House Passes 2015 THUD ... $1.4 billion for Amtrak

[wolfspirit]

SEPTA shows a huge increase for the next FY for trackage rights, and they don't touch the expensive areas of the NEC (tunnels, etc.). I didn't poke at NJT's budget, but if SEPTA goes from $28 million in 2014 to $35 million in 2015, I would guess that other thansit agencies are having the same expense inrease (in percentage terms, at least).

 

[neroden]

Rather than looking at the monthly reports and guessing on cost shuffling between the NEC and the other business lines, I suggest that one should read the combined FY14 budget FY15 Budget justification and FY13-18 Five Year Financial Plan. The FY14 budget specifies the budgeted allocation of direct and shared costs in detail for the NEC (table Exhibit 4-2), State supported corridors (Exhibit 5-2), and the LD trains (Exhibit 6-2).

There's some very questionable allocations there. Everyone who said the LD trains were being assigned too much overhead is clearly correct.

* There is $37.5 million allocated to "MoW Support" for the long-distance trains, which have practically no right of way of their own to maintain. (Post Road Branch outside Albany, I suppose, but that cannot account for $37.5 million.) By contrast, the NEC only gets assigned $93 million, and the states (which actually have right-of-way to maintain) a mere $38 million. This is an obvious and gross misallocation.

* The LD business line is assigned $43 million for data processing services (IT), vs. $28 million for the state corridors and $25 million for the NEC, which seems obviously bogus to me.

* Likewise, the LD business line are assigned $81.2 million in "other general and administrative", vs. $64.7 million for the states and $50.3 million for the NEC, which is again obviously bogus.

* MoE Supervision Training and Overhead assigns *$114.4* million to the LD trains, versus $71.4 million to the states, and $49.9 million to the NEC. This is sort of plausible now given the wide variety of equipment (including Heritage cars) n the LD trains versus the uniform fleet on the NEC -- but it's completely unbelievable in 2018, when the Heritage cars will be gone and the state-supported fleet will be more motley than the LD fleet, and the plan continues to overload the LD division with this overhead in 2018.

There are others where I find the allocation suspicious ("Communication", "Police/Environmental and Safety"), etc., but these four, I simply don't believe at all.

At least $100 million in overhead should be reallocated away from the long-distance business line, arguably much more.

There are some reasons for this misallocation: according to the same document, the LD business line is currently assigned the job of operating major terminals outside the NEC (Chicago included), crew bases, service and inspection, etc. Clearly they are not charging the state corridor line enough for this.

It gets worse: in that document, Amtrak apparently proposes to increase the overhead artificially loaded onto the LD line by $71 million by 2018. So although the trains are projected to improve their *actual* performance by $34 million over the same period... yeah, not cool, Amtrak. My best guess there is that Amtrak didn't actually do any real projections for overhead allocation, and just used some arbitrary inflation percentage or something.

Looking at that document, it's actually even worse than I thought. While the long-distance trains were saddled with only $439.7 million in overhead in 2012, they have been artificially saddled with $507.7 million in overhead in 2014.

The people who say that Amtrak is sandbagging the long-distance trains with bad accounting are correct. This mostly hurts the eastern trains, since the western trains still require large yearly subsidies.

The 2014 overhead number (even though it was a budget estimate) allowed me to redo my spreadsheet on the long-distance trains. Assuming that this is the correct 2014 overhead number, and that the proportions allocated to each individual route have remained the same (again a dubious assumption), I find that the true profits (based on direct costs) of each route are:

Silver Star $4.8 million PROFIT
Cardinal ($2.8 million loss)
Silver Meteor $15.3 million PROFIT
Empire Builder $0.7 million PROFIT
Capitol Limited ($4.9 million loss)
California Zephyr ($16.4 million loss)
Southwest Chief ($16.9 million loss)
City of New Orleans ($6.8 million loss)
Texas Eagle ($8.1 million loss)
Sunset Limited ($16.3 million loss)
Coast Starlight ($9.4 million loss)
Lake Shore Limited $3.7 million PROFIT
Palmetto $7.2 million PROFIT
Crescent ($7.3 million loss)
Auto Train $35.5 million PROFIT
Let me make this clear again: cancelling Auto Train, LSL, Palmetto, Silver Meteor, or Silver Star would be bad for Amtrak's bottom line. They clearly make enough to cover all the avoidable costs which are listed in "shared costs". All of these make enough to cover their yard operations, easily. The Star/Meteor/Palmetto also clearly make enough to cover the cost of the stations shared only by them, so even cancelling all three of them would be *bad for Amtrak's bottom line*. (The LSL shares most of its stations with corridor trains, and can surely cover the cost of the other 9 stations shared with the Capitol Limited, since only 3 are staffed. The Auto Train has no shared stations.)
If cancelling a train is bad for the bottom line, that train is profitable by any sane standard.

Another way you know that the "fully allocated" numbers are bogus is that "NEC Special trains" are repeatedly listed as having a large cost per rider. Obviously Amtrak wouldn't run them at all if they weren't profitable, so they are profitable. This is just stupid overhead misallocation.

----

FWIW, I anticipate an improvement of about $2.4 million each for the Star, Meteor, and LSL when they get new Viewliner sleepers (a very rough guess); about $1 million for the Crescent and about $0.8 million for the Cardinal. I have no way to estimate the reduced maintenance cost from retiring the Heritage cars, but it should be substantial, and maybe it will also reduce the amount of MoE overhead allocated to these routes. Finally, a daily Cardinal should give a boost of roughly $6.5 million, which would make the Cardinal profitable before overhead as well.

On other points, it looks like a daily Sunset Limited would cost $14 million to run vs. $16.3 million currently for three-a-week -- it would likely do better financially than the SWC or the CZ. In short, daily service would be good for the bottom line. However, cancelling the Sunset Limited would also be good for the bottom line, so there's that.

Also, the Empire Builder appears to be doing much better financially than you'd expect (given the massive delays and crashing ridership and so forth). This is probably due to recovering substantial penalties from BNSF for the delays; it looks to me as if BNSF is actually fully compensating Amtrak for the lost revenue, and perhaps more. This might explain why CN is the one getting the STB complaints, while Amtrak is being very polite to BNSF... the delays on the Empire Builder may actually be good for Amtrak's bottom line, while the delays on CN are bad for it.

----

Regarding the capital projects allocated to the long distance business line, they're broken down by route on page 85, but the allocation is a complete head-scratcher. The largest amount is allocated to the Star ($131.1)/Meteor($135.7)/Palmetto($101.4), for a total of $368.5 million. The EB ($125.1), CZ ($126.4) and SWC ($113.3) also get large amounts. The least goes to the Cardinal, with $18.8 million, followed by the Crescent with $49.2 million. The LSL is allocated $77.2 million, the Coast Starlight $72.5 million. I don't get it at all.

 

[neroden]

If you went strictly by ridership numbers, the LD trains would get about 1/6-1/7 of the reservation system. If you go by share of revenue for something, it's about 1/4. If you go by passenger miles, I don't recall but I think it would put a bit more towards the LD system...and I don't recall where train miles (or car miles) puts you. And on the other end, if you went by train frequencies I think LD would get about 10% of the overhead.

FWIW, Amtrak is supposed to allocate according to how much of the work of the allocated department is used by that line of business. Now, sleeper bookings may require more complicated IT work than other bookings, so I could imagine allocating a disproportionately high amount to the LD trains.

But the fact is that they're getting allocated 45% of Amtrak's IT costs, which is obviously wrong -- that's *way* too high. That needs to be reallocated back to the NEC and the state corrdors.

----

If Amtrak were being run like a business (yeah, we all know it isn't, thanks to Congressional micromanagement), top priorities for the long-distance division would be:

(1) make the Cardinal daily, converting a money-loser into a profitable train

(2) deploy new Viewliner sleeping cars for incremental profits

(3) implement efficient point-of-sale procedures in the dining cars for incremental profits; the dining cars are turning tables over more slowly than they were in the 1950s.

(4) implement WiFi on the long-distance trains for increased ticket prices & thus profits

(5) implement the Capitol Limited / Pennsylvanian through cars for incremental profits; in practice, the dominant ridership would soon be on the Pennsylvania side, and the DC section would shrink

(6) implement the PIP-proposed rescheduling of the Lake Shore Limited (& Capitol Limited), allowing LSL-Florida connections and generating incremental profits;

(7) Get the Texas Eagle onto TRE as soon as possible;

(8) evaluate restoration of a third NY-Florida train;

(9) convince Michigan to run a daily train from Chicago to Toledo via Dearborn, connecting to the LSL/CL (and generating substantial connecting revenue)

(10) Implement the daily Sunset Limited / Texas Eagle run-through plan. This was supposed to cost an additional $2.95 million/year back in 2009. If you use the substantially higher revenue base from 2013 (since 2014 was full of bustitutions), it should cost more like $0.75 million/year. In a few years, with continued revenue growth faster than cost growth, it should be positive for Amtrak's bottom line; the *hour* savings from switching to the TRE should help in this matter as well. This route now has bigger intermediate cities than the SW Chief, so daily Chicago-LA service on this route may have more growth potential and may have better operating results than the SW Chief. It also provides a 'safety' route in case the SW Chief can't be preserved.

(10) buy and deploy new Viewliner glasstop cafe/lounge/obs cars for incremental ridership & improved ticket prices. Since the Amfleet II cafes are generally considered deplorable, they are probably reducing people's willingness to pay; glasstops are a selling point. Convert the old cafes to overflow coaches.

(11) buy and deploy new Viewliner coaches for incremental ridership & profits

Micromanagement, misallocation of overhead, and mixing up the profitable Eastern and unprofitable Western trains, may prevent Amtrak from doing these things.

 

[Anderson]

IIRC, the LSL reschedule still doesn't connect with the Meteor (it would have connected back in the 90s when the Meteor left at around 1900, but now that it leaves just after 1500 that's no bueno).

As far as I can tell:
(1), (2), and (3) are doable. (4) has a reliability problem on some routes (you could probably make it work on the Silvers and possibly the LSL; the Cap is going to lose service for much of Cleveland-Cumberland, while the Western trains probably can't hold a connection for much of the route and the result is likely going to just be irate passengers when the connections go down.

On (5), actually I think DC would retain most of its ridership...the Cap is, after all, the only legal daily connection to the Florida trains (the Cardinal requires a bus transfer, which is probably a turnoff, and there's a capacity issue there to boot).

On (6), see above about the connection. To get the LSL into NYP in time to reliably connect with the Meteor right now, you'd need it there by 1200-1300. Assuming that it would run as fast as the WB LSL (call it 19 hours plus one for the time difference), you'd be departing CHI at something like 1600-1700...which gets a bit tight with the current OTP out west.

On (7) and (8), I agree. With the third FL train, you can ditch the Savannah crew base and move that down to Miami, and an extension into FL should be good for a decent number of riders. You'd probably add at least some intrastate traffic aiming for that morning time into Miami, too.

(9) I don't hold out much hope for. Folks in MI want this, but the tracks stink. That said, it seems possible you might get this to happen sometime when MI adds a train.

(10) should happen; oddly, the issue here might well be the risk of a spike in ridership "crashing" both trains with a spike in demand that Amtrak can't cover.

(10-2) and (11) are ones I agree with strongly. I believe that on such an order, VA (among others...NC, PA, NY, CT, VT, MA, and ME are other candidates) would probably go in with Amtrak for some cars, so a base order of 100 coaches and some cafes could easily swell. There are lots of options for what you could do with a cafe-type order (including having some limited revenue space...several bedrooms in one end, for example...or coach/business class-type space), and you could probably embed another 10-25 sleepers and/or bag-dorms in such an order as well. It isn't just the LD trains that are slamming into capacity issues...the Regionals are also starting to get a bit out of hand in terms of demand and pricing, and being able to redeploy Amfleets to the state corridors around NYP would allow those Amfleets to be redeployed on the NEC (adding a car or two to a bunch of Regionals, if not adding a few Regionals to the timetable near peak times).

On the NEC point, Amtrak seems to be under the illusion that they can divert everyone to the Acela...and frankly, that dog ain't gonna hunt. Some riders will spring for a $150 one-way ticket, but there are plenty of others who that prices out and Amtrak will simply divert to a bus service.

 

[neroden]

Well (heh) when I said WiFi on the long-distance trains, I was actually thinking of standardizing the offering on the Empire Corridor, Richmond corridor, etc. The LSL acts as one of the Empire Service trains, effectively, so you see my point. Just getting WiFi on the LSL from Buffalo east, the Silver Service as far south as Raleigh, the Cardinal to Charlottesville, and the Crescent to Atlanta would be a big deal. The sections of the route without WiFi could be noted in the schedule.

 

[Anderson]

Well (heh) when I said WiFi on the long-distance trains, I was actually thinking of standardizing the offering on the Empire Corridor, Richmond corridor, etc. The LSL acts as one of the Empire Service trains, effectively, so you see my point. Just getting WiFi on the LSL from Buffalo east, the Silver Service as far south as Raleigh, the Cardinal to Charlottesville, and the Crescent to Atlanta would be a big deal. The sections of the route without WiFi could be noted in the schedule.

True. If you just shot for the LSL, Crescent, and Silvers overall that would also be a decent improvement (and likely a modest ridership driver as well). Honestly, just going for wifi-ing up the Viewliners and remaining Amfleets should do the trick...you'd probably need to note bad wifi service on a few segments, but it wouldn't be the end of the world.

Also, the Meteor, at least, parallels the Palmetto to SAV...and the Palmetto has wifi.

 

[wolfspirit]

On the NEC point, Amtrak seems to be under the illusion that they can divert everyone to the Acela...and frankly, that dog ain't gonna hunt. Some riders will spring for a $150 one-way ticket, but there are plenty of others who that prices out and Amtrak will simply divert to a bus service.

Generally I see Acela's sold out before other NER trains (with the exception of Keystone #640, which holds the single best timeslot for a morning PHL -> NYP train). I'm not looking in the northern end of the NEC on a regular basis, but I don't see much evidence of that in the middle of the NEC. (Currently 2 trains sold out for tomorrow, 1 keystone, 1 Acela).

EDIT: I use the PHL->NYP trains on a regular basis, so those are the only ones that I have an idea of sold out status on a regular basis.

 

[Anderson]

On the NEC point, Amtrak seems to be under the illusion that they can divert everyone to the Acela...and frankly, that dog ain't gonna hunt. Some riders will spring for a $150 one-way ticket, but there are plenty of others who that prices out and Amtrak will simply divert to a bus service.

Generally I see Acela's sold out before other NER trains (with the exception of Keystone #640, which holds the single best timeslot for a morning PHL -> NYP train). I'm not looking in the northern end of the NEC on a regular basis, but I don't see much evidence of that in the middle of the NEC. (Currently 2 trains sold out for tomorrow, 1 keystone, 1 Acela).

EDIT: I use the PHL->NYP trains on a regular basis, so those are the only ones that I have an idea of sold out status on a regular basis.

That's true, but the Acelas also have half of the capacity of a Regional (302 seats [260 BC, 42 FC] for the Acela, versus 494 to 638 (for 8-10 car Regionals, respectively).

 

[afigg]

BTW, for the record, the US Senate passed the $1.1 trillion Omnibus appropriations bill, so it is headed to Obama for his signature. The FY15 funding of $1.39 billion total for Amtrak is set. Now we wait for the FY15 budget and FY15-FY19 Five Year Financial Plan to be posted to see what plans and revenue & cost allocations are changed from over the next 5 years from FY14.