Poking around at Amtrak financials again

I've been poking around at the long-distance train financials again. Comparing 2012 "core contribution/loss" numbers to Boardman's presentation on a direct-costs basis, which it looks like is probably based on 2012 numbers, it looks as if the following amounts of overhead were allocated to each train:

Silver Star $38 million (44.1 - 6)

Cardinal $11 million (16.9 - 6)

Silver Meteor $38 million (36.9 + 1)

Empire Builder $45 million (56.0 - 11)

Capitol Limited $18 million (23.6 - 5.5)

California Zephyr $38 million (68.7 - 31)

Southwest Chief $38 million (65.2 - 27.5)

City of New Orleans $16 million (20.6 - 5)

Texas Eagle $19 million (33.1 - 14)

Sunset Limited $18 million (40.9 - 23)

Coast Starlight $37 million (53.8 - 17)

Lake Shore Limited $30 million (31.6 - 1.5)

Palmetto $15 million (10.9 + 4)

Crescent $33 million (40.6 - 8)

Auto Train $32 million (32.4 + 0)

I can't make much sense of this. Shared stations are part of this, but they certainly don't account for the entire pattern. It doesn't line up with rolling stock usage, either. Maybe it's partly train-mile-based allocation? And is it coincidence that the Star and Meteor have *exactly the same* allocated overhead?

Anyway, if I back out this 2012 level of allocated overhead from the August 2014 results, to try to derive direct-costs losses for YTD 2014, I find:

Silver Star $4 million surplus

Cardinal ($2 million loss)

Silver Meteor $12 million surplus

Empire Builder ($1 million loss)

Capitol Limited ($4 million loss)

California Zephyr ($18 million loss)

Southwest Chief ($17 million loss)

City of New Orleans ($6 million loss)

Texas Eagle ($9 million loss)

Sunset Limited ($15 million loss)

Coast Starlight ($10 million loss)

Lake Shore Limited $2 million surplus

Palmetto $6 million surplus

Crescent ($7 million loss)

Auto Train $31 million surplus

Now, these numbers still look close to the *right order* to me, just as the numbers in Boardman's presentation did. They seem to roughly correspond to online city size, ridership, train length, load factors, cost of operations, etc. in a way that other commonly referenced numbers do not.

(Using the August numbers will overstate the success of every train, because some of the overhead allocation will creep in in December, but the relative order should still hold, more or less. Using two-year-old overhead numbers for subtraction may understate the success of every train, because overhead has probably gone up some. Or it may overstate the success of every train because overhead may have been reallocated toward the NEC and state corridors, thanks to increased service there. Or overhead may actually have dropped. Who knows? There's also still probably something screwy in the Sunset Limited/Texas Eagle accounting. And I don't trust the Auto Train numbers because I suspect overhead allocation procedure on that has changed substantially between 2012 and 2014. However, the *relative order* of the trains is probably still *mostly* right.)

Make no mistake: there's a lot of overhead to cover. $416+ million. Some unknown amount of it is arguably "avoidable" or belongs in direct costs -- it's not clear how "shared stations" are treated, for example. Most of the overhead is not like that, though; it's unavoidable overhead. (There's also very likely to be overhead buried even in "direct costs", unfortunately.)

The entire premise of my calculation here, of course, is that allocated costs haven't changed, and that any drop in costs is a real drop in direct costs. I have no way of telling whether this is true, however, and I actually suspect that it isn't.

Revenues have been flat in most of the "long distance" department from 2012 to 2014, and actually down significantly on the Builder and Eagle (though way up on Auto Train thanks to the extra coach).

Costs for the "long distance business line" dropped by $39.8 million from 2012 to 2013, and by a further $91.6 million from August 2013 to August 2014. I am still suspicious that much of this change is overhead.

Costs in *every* business line dropped very substantially from August 2013 to August 2014, which might point to the elimination of large sources of overhead. (Delayed benefits from e-ticketing, perhaps? Such as downsizing or elimination of the center which processed the paper tickets?).

It's also worth noting that for the last several years Amtrak has been -- on a "fully allocated" basis, of course -- losing money on freight and commuter service and "ancillary customers". This "loss" is going up, so it's possible that overhead has been reallocated *there*. (Perhaps due to increased commuter service or something.)

Overhead allocation is magic and they will not share it with us simple mortals.

If they did the sun would fail to rise and the end of the world would be upon us.

At least a portion of the overhead allocation might also have to do with a desire to obfuscate the total centralized cost, by instead blaming individual trains for not so well managed central budget items. This was an old trick used on certain other railroads like one which had a P and a C in its logo. That might suggest a reason for the reluctance to generally report the way Boardman did in his slide set.

If Amtrak's accounting is like the municipal and school district budgets I used to wrestle over as a reporter, you'll never figure it out since the budgets are designed to be overly complicated. But, I'm terrible at math.

neroden said:

Costs for the "long distance business line" dropped by $39.8 million from 2012 to 2013, and by a further $91.6 million from August 2013 to August 2014. I am still suspicious that much of this change is overhead.

Costs in *every* business line dropped very substantially from August 2013 to August 2014, which might point to the elimination of large sources of overhead. (Delayed benefits from e-ticketing, perhaps? Such as downsizing or elimination of the center which processed the paper tickets?).

Click to expand...

There have been a number of changes over the past several years that might be contributing to the changes in total allocated costs.

1. The cost sharing formula hammered out by the states, Amtrak, the FRA, and accepted by the STB which took effect for the state supported corridors in FY2014. The formula could have shifted shared costs around or eliminated some book keeping costs. In the August 2014 MPR, the total allocated national system cost compared to Aug 2013 has declined by $149.2 million excl OPEB, etc and $316.4 million for the fully allocated cost.

2. Restructuring into 3 separate business lines for NEC, state supported corridors, and LD trains. The overhead costs for the managers and staff that moved to the NEC and state corridor divisions were likely partially allocated to the LD trains and now no longer are. Meanwhile the LD division has to fully support the overhead of its managers, but the LD trains could have come out ahead with lower overhead.

3. There were a number of layoffs and personnel reductions in management and maintenance several years ago specifically to reduce overhead. Due to separation and termination costs, the savings from the staff cuts may not have shown up until the FY14 budget.

4. Early Buy-outs of equipment leases. From FY10 to FY13, Amtrak exercised early buy-out options on leases for the 50 Viewliners, close to a couple of hundred Superliners, and a large chunk of the P-42 fleet. Eliminating the debt load and lease payments on the equipment could have lowered the allocated costs, even though Congress was funding the debt service payments in its annual appropriations. If Amtrak can manage to purchase the 130 Viewliner IIs out of annual capital budget without a loan, that should keep the equipment costs down for the single level LD fleet.

There has been a rather large reduction in the total allocated cost for the AutoTrain. Curious how that was justified, but the AT costs also jumped up a few years ago, so it may have been a matter of changing and correcting cost allocations for the AT.

Regardless of how and why, in the latest reports, the total allocated costs for the LD trains has shrunk while revenue has held up despite the service disruptions and delays. That helps the survival prospects for the LD trains.

I believe the Auto Train cost spike was down to the station overhauls (which may have been spread over 2-3 years in a somewhat simplified manner...that is pretty standard in accounting and is also the genesis of the infamous $25,000 toilets and whatnot: Someone simply divides the cost by parts or periods of time rather than looking at how/when the cash is spent

Edit: Speaking more generally, and repeating stuff that has already been said, Amtrak needs to focus on getting more equipment into service on a number of routes. It's pretty clear that to at least some extent, LD numbers aren't falling because there is a mass of pent-up and unmet demand, particularly at not-top-bucket pricing points.

Anderson: This is why I hate accrual accounting. You can do it well, but you have to *want* to do it well. In accrual accounting, the Auto Train station expenses should have been spread out for accounting purposes over the entire lifetime of the improvements (upwards of 50 years probably). But nobody ever bothers to account for it that carefully, of course.

In the alternative, cash accounting, they should of course have been listed as capital improvements and be completely outside the operating budget. Surely that's what Amtrak usually does?...

I will certainly agree with you that Amtrak needs to get more rolling stock into service on a number of routes. This is certain to have direct benefits to the bottom line.

Afigg: thanks for the list of possibilities.

I think assigning the managers to specific business lines could account for the transfer of overhead from the LD trains to the NEC between 2012-2013.

The massive drop in all overhead from 2013-2014 has to be real, though, whether it's delayed-response to the layoffs, delayed-response to the e-ticketing conversion, or elimination of interest. (Interest is supposed to be broken out separately, so I think it isn't that.)

But even with accrual accounting you have to go back a true it up. The next month you have to make adjusting entries to give a true number. I done this hundreds of times.

neroden said:

I've been poking around at the long-distance train financials again. Comparing 2012 "core contribution/loss" numbers to Boardman's presentation on a direct-costs basis, which it looks like is probably based on 2012 numbers, it looks as if the following amounts of overhead were allocated to each train:

Silver Star $38 million (44.1 - 6)

Cardinal $11 million (16.9 - 6)

...

I can't make much sense of this. Shared stations are part of this, but they certainly don't account for the entire pattern. ...

Anyway, if I back out this 2012 level of allocated overhead from the August 2014 results, to try to derive direct-costs losses for YTD 2014, I find:

Silver Star $4 million surplus

Cardinal ($2 million loss)

...

Now, these numbers still look close to the *right order* to me, just as the numbers in Boardman's presentation did. They seem to roughly correspond to online city size, ridership, train length, load factors, cost of operations, etc. in a way that other commonly referenced numbers do not.

...

Make no mistake: there's a lot of overhead to cover. $416+ million. Some unknown amount of it is arguably "avoidable" or belongs in direct costs -- it's not clear how "shared stations" are treated, for example. Most of the overhead is not like that, though; it's unavoidable overhead. (There's also very likely to be overhead buried even in "direct costs", unfortunately.)

The entire premise of my calculation here, of course, is that allocated costs haven't changed, and that any drop in costs is a real drop in direct costs. I have no way of telling whether this is true, however, and I actually suspect that it isn't.

...

Click to expand...

It's all beyond my poor power to calculate!

But let me direct your attention to the PRIIA PIP report on the Cardinal.

Because of the discussion of taking the train daily, it has some info

that other PIP reports do not.

http://www.amtrak.com/ccurl/536/878/PRIIA-210-Cardinal-PIP.pdf

See page 24, under Appendix C: Financials and Expense Description

For FY 2010 base, it gives Stations a cost of $1.8 million total for the

current 3/7 Cardinal, and for a daily Cardinal, an estimated $2.0 million

(but shouldn't it add up to $2.2 million? Oh Lordy).

These from Total Direct and Shared Costs of $26.8 million as of FY 2010,

and an estimated $37.9 million expenses for the daily Cardinal (yielding

a Core Contribution/Loss of ($21.6 million).

I liked to think that sharing more station costs could help a lot. Charlottesville

used to have Station Costs/7 Crescents + 3 Cardinals = 10, then it became

Station Costs/7 Crescents + 3 Cardinals + 7 Lynchburgers = 17, with the

Lynchburger thus taking on a little cost from the LD trains. But very little

savings, after all, is how it looks from the PIP figures, if the entire route's

Station Costs, from Penn to Union Station, aren't even $2 million.

So I don't think Station Costs are enuff of a deal to tilt the figures on

the Operating Profit & Loss statements.

Seems like the apparent improvements in operations when LD routes

are overlaid by corridor trains come from the added revenue due to

ridership induced by the convenience of added frequencies, rather

than any cost sharing.

The station situation is complicated. Assuming a fair allocation, WAS-NYP and CHI will allocate a minimal fraction of their costs to the Cardinal to begin with (PHL-NYP have somewhere close to 50 Amtrak trains per day, to say nothing of commuter trains; WAS gets pretty close as well...and CHI has a similar situation with Amtrak and Metra). So the Cardinal isn't going to have much of an impact here in any event...

Transitioning from 3x weekly to daily, the Indiana stations would mostly not need to add any operations (the Hoosier State already rounding service up to daily). CIN would need to add one day of operations (the station is open six nights per week) while the WV stations would need to add four days (since the Cardinals are on the same day each way there). However, only three of the stations in WV (Huntington, Charleston, and Prince) and six of the stations between WAS and CHI (Charlottesville, Cincinnati, and Indianapolis are the others) are staffed. So you'd be adding something like 13 shifts to station staffing (4 each to the WV stations, one to Cincinnati, and none to Charlottesville or Indianapolis...and probably near zero added other expenses beyond that).

Finally, on the odd addition: I assume that's a bunch of rounding adding up to look odd.

MikefromCrete said:

If Amtrak's accounting is like the municipal and school district budgets I used to wrestle over as a reporter, you'll never figure it out since the budgets are designed to be overly complicated. But, I'm terrible at math.

Click to expand...

They're designed to let the top administrators load the school budget down with debt but make it look like they saved money so they get a bonus and nobody questions bizarre ballooning admin staff, pointless admin projects, and admin pay and benefits.

If anyone complains about spending, cut the busing budget or blame teachers and destroy any income for brand new teachers while spinning rhetoric about the "tenured dead wood". The dead wood in admin building (suspiciously larger than the high school) is not dead wood but genius results-driven visionary professionals that somehow are vitally important in some ineffable way to the individual operation of each individual school.

/not bitter or anything

IMO, LD trains are part of an eco-system (which they support) that includes LD coach bus service. LD trains would have better metrics with better frequencies, enough equipment for demand, and with more frequencies for the highest demand city-pairs (for many trains this is on end points). Likely this is why the Eastern routes generally outperform the Western routes.

When the Amtrak trains get full and the prices high, the bottom feeder multi-national, non-unionized, pack-'em-in non-stop coach services are there to fill the gap. Cheap-o buses have depressed demand on Midwest corridor services in the last few years. Amtrak has hit a capacity wall. What really hurts is that if some portions of an LD trip are selling out, but not the whole trip, then they are turning away paying customers AND running around for hundreds of miles with empty seats. Lose-lose.