July Performance Reports

[henryj]

Anderson, Afigg and henryj,

Thanks for the very detailed information. I'm amazed at how much time and trouble you fellows spend at your "hobby" lol!

From my vantage point, long distance train riders are non-flyers (like me), disabled or obese who may have trouble traveling on airplanes (my rather large brother-in-law had to buy two seats on the plane on a recent trip to Tampa), people from "railroad families" who have a love affair with anything train, and people who are just curious as to train travel.

Recently the high cost of gasoline and the TSA hassles at the airports have added ridership to Amtrak. Business travelers are almost non-existant...unless they fall into one of the groups previously mentioned. I believe the railfans are an aging bunch...just from personal observation. There was a company from Michigan (I believe) who toured Pennsylvania back in the nineties. They set up open houses in hotels in Allentown, Reading, and Lancaster on successive nights trying to drum up business for their railroad tours. My wife and I attended the Reading meeting (actually held at the Holiday Inn in Morgantown, PA). I'd estimate they had about 120 people show up. My wife and I were probably the youngest one there. We were in our fifties lol!

Since we have a small percentage of the populace to draw from, it is important for each new rider to thouroughly enjoy his (her) Amtrak trip. Not only for repeat business but for the new rider to tell friends and family how great Amtrak is. That is why I am an advocate for cheerful, polite waitstaff. Clean restrooms. Clean stations and clean ROW.

Thanks again.

Thanks for the complement. It is just a hobby of course, but an interesting one. Thanks again.

 

[Train2104]

I'm honestly surprised by the Zephyr losing more than the Chief, both of which are worse than the SUnset.

 

[henryj]

I'm honestly surprised by the Zephyr losing more than the Chief, both of which are worse than the SUnset.

The SWC is the one that surprises me. But it's all tied to revenue, maintenance costs allocations and labor. Maintenance is $1 a car mile and labor is by the hour. The SWC just doesn't bring in enough revenue in spite of the high costs of rooms. The CZ and the EB both bring in around 20 mil in sleeper revenue, the SWC only 15.6 mil. The EB brings in coach revenue of almost 32 mil, the CZ 20 and the SWC 21. Thus the EB only loses 1.5 mil in spite of it's on time problems, whereas the CZ and the SWC lose 23 mil and 16 mil respectively. The Sunset only runs three times a week as does the Cardinal. The answer to all this seems to be more capacity, particularly in the peak months to increase revenue and perhaps fewer on board personnel to bring these trains up to the break even point on avoidable or operating costs. Overall LD trains still recover 85% or more of their operating costs which isn't that bad. Amtrak overhead costs I just don't mess with as Amtrak has a certain level of overhead and they have to put it somewhere. It's all allocated anyway and is basically meaningless unless you are going to just whole sale discontinue trains. Otherwise it's not going to change.

 

[afigg]

A few other thoughts of note, this time regarding the Silvers:

(1) FEC is going to collapse the ORL-MIA market when that comes on line, and collapse it /hard/. The Meteor might retain some residual business, but the FEC is going to be two hours faster than the Meteor NB and not subject to through delays SB. If the FEC goes through to Tampa, I wouldn't be surprised to see the Star drop 100,000 passengers/year.

(2) On the other hand, assuming that Amtrak can get a section of one (or both) Silvers onto the FEC, that should make up at least some of the difference. This combination leads to...

(3) It's probably going to make sense for Amtrak to revive the old East Coast/West Coast split in the Silvers (which was the case until the early 90s) and turn one section of each at Tampa and the other section at Miami. This probably won't happen, but a combination of #1 and #2 are going to gut business ORL/TPA-MIA.

Hard to say what the full effect of the AAF Miami to Orlando service is going to be on the Silvers. AAF will go to Orlando Airport while Amtrak will be at the Orlando SunRail station. Different stations may not entirely overlap on the local markets. If it succeeds, AAF will carry a lot more passengers between Miami and Orlando than the Silvers ever could. While AAF may erode Amtrak's Orlando to Miami ridership, more people taking AAF in Florida could result in greater awareness of the Amtrak Silvers and get more Florida residents taking the Silvers longer distances to stops north of FL.
If AAF extends its new 125 mph corridor to Tampa and extends service to Jacksonville, yea, that is going to take much of the Silver intrastate business on the current routes. Connections between the CSX tracks and the new FEC tracks could provide new faster alternate routes for the Silvers - if the FEC/AAF is at all interested. There are a lot of unknowns on the FEC/AAF plans beyond running the downtown Miami to Orlando Airport corridor service. First, we will see if AAF can build and succeed with the Miami-Orlando service.

 

[haolerider]

I'm honestly surprised by the Zephyr losing more than the Chief, both of which are worse than the SUnset.

The SWC is the one that surprises me. But it's all tied to revenue, maintenance costs allocations and labor. Maintenance is $1 a car mile and labor is by the hour. The SWC just doesn't bring in enough revenue in spite of the high costs of rooms. The CZ and the EB both bring in around 20 mil in sleeper revenue, the SWC only 15.6 mil. The EB brings in coach revenue of almost 32 mil, the CZ 20 and the SWC 21. Thus the EB only loses 1.5 mil in spite of it's on time problems, whereas the CZ and the SWC lose 23 mil and 16 mil respectively. The Sunset only runs three times a week as does the Cardinal. The answer to all this seems to be more capacity, particularly in the peak months to increase revenue and perhaps fewer on board personnel to bring these trains up to the break even point on avoidable or operating costs. Overall LD trains still recover 85% or more of their operating costs which isn't that bad. Amtrak overhead costs I just don't mess with as Amtrak has a certain level of overhead and they have to put it somewhere. It's all allocated anyway and is basically meaningless unless you are going to just whole sale discontinue trains. Otherwise it's not going to change.

If you increase capacity, how can you have fewer on-board staff? Doesn't make logical sense!

 

[Nathanael]

So, keeping it simple by ignoring the effects of a big price increase on ridership, Amtrak would to more than double the prices on the CZ to get the train to breaking even - for the fully allocated costs which is different from the direct costs.

Unfortunately Amtrak has only published direct costs *very* occasionally; they are not published regularly.
There was a bar graph of "revenue minus avoidable costs" in a presentation by Boardman sometime in the last year or two, as part of the political campaign to convince Congress that cancelling trains is stupid and will not lead to profit.

At that time, if I remember correctly, the Auto Train and Palmetto had positive numbers, the Silver Star, Silver Meteor, Lake Shore Limited had very small negative numbers (close to break-even); the Capitol Limited, CONO, and Crescent had somewhat larger negative numbers; and the other trains had much larger negative numbers. Conclusion: The two-night western trains appear to perform more poorly on an avoidable costs basis. We have no idea what Amtrak counted as avoidable costs for that report, however. Personally, I suspect (from some other stuff I've read) that any station which is only served by one train is counted as avoidable cost, while a station which is served by multiple trains is an allocated cost. This would account for that result entirely.

Most of the cost of train service is in shared costs; it's just a fact of railroading. This means there are major benefits to scale at all levels, from longer trains, to more trains per day on the same route, to routes sharing track.

 

[Nathanael]

I pretty much get the same results, maybe in a little different order, but most of the trains are so close to each other it makes little difference. I get the same big losers and the same winners. Except the LSL. I have no idea why they get such good results on that one and I get such bad results. There must me something we just don't know about the train's costs. In the LSL's PRII report they state it gets around a 75% cost recovery, that means is still loses a ton of money. I get a 65% cost recovery. I think they just made a mistake on the graph.

I suspect Amtrak is basically correct about the LSL. The LSL is unusual among long-distance trains due to (1) the split at Albany into New York and Boston sections, and (2) the fact that most of its length is covered by other trains. The only "LSL-only" stations are Bryan, OH; Erie, PA; and Pittsfield, MA -- every other station is already being served by another train, often several. It's also a *long* train, distributing the OBS, crew, and fuel costs over more passengers. (And the route is flat so it should get good fuel efficiency.)

You'll note that the Silver Service trains which have good numbers have similar characteristics.

I suspect that even with "direct" or "avoidable" costs there's still an issue of potentially-shared costs. The Sunset Limited has to carry the costs of idle OBS workers waiting for the next train. The LSL never has to carry these costs, due to the much more intensive availability of OBS work out of New York and Chicago. There's probably lots of other little things like that.

---

I suspect the California Zephyr numbers are as bad as they arebecause of the scenic "Moffat Tunnel" section, frankly: it's sloooow going over the mountains. That means the train is often half-empty from Glenwood Springs to Reno (even while being full on the east and west ends) and Amtrak has commented on this. Running over the mountains is fuel-intensive, too; it's the most expensive part to operate but raises the least revenue. I don't know if there's any way to separate out the numbers for different parts of the California Zephyr, but I suspect the Denver-Chicago portion, which makes much better time and often needs an extra coach, is doing as well as the eastern trains. I read somewhere that over half the train seats turn over in Denver.

 

[afigg]

Unfortunately Amtrak has only published direct costs *very* occasionally; they are not published regularly.

There was a bar graph of "revenue minus avoidable costs" in a presentation by Boardman sometime in the last year or two, as part of the political campaign to convince Congress that cancelling trains is stupid and will not lead to profit.

Anderson posted a link to the Boardman presentation earlier in this thread: http://www.amtrak.com/ccurl/401/881/Amtrak-CEO-Boardman-House-T&I-testimony-Mar-05-2013.pdf.

Several of the FY2010 PRIIA PIP reports for the LD trains broke out the expenses. Fuel cost is a factor for the trains, but crew and maintenance costs are the biggest cost components. For the CZ, in FY09 or FY10, total direct and shared costs were $101.1 million with the fuel cost component at $12 million. T&E was $17.4 million, OBS was $13.5 million, Maintenance of Equipment was $16.5 million (maintenance could cover a wide range of equipment cost in this context). What drives up the costs of the western LD trains are the 2 day trips with higher crew and equipment costs but less opportunity to turn seats and sleeper rooms over, not so much the fuel cost of running over the mountains.

 

[henryj]

I pretty much get the same results, maybe in a little different order, but most of the trains are so close to each other it makes little difference. I get the same big losers and the same winners. Except the LSL. I have no idea why they get such good results on that one and I get such bad results. There must me something we just don't know about the train's costs. In the LSL's PRII report they state it gets around a 75% cost recovery, that means is still loses a ton of money. I get a 65% cost recovery. I think they just made a mistake on the graph.

I suspect Amtrak is basically correct about the LSL. The LSL is unusual among long-distance trains due to (1) the split at Albany into New York and Boston sections, and (2) the fact that most of its length is covered by other trains. The only "LSL-only" stations are Bryan, OH; Erie, PA; and Pittsfield, MA -- every other station is already being served by another train, often several. It's also a *long* train, distributing the OBS, crew, and fuel costs over more passengers. (And the route is flat so it should get good fuel efficiency.)

As much as we might want it to be so, the LSL does not make any money and here is why. Revenue for the Meteor, sleeper 9.6 mil, coach 23.5 mil total 33.1 mil and total overall 35.4. The Star is somewhat less. The LSL by contrast has sleeper revenue of 7.7 mil and coach revenue of 16.1 and total of 26.7 and total overall of 28.5. Ridership on the Meteor is actually less. Operating costs per my calculations are 36.2 for the Meteor and 38 mil for the LSL. Amtrak can finagle the costs anyway they want, but the revenues for the LSL are way below the Meteor and the Star and about on par with the Crescent.

 

[AlanB]

Operating costs per my calculations are 36.2 for the Meteor and 38 mil for the LSL.

And right here we see one of the problems with your calculations, Henry. Sorry! But there is no way that can be true.

Just based upon crew costs alone, the Meteor's costs have to be higher. On the Meteor, assuming it runs on time, you are paying your OBS crew for 28 hours. That's 3 sleeping car attendants, 2 LSA's, and usually 3 people in the dining car, plus 2 coach attendants for a total of 10 people. Let's assume an average of $20 per hour, that's probably low, but still for our purposes it works. Therefore, for 1 trip just the OBS costs are $5,600.

On the LSL, we have 3 OBS on duty for 22 hours; = $1,320 on the Boston section. For the NY section we have 7 employees on duty for 18 hours; = $2,250. Grand total for the LSL is $3,840 in OBS salaries. And the operating crew salaries would be equally lower.

So for the 2 trains per day, per year, total OBS for the Meteor are $2.044 Million. Total OBS for the LSL is $1.401 Million. That's $640,000 more in OBS costs for the Meteor. Throw in at least $50K more for operating crew costs on the Meteor, more fuel burned due to longer operations, more meals served, and the operating costs have to be higher for the Meteor than for the LSL.

 

[Anderson]

Well, and there's also the four-day cycle on the Silvers instead of three days on the LSL. The LSL does a same-day turn in CHI; the Silvers don't do one at either end.

 

[henryj]

Operating costs per my calculations are 36.2 for the Meteor and 38 mil for the LSL.

And right here we see one of the problems with your calculations, Henry. Sorry! But there is no way that can be true.

Just based upon crew costs alone, the Meteor's costs have to be higher. On the Meteor, assuming it runs on time, you are paying your OBS crew for 28 hours. That's 3 sleeping car attendants, 2 LSA's, and usually 3 people in the dining car, plus 2 coach attendants for a total of 10 people. Let's assume an average of $20 per hour, that's probably low, but still for our purposes it works. Therefore, for 1 trip just the OBS costs are $5,600.

On the LSL, we have 3 OBS on duty for 22 hours; = $1,320 on the Boston section. For the NY section we have 7 employees on duty for 18 hours; = $2,250. Grand total for the LSL is $3,840 in OBS salaries. And the operating crew salaries would be equally lower.

So for the 2 trains per day, per year, total OBS for the Meteor are $2.044 Million. Total OBS for the LSL is $1.401 Million. That's $640,000 more in OBS costs for the Meteor. Throw in at least $50K more for operating crew costs on the Meteor, more fuel burned due to longer operations, more meals served, and the operating costs have to be higher for the Meteor than for the LSL.

Cool Alan. I am open to change. I put in 10 OBS for the Meteor. I used $50 an hour because you have to double to include benefits. I used 28 hours for it also. On the LSL I used 23 hours for the Boston section and 19 for the NY section. I had used OBS crews of 3 and 8, but I can reduce the NY section to 7. So now we have OBS for the Meteor of 8.4 mil and the LSL of 6.1 mil. For T&E I always use 4 man crews and I used a higher rate for the LSL, $115/hr for the LSL vs $95/hr for the Meteor. The Meteor is still less because the LSL has the Boston section to contend with as a separate train and the rate diff. Meteor 6.4 mil and LSL 7. Maintenance costs at 11 cars per train(incl engines) for the Meteor are 9.3 mil and for the LSL 15 cars, 10 to NY and 5 to Boston, total 9.2 mil. Rent is less for the LSL as is fuel. Total Op costs for the Meteor now are 38.7 and for the LSL 37.4. Thanks, I am glad to get some accurate information finally. Still doesn't explain the LSL low revenue production. Other parameters I use are $3.50/gal for fuel and $5 a train mile for track rent. All the other costs are really insignificant compared to these. I used station, diner and switching costs from a published Amtrak report. The Meteor now shows an operating loss of 3.3 mil and the LSL 8.9 mil. Hope this helps.

 

[Anderson]

Henry,

You should probably add in hotel costs for the Meteor/Star crews, since the OBS are going to overnight at the "back end" of the trip. IIRC this isn't a cost for the LSL (or Cap) because of the same-day turns in CHI. I'm also not sure if there's held-away time involved there.

 

[henryj]

Henry,

You should probably add in hotel costs for the Meteor/Star crews, since the OBS are going to overnight at the "back end" of the trip. IIRC this isn't a cost for the LSL (or Cap) because of the same-day turns in CHI. I'm also not sure if there's held-away time involved there.

It's all in the hourly rate. Why overly complicate it as it just an estimate anyway and it probably wouldn't make much diff. When you are talking 10's of millions, this little stuff really doesn't matter much. We are at 76% recovery for the LSL which is around what the had in it's PRII and 91% for the Meteor. I do have a question for Alan........why do these trains need two LSA's?

 

[AlanB]

I do have a question for Alan........why do these trains need two LSA's?

One for the dining car and one for the cafe car.

You must be an LSA to handle cash and inventory for Amtrak.

 

[Eric S]

Henry,

You should probably add in hotel costs for the Meteor/Star crews, since the OBS are going to overnight at the "back end" of the trip. IIRC this isn't a cost for the LSL (or Cap) because of the same-day turns in CHI. I'm also not sure if there's held-away time involved there.

When an OBS crew does a same-day turn (like the LSL in Chicago), do they lay over in a hotel (at a day rate)?

 

[AlanB]

I put in 10 OBS for the Meteor. I used $50 an hour because you have to double to include benefits. I used 28 hours for it also. On the LSL I used 23 hours for the Boston section and 19 for the NY section. I had used OBS crews of 3 and 8, but I can reduce the NY section to 7. So now we have OBS for the Meteor of 8.4 mil and the LSL of 6.1 mil.

So let's see, for the Meteor you use the actual running times to calculate your numbers, but for the LSL you use an extra hour. Is this so you can keep the expenses higher than they need to be?

For T&E I always use 4 man crews and I used a higher rate for the LSL, $115/hr for the LSL vs $95/hr for the Meteor.

AFAIK, the LSL uses largely 3 man operating crews. The Meteor does so too for part of the run, but I do believe that there are a few stretches where it's 4 man operating crews.

The Meteor is still less because the LSL has the Boston section to contend with as a separate train and the rate diff.

Huh? :unsure:

Maintenance costs at 11 cars per train(incl engines) for the Meteor are 9.3 mil and for the LSL 15 cars, 10 to NY and 5 to Boston, total 9.2 mil.

Again, padding the numbers in favor of the Meteor for some reason. The Meteor is 12 cars, if you include engines.

 

[AlanB]

Henry,

You should probably add in hotel costs for the Meteor/Star crews, since the OBS are going to overnight at the "back end" of the trip. IIRC this isn't a cost for the LSL (or Cap) because of the same-day turns in CHI. I'm also not sure if there's held-away time involved there.

When an OBS crew does a same-day turn (like the LSL in Chicago), do they lay over in a hotel (at a day rate)?

Yes.

 

[billthebarn]

Alan,

Do the OBS get paid for every hour they are on the train? In your example, I believe you used $20 per hour X 22 hours. Isn't sleep time deducted?

Thanks

 

[AlanB]

Alan,

Do the OBS get paid for every hour they are on the train? In your example, I believe you used $20 per hour X 22 hours. Isn't sleep time deducted?

Thanks

Bill,

I think that they're paid bumper to bumper post; but I'm not 100% positive.

Regardless, the results wouldn't change if that's not true, as 4 hours would be deducted from both crews if they aren't getting paid while sleeping.

 

[Paulus]

Alan,

Do the OBS get paid for every hour they are on the train? In your example, I believe you used $20 per hour X 22 hours. Isn't sleep time deducted?

Thanks

They get four hours of unpaid sleep time per night. I'm not sure what, if any, time is paid before departure or after terminus however. I do know that they have some duties and it would be odd not to be paid for them.

 

[Anderson]

Henry,

You should probably add in hotel costs for the Meteor/Star crews, since the OBS are going to overnight at the "back end" of the trip. IIRC this isn't a cost for the LSL (or Cap) because of the same-day turns in CHI. I'm also not sure if there's held-away time involved there.

It's all in the hourly rate. Why overly complicate it as it just an estimate anyway and it probably wouldn't make much diff. When you are talking 10's of millions, this little stuff really doesn't matter much. We are at 76% recovery for the LSL which is around what the had in it's PRII and 91% for the Meteor. I do have a question for Alan........why do these trains need two LSA's?

That's fair. If HAT+hotel+food came to $150/person, that would be another $1500/trip.

On the LSL: Part of the lower revenue production there (vis-a-vis the Meteor) is in the schedule, I suspect. Basically, the WB train has nice times and is a decent choice for getting to CHI. The EB train, on the other hand, basically burns a whole day getting to NYP. At least if I had to guess (and Amtrak broadly concurs), this generates empty space EB. The Meteor is more symmetrical in this regard. In this vein, the Meteor has (at least for the last ten months) 35,166 sleeper pax vs. 29,007 on the LSL. That comes to about a $1.8m difference in sleeper revenue.

Edit: One other point: The Meteor is likely helped by the fact that trips from the NEC into VA get billed at NEC rates. As such, a NYP-RVR ticket frequently costs as much as or more than a RVR-MIA ticket would.

 

[Devil's Advocate]

I find these report reviews fascinating but the disputes over accounting theory get old after a while.

Does anyone know a trained analyst who could help dig up some of the missing data and flesh out some of the assumptions?

 

[Anderson]

I find these report reviews fascinating but the disputes over accounting theory get old after a while.

Does anyone know a trained analyst who could help dig up some of the missing data and flesh out some of the assumptions?

My understanding is that the problem is finding the data. What would probably be most useful would be a detailed case study on 2-4 routes. Mind you, the theory fights would still rage on (and they wouldn't be invalid), but you could group costs and basically do a "choose your own accounting" game and see how things stack up under several assumptions.

 

[haolerider]

I find these report reviews fascinating but the disputes over accounting theory get old after a while.

Does anyone know a trained analyst who could help dig up some of the missing data and flesh out some of the assumptions?

My understanding is that the problem is finding the data. What would probably be most useful would be a detailed case study on 2-4 routes. Mind you, the theory fights would still rage on (and they wouldn't be invalid), but you could group costs and basically do a "choose your own accounting" game and see how things stack up under several assumptions.

All the assumptions in the world make little or no difference, unless you can get into Amtrak's accounting system to determine how costs are allocated. There is (and almost always has been) controversy within various Amtrak departments as to cost allocation. You need to have a complete understanding of staffing coast, benefit costs, pensions costs, management expenses, etc. to properly examine the revenue vs. expenses of each train....and they vary tremendously. I understand the desire to try to determine what can be done to improve the overall profit/loss situation within Amtrak, but continuing to "estimate" costs and allocations is just like trying to swim upstream......you might make a bit of headway, but eventually you are probably going to find yourself exactly where you started!