus]

Amtrak changed the type of accounting it billed the Surfliner on. All of the California trains are expensive, but at least some of that is due to all of the buses.

 

[neroden]

As far as I can tell, the vast majority of trains are profitable before overhead. It's hard to sort out which state trains are and which state trains aren't, though, and I've given up trying to figure it out. Without seeing direct-costs numbers for individual state trains I see no way to do it.

The sheer size of overhead does mean that, as Woody says, "the cure for what ails Amtrak is more Amtrak" -- more trains on the same routes, to spread that overhead around.

 

[twa904]

I have a couple of questions about the way info is reported.

1st on page A-2.1 Does it make sense to report Food & Beverage Revenue broken out per 100 pax miles?

2nd On Page A-3.4 Ridership on the Crescent is reported as 24,152. On page A-3.5 Sleeper Class passengers on the Crescent is reported as 2,485. My question, are the number of sleeper pax

already included in the total of 24,152?

 

[neroden]

2nd On Page A-3.4 Ridership on the Crescent is reported as 24,152. On page A-3.5 Sleeper Class passengers on the Crescent is reported as 2,485. My question, are the number of sleeper pax

already included in the total of 24,152?

The answer to this question is "yes". Can't answer your other question.

 

[afigg]

I have a couple of questions about the way info is reported.

1st on page A-2.1 Does it make sense to report Food & Beverage Revenue broken out per 100 pax miles?

Do you think the F&B revenue should be by 1000 passenger miles, a million miles, or by per mile? The per 100 mile baseline puts the values into an easy to follow range with F&B revenue of $2.85 per hundred passenger miles. Per mile, that would read $0.0285 or per 1000 miles, $28.5. So per 100 passenger miles make the F&B numbers meaningful.

 

[afigg]

I wonder how much of the increased wage expenses is due to poor OTP on the long distance trains making for a lot of OT.

No way to tell, but my wild-ass-guess is "most of it".

Digging into the October report, that does not appear to be correct. The expense breakout in the Consolidated Income Statement on page A-4.1 shows that Wages & Overtime for October were within the FY15 budget. The bulk of the increase under "Salaries, Wages and Benefits" was for Employee Benefits Expenses, $8.7 million above budget and $12.5 million over FY14. So there were unexpectedly large benefit expenses in October for pension, medical, whatever. Might have been a one time payment thing. We'll see when the November 2014 MPR drops, perhaps at the end of the 1st full workweek in January.
The November 2014 report will provide a better 2 month baseline for ridership and revenue. If November was another good month for increased system ridership, the odds of a solid FY15 for positive ridership and revenue numbers for political and press coverage are good.

 

[twa904]

I have a couple of questions about the way info is reported.

1st on page A-2.1 Does it make sense to report Food & Beverage Revenue broken out per 100 pax miles?

Do you think the F&B revenue should be by 1000 passenger miles, a million miles, or by per mile? The per 100 mile baseline puts the values into an easy to follow range with F&B revenue of $2.85 per hundred passenger miles. Per mile, that would read $0.0285 or per 1000 miles, $28.5. So per 100 passenger miles make the F&B numbers meaningful.

Food & Beverage Revenue is just that. I'm not saying it should broken out any differently. But why break it out in miles at all?

 

[jis]

I wish that they were able to publish everything in terms of seat miles or passenger miles. AFAICT they can't at present quite figure out the real CASM on a train by train basis, let alone F&B. RASM seems to be something they have a handle on on a per train basis.

 

[neroden]

The problem is that CASM *isn't* per seat mile. Most of it's fixed overhead!

RASM is actually derived per passenger mile.

I'd prefer to see a breakdown which really separates the fixed costs and the variable costs out. I think it makes it much easier to make the case that expansion is what's needed.

"Look here, this is the fixed cost of running a reservations system, stations, and maintenance shops. Now over here, separately, is the cost of running and maintaining a train. We're already running one a day, so if we ran two a day, *those fixed costs don't change*,..."

 

[jis]

I do agree with that.

 

[Anderson]

I'll agree as well. While I know the situation is more complicated than that (i.e. adding a second train would, in some cases, add to station expenses and the like while the number of sets needed for 2x daily operation doesn't always scale perfectly vs. 1x daily), it would be nice to see that sort of a breakout.

 

[jis]

I'll agree as well. While I know the situation is more complicated than that (i.e. adding a second train would, in some cases, add to station expenses and the like while the number of sets needed for 2x daily operation doesn't always scale perfectly vs. 1x daily), it would be nice to see that sort of a breakout.

But as we discussed in the other thread, that wish is a rtwo edged sword. It may or may not be in the best interest of Amtrak as a business to make such information public. Companies in competitive environment typically rollup suitable groups of products into single account and report on that, so that they keep significant flexibility to take loss on one individual product while developing it while funding it from other cash cow. In balance considering everything I am not sure where I would (or we should) land myself (ourselves) in this spectrum of possible granularity of reporting.

 

[neroden]

Companies in competitive environment typically rollup suitable groups of products into single account and report on that, so that they keep significant flexibility to take loss on one individual product while developing it while funding it from other cash cow.

Sure, that makes sense. Unfortunately, I think loading the overhead onto the product lines is creating a different form of distortion which is causing a lot of the products to look more expensive than they are.
I may be wrong, but I think this is a good pitch: "Congress needs to subsidize the basic core underpinnings of the system, our reservations and computer system and our maintenance / repair shops, just like Congress pays for air traffic control and weather reports for air pilots, and just like the government pays for the cost of maintaining, clearing, and policing roads."

Maybe 20 years ago this wasn't a good pitch to make to Congress -- it's quite possible that the Amtrak system required subsidies well above and beyond overhead back then, I don't know. But right now, it seems like the subsidies are nearly entirely for that overhead.

If the entire long-distance segment were eliminated, it looks like there would be roughly $500 million in costs to allocate to other business lines, probably $250 million to the states and $250 million to the NEC. Ouch

The overhead allocation is deeply misleading, because it puts a target on profitable trains like Auto Train; it discourages Amtrak from investing in things like WiFi on the LSL & Silver Service.

 

[Anderson]

I'll agree on the small line items point...it's more that I'd like to see a more clear disentangling of overhead from the three ridership segments (and from "other" as well, while we're at it).

 

[jis]

I agree that there is a need to get precise info on how much and where the overhead that is allocated is coming from and what precise algorithm is being used to allocate them to BUs. This can be done without exposing too much per train info I think, unless of course there is too much per train or per route special casing going on. If that is the case then that needs to be exposed and corrected.

 

[Paulus]

I wonder how much of the increased wage expenses is due to poor OTP on the long distance trains making for a lot of OT.

No way to tell, but my wild-ass-guess is "most of it".

Digging into the October report, that does not appear to be correct. The expense breakout in the Consolidated Income Statement on page A-4.1 shows that Wages & Overtime for October were within the FY15 budget. The bulk of the increase under "Salaries, Wages and Benefits" was for Employee Benefits Expenses, $8.7 million above budget and $12.5 million over FY14. So there were unexpectedly large benefit expenses in October for pension, medical, whatever. Might have been a one time payment thing. We'll see when the November 2014 MPR drops, perhaps at the end of the 1st full workweek in January.
The November 2014 report will provide a better 2 month baseline for ridership and revenue. If November was another good month for increased system ridership, the odds of a solid FY15 for positive ridership and revenue numbers for political and press coverage are good.

Looks like it'll be a good one

November ridership and revenue reports for Amtrak California are in and boast impressive gains. The Pacific Surfliner, the second busiest intercity rail corridor in the nation, set records for the highest ridership and revenue in over a decade.

The San Joaquin, the fifth busiest corridor in the nation that travels from both Sacramento and San Francisco to Bakersfield, had the second highest ridership and revenue in over a decade.

“Intercity passenger rail is an important component of a sustainable, multimodal transportation system for today and the future,” said Caltrans Director Malcolm Dougherty. “More and more Californians are increasingly seeing train travel as a smart option.”

Traditionally, declining gas prices have had a negative impact on ridership. Not this year. Gas prices for the Thanksgiving holiday this year were just over $3 a gallon, down nearly 15 percent from last year, but ridership and revenue were up on all Amtrak California routes.

 

[neroden]

FWIW, I'm going to make a prediction.

I predict that the effect of lower gas prices will be seen only in certain regions of the country.

If it didn't make a dent in ridership in California, it shouldn't make a dent in ridership in the Northeast, including upstate NY and Vermont (where gas prices are higher than in California), and western PA. The PacNW ridership should also be unaffected by gas prices.

There may be an effect on the Chicago hub routes. We'll see. Gas is significantly cheaper there than it is in the Northeast or West Coast. It's at its cheapest in Missouri, so the Missouri River Runner would be the most likely to be affected.

Gas is particularly cheap in Missouri and Texas; also in Ohio, Indiana, and Kentucky. I might expect lighter loads on the Texas Eagle, and on the Cardinal through Indiana & Kentucky. The LSL/CL don't really get their business from Ohio & Indiana so they'll probably be fine.

(As an aside, it obviously costs quite a bit to ship gasoline from Texas to the East or West Coasts. Those price differences are larger than the difference in tax rates, and you'd expect them to be arbitraged out of existence if the transportation costs were lower.)

 

[Paulus]

Part of it may be that there is increased driving, but with it increased congestion and more people taking the train for that reason. Or, alternatively, fewer dollars are spent on work commute and more dollars are available for leisure transportation (and thus Amtrak).

(As an aside, it obviously costs quite a bit to ship gasoline from Texas to the East or West Coasts. Those price differences are larger than the difference in tax rates, and you'd expect them to be arbitraged out of existence if the transportation costs were lower.)

There's no pipeline to the PADD V states on the West Coast, so domestic oil is pretty much all California or Alaska. There's some coming by train now, but it's not that big and of course it's crude rather than refined.

 

[Paulus]

Here's the day by day breakdown for the San Joaquin and Surfliner Thanksgiving ridership numbers. Absolutely incredible growth for the Surfliner, especially on Black Friday.

Edit: Surfliner added 3,100 more Thanksgiving seats in 2014 than it did in 2013 (7,300 total extra seats) but added 6,319 more riders.

 

[Anderson]

I agree that there is a need to get precise info on how much and where the overhead that is allocated is coming from and what precise algorithm is being used to allocate them to BUs. This can be done without exposing too much per train info I think, unless of course there is too much per train or per route special casing going on. If that is the case then that needs to be exposed and corrected.

Agreed. I think you can make a case for splitting the Auto Train (which is sui generis) and the Palmetto (which has no sleepers or diner and therefore behaves a lot more like a long daytime corridor train such as the Vermonter) from the other 13.

 

[neroden]

(As an aside, it obviously costs quite a bit to ship gasoline from Texas to the East or West Coasts. Those price differences are larger than the difference in tax rates, and you'd expect them to be arbitraged out of existence if the transportation costs were lower.)

There's no pipeline to the PADD V states on the West Coast, so domestic oil is pretty much all California or Alaska. There's some coming by train now, but it's not that big and of course it's crude rather than refined.

A little googling tells me that it *is* expensive to get oil to the east coast (PADD I)

http://www.refinerlink.com/blog/US_PADD_Overview/

This explains why the west & east coast prices stay up when Texas and the Midwest drop.

----

Anderson, FWIW, my spreadsheet estimates (with overhead "backed out") currently give before-overhead profits of:

Auto Train -- $35.5 million

Silver Meteor -- $15.3 million

Silver Star -- $4.8 million

Palmetto -- $7.2 million

These may be wrong if overhead has been shifted around in the last couple of years.

I am increasingly of the opinion that the correct analytical structure, from a planning perspective, is actually geographical. Trains on nearly the same route behave pretty similarly, regardless of amenities (even if the amenity is "bring your car"). Segregation into "commuter rail", "corridor", and "long-distance" is artificial and confusing.

Spreadsheet also guesses profits for:

Lake Shore Limited -- $3.7 million

Empire Builder -- $0.7 million (this is probably wrong, remember the error level in these estimates)

The other long-distance trains are still losing money before overhead allocation. (Cardinal should be profitable if daily; Capitol Limited may be profitable if connecting revenue is counted, and should be profitable with Pennsylvanian through cars.)

If there's some desire to cross-subsidize "emerging" routes from the profits from successful routes, you probably don't want to split the Auto Train and Palmetto out separately, it'll make the Crescent (for example) more of a target. The Auto Train is actually getting to the point where it can cover more than its allocated overhead.

(Even without the Auto Train, it looks like the Florida & LSL profits cover the avoidable costs of the CL, Cardinal, Crescent, CONO, and TE. But there's that pesky overhead.)

 

[afigg]

FWIW, I'm going to make a prediction.

I predict that the effect of lower gas prices will be seen only in certain regions of the country.

If it didn't make a dent in ridership in California, it shouldn't make a dent in ridership in the Northeast, including upstate NY and Vermont (where gas prices are higher than in California), and western PA. The PacNW ridership should also be unaffected by gas prices.

There may be an effect on the Chicago hub routes. We'll see. Gas is significantly cheaper there than it is in the Northeast or West Coast. It's at its cheapest in Missouri, so the Missouri River Runner would be the most likely to be affected.

The big drop in gasoline prices (below the price ranges of the past ~3 years) did not really begin until October. The Thanksgiving ridership on the CA corridor (and NEC) may have a number of passengers who made their plans and brought their tickets before the price drop. There is a lag factor to consider, so the CA Thanksgiving ridership numbers may be too early for the effect of the price drop for oil to show up. We will have to wait a few months and see what happens.
Still, the big numbers for the CA corridors are encouraging that the ridership weakness of the past year or two in CA is over. If the CA corridors and NEC ridership hold up and have a good year, they are a big enough piece of the system, that the system ridership and revenue numbers for FY15 will hold up.

As for the Chicago hub routes, the track work and consequent delays that are going to occur next Spring and Summer are going to continue to likely hold down ridership for the IL and MI corridors. Add in low gasoline prices for at least few months and competition from bus services, solid growth for the Chicago corridor routes may have to wait until FY17.

The good news for the LD trains is that overall OTP has been better since October. But I expect that it will take a while for the CL and LSL ridership to fully recover from the NS meltdown as well as the EB.

 

[neroden]

The good news for the LD trains is that overall OTP has been better since October. But I expect that it will take a while for the CL and LSL ridership to fully recover from the NS meltdown as well as the EB.

It'll take a while for CL ridership to recover, yes.

From what I can tell, LSL ridership seems to have barely dropped, amazingly, though I suspect the meltdown will show up in the form of lower revenue and lower ticket yields. I think there's huge suppressed demand along the LSL route. Rochester, Syracuse, Schenectady, and Albany platforms are all starting to run up against hard limits as to how much they can be lengthened, so.... second frequency per day, I think! Or divert the through traffic onto the "Broadway Limited", which needs to come back (if only in the form of CL-Pennsy through cars).