For the 12 months ending Sept 2007, Amtrak had a load factor slightly less than 50%. So, when critics see this, they start making cracks about "half empty trains". On the whole, they would be correct! But, as we know, its not quite that simple. A train operating from A to B to C to D might have a lot of passengers from A to B and C to D, but not so many from B to C.
Still, it's a problem - or maybe an opportunity. This is "low hanging fruit". How do you fix it?
1. Adjust fares to fill seats in the lightly used segments? Can Amtrak do this more than they already do it?
2. Drop and add equipment? Costs vs. benefits? What are the barriers? For example, Amtrak used to add and drop 2 coaches in Atlanta on the Crescent. NS used to handle the switching and storage. Amtrak doesn't do that any more. That causes them to use 4 more coaches.
3. Add frequency to the heavy portion of the route? I notice that those IL trains with doubled frequency did not double ridership. I am surprised by this. I would have thought they should have more than doubled. Are the travel plans of medium haul leisure travellers that flexible? Aren't these trains attacting any business travellers? Why or why not?
-Don (Random stuff, mostly about trains - what else? http://blerfblog.blogspot.com/)
Actually, if their load factor is about 50 percent, Amtrak is doing a strong business.
Airlines, I believe, run a higher load factor, but they play all kinds of games of advance purchase fares and witholding comfortable or safer (i.e. exit row) seats from the cheap-fare customers and overbooking and the whole gamut. If Amtrak is on average at 50 percent load factor, they are probably turning away business right and left because they are full at the popular travel times.
The point I want to make about load factor is that while people do auto vs common-carrier mode comparisons and allow that a car is carrying usually only one person (the evil sole-occupant automobile trip), a common-carrier conveyance can't run full all of the time either on account of variable demand. If I highway is running below capacity in an off peak time, there is empty concrete -- if a train is below capacity in off-peak, there are empty seats. In terms of energy-efficiency comparisons, where people compare a full gallery-car commuter train against a sole-occupant auto and proclaim the energy efficiency of trains is missing a point.
I guess the only other thing about load factor is that planes, especially, have to play games to fill up the seats and meet peak demand, but trains famously can add more cars. Amtrak doesn't seem to vary consists with demand very much.
If GM "killed the electric car", what am I doing standing next to an EV-1, a half a block from the WSOR tracks?
Paul Milenkovic wrote: Actually, if their load factor is about 50 percent, Amtrak is doing a strong business.Airlines, I believe, run a higher load factor, but they play all kinds of games of advance purchase fares and witholding comfortable or safer (i.e. exit row) seats from the cheap-fare customers and overbooking and the whole gamut. If Amtrak is on average at 50 percent load factor, they are probably turning away business right and left because they are full at the popular travel times.The point I want to make about load factor is that while people do auto vs common-carrier mode comparisons and allow that a car is carrying usually only one person (the evil sole-occupant automobile trip), a common-carrier conveyance can't run full all of the time either on account of variable demand. If I highway is running below capacity in an off peak time, there is empty concrete -- if a train is below capacity in off-peak, there are empty seats. In terms of energy-efficiency comparisons, where people compare a full gallery-car commuter train against a sole-occupant auto and proclaim the energy efficiency of trains is missing a point.I guess the only other thing about load factor is that planes, especially, have to play games to fill up the seats and meet peak demand, but trains famously can add more cars. Amtrak doesn't seem to vary consists with demand very much.
I can vouch for the load factor point.
I once flew from Green Bay to Milwaukee on one leg of a trip home; 8 people aboard a 150 seat MD80. The crew sat and relaxed; I spoke them and they said this was their daily milkrun. When this happens on an Amtrak segment we can all but be sure of the comment that will be made.
What this goes to show me is that you can cook the books or spin an impression just about any way you want in order to make your point. When I drive 300 miles alone in my car it could be said that I am operating at somewhere between 20-25% efficiency. Some might even say it is worse than that because you could possibly stick two small people in my cargo area. This is a stupid argument, but I could also say that this is to me, nearly as stupid as the arguments about money losing Amtrak. I really don't see the point unless this public outlay will somehow delay our first manned trip to Mars, or, slow down the building of a fence across the entire border with Mexico, or compete with a brand new testing program in schools so that no school district gets left behind.
I've been following the activities of a group that is seeking to reintroduce Amtrak service in Kansas during daylight hours. This group has put together a preliminary study with cost estimates that indicate that two trains could be put on the tracks each direction for about $7 million dollars (and these appear to be credible figures). The cost to do this could also be largely funded with the money that may come from the Amtrak bill that NARP reported as passing in Congress last month (with federal funding that would pay 80%). That would be approximately $1.4 million dollars that the state would have to come up with, and this is what I paid for my house. If Kansas can't come up with that amount of money to add a nice segment to the national rail system then that State and our country ought to throw in the towel right now. Let's just keep dreaming about a manned mission to Mars, or an improvement package for passenger rail that may occur 43 years from now (I should be dead by then), and my grown kids will be grandparents and will probably prefer to ride around in little saucers between space needles).
Just to add a data point to the discussion, the airlines as a whole are running above 80% load factor this year. 80% is the generally accepted benchmark for "full" for the airlines. Airtran actually posts their break-even load factor on their web site - it's in the mid 70s.
I know Amtrak can't afford much extra equipment for seasonality, but I wonder if they aren't doing their train sizing for maintenance convenience rather than service demand. There are cost trade-offs, for sure, but....
Airliners are the ultimate "fixed consist". I would think that the variable nature of a train consist should yield higher load factors - at least in theory.
Would it be helpful for Amtrak to play more games with the fares in order to get seats filled? Don't they already do this to some extent?
I got into some heated debates about load factors on another board. One gentleman repeatedly argued that 50% load factors were sufficient justification for eliminating long distance trains.
Its easy to see how full a plane is by looking at the load factor, because nobody gets on or off en route. But since trains pick people up and drop people off along the way, it's not as clear.
Load factors only measure the average load on a given run. They don't tell you how full the train actually gets. For example, Train 14 never leaves Los Angeles full. If it did the passengers farther down the line would be out of luck. But it usually is completely full by the time it departs Sacramento. Once the train gets into Oregon, more people start getting off than are getting on, so that by the time it reaches Seattle it may only be a quarter full. So on average the train is only 50% full, but for part of its run it is at or near 100%, while at other points it is at or near 25%. (NB: I don't know the actual numbers, these are just examples for illustration purposes.)
Some people actually argue that the Starlight should therefore only operate between Sacramento and Portland because thats the only part of the run that operates near capacity. But then the train wouldn't have the passengers that boarded south of Sacramento, and so sales would be lighter.
So a train starts out partially filled, fills up, then gradually empties again. What is important to understand is that the principle is the same no matter how long the route. It works the same for a run of 200 miles as a run for 2,000 miles. I even observe the same thing on local city bus routes of 5-10 miles.
If we were to judge certain commuter railroads by average load factors over the whole week, the judgment would be that some lines would have to close while others would have to run unidirectionally at peak hours only. Judgments against subway and light rail lines would be even more draconian...e.g. the New York Subway's shuttle service to Rockaway Beach, where I was once one of less than a handful of passengers on a five-car train (and this shuttle operates 24/7, too).
Incidentally, the lowest load factor that I'm aware of for the Shinkansen is about 68 percent. Average load factors for the TGV are around 75 percent. Overall, Deutsche Bahn's "long-distance" passenger transportation load factors came in at 43.2 percent back in 2005 (this includes traditional InterCity trains, which are hauled by the Bombardier-built ALP-46-lookalike 101-class at 137 mph top speed, and the ICE-T tilt-train service, not just the ICE trains that travel the NBS corridors). Amtrak's on a par with their overseas counterparts, in terms of load factors at least.
I too have been puzzled about the fixed consist philosophy. Especially when you may be tight on equipment (think Superliners), it would seem to me the cost of switching in and out an additional coach or two would be far exceeded by the additional revenue produced. I understand that this could only be done at certain stops that already have longer dwell times.
In Amtrak's defense, I have heard this is exactly whet they are doing with the Empire Builder between Chicago and the Twin Cities.
In my earlier message I mentioned how it is possible to cook the books or spin the perception any way you want. Here's another one, that uses or taps into the situation that you describe:
Let's say (for the case of argument) an Amtrak train (any one of them) has a consist that will seat 200 people. Between any two points it may or may not approach full, but over the entire route, point-point they board 800 people. I could say that the train ran at 400%. Now, if there were at any one point a maximum load of 150 passengers and you reduced the capacity of equipment to just under that, then it could be said that the train is running over capacity (100% +). And this would be indisputable, but also a bogus evaluation of the actual efficacy of the service.
The next thing you could get into (to equate) would be the actual cost per passenger mile versus airplane, bus and car (and don't leave out bicycles and motor scooters as motor scooters also often run at only 50% capacity.) Well, my ludicrous examples sort of touch on where these arguments tend to go. I really grow weary of the some of the comparisons which always trend towards arguments to disqualify passenger rail service. This about sleepers and diners makes me weary and sick in the stomach. Come one and really, does a $79 airline ticket really cover all of the costs of a single flight, or is it more gimmic wherein airport, infrastructure, security, personnel and every other cost is all an completely covered by $79?, yeah, sure.
(edited for language, Ken)
That sounds like a great idea because two sets of coaches can serve the Minneapolis-Chicago cycle while more than two train sets (is it 6?) are required for the full Seattle-Chicago run. Does anyone know what cars and what types from what part of the consist they switch at Minneapolis?
Paul Milenkovic wrote: In Amtrak's defense, I have heard this is exactly whet they are doing with the Empire Builder between Chicago and the Twin Cities.That sounds like a great idea because two sets of coaches can serve the Minneapolis-Chicago cycle while more than two train sets (is it 6?) are required for the full Seattle-Chicago run. Does anyone know what cars and what types from what part of the consist they switch at Minneapolis?
There is a substantial historical database on this idea, the Milwaukee Road's Olympian Hiawatha, which was ultimately consolidated with the AM and PM Hiawathas. West of St. Paul, the trains split, and the OH continued on with six sets of equipment on line including a huge dome car and Creek Observation cars. A lot of expensive equipment, and yet a relatively low load factor.
The loss on the direct operating costs of the OH were low even though the railroad kicked and screamed about them. At $1.2 million in 1960, the RCAF in 2006 dollars would be something around $4 million in losses today with a load factor considerably lower than today's Empire Builder which loses about $9.5 million.
Which brings up an interesting conundrum about load factor and Operating Ratios (Direct Operating costs/Revenue).
The NEC has an overall load factor of 48%. OR of 35.8%
And, there's nothing special about the Northeast. The Hiawatha Service has a load factor 38.6%. OR of 40.3%
The average Short Distance train has a load factor of 40%. OR of 72.5%
The average Long Distance train has a load factor of 57%. OR of 134%
You will note that there appears to be an inverse correlation between load factor and profitability. That is something of an artifact as it describes a key difference between long haul and short haul trains. But, the moral of the story is that load factor is not the key determining factor to profitability. It certainly ought to be important, but I would be inclined to suspect that ability to handle peak load is more important, which leads to the lower operating ratios even as the load factor deteriorates -- as the higher revenues more than offset the costs.
My last three proposed trips on the Empire Builder were either rescheduled or not taken due to "sold out" status. With a lower load factor, the Empire Builder would undoubtedly generate more revenue and more overall riders, it just depends on the cost of offering the additional equipment, but it would generate more revenue and more riders and determing the beakeven point to that appears to be a key. That's where long distance trains run into leverage problems -- Amtrak doesn't add just an additional car to the Empire Builder, it adds six or more cars to the EB fleet.
And when each of those cars costs substantially more to maintain than a heavily-used 4,000 hp diesel locomotive , there is probably no hope.
I rode the Empire Builder from Milwaukee to Portland last December. The train had two engines, four coaches, two sleepers, a dinner, a lounge car, and a baggage car departing Milwaukee. Interestingly, it also had a private business car on the rear of the train. It was nearly full as far as Minneapolis, where a large number of the passengers left the train. The load was pretty light departing Minneapolis.
At Minneapolis the private car was pulled off the train by a private switcher; I think it is owned by a rail enthusiast group in Minneapolis that leases the car for special events. In any case, if I remember correctly, the person who appeared to be in charge of the private car told me that it cost several hundred dollars to pull the car off the Empire Builder and move it to its storage site.
At Spokane the train was split. The baggage car, dinner, one sleeper, two coaches, and of course one of the engines went on to Seattle, while the remainder of the equipment (four cars) went to Portland. There was a little switching at Spokane, but it was not much.
I am sure that there are more cars involved in the summer, but I suspect not too many more. The train does not run in sections from Minneapolis to the west coast.
Samantha wrote: The train does not run in sections from Minneapolis to the west coast.
The train does not run in sections from Minneapolis to the west coast.
Well, not "sections" in the understood sense, but the EB is somewhat unique. From Chicago, you are riding two sections consolidated into one train, but with separate train numbers depending on how you booked your ticket, Amtrak train # 7 (to Seattle) and Amtrak train # 27 (to Portland). Presumably, #27 is a derivative of 2nd Section, Train #7.There are generally three trains westbound of these combined sections and three eastbound at any point in time, and each will have an appropriate designation after the train number, usually the date of departure from Chicago or Seattle and Portland, and each of those does, in fact, operate in two sections Spokane/Seattle, Spokane/Portland, although they are designated by the separate train numbers rather than sections.
It may be possible for a total, at any given point in time, of possibly four westbound EBs and four eastbound EBs. This is off the top of my head as I don't have a timetable handy to check the timing, nor a particular interest in the train order designations for Amtrak, but that would be my guess, that you would have an Amtrak #7C20 (Train # 7, Chicago, departing December 20), 7C21, 7C22, a 27C20, 27C21, and 27C22 probably run together as 7C20, 7C21 and 7C22 and split at Spokane, plus eastbound an 8S20 Train # 8, Seattle, December 20), 8S21, 8S22 from Seattle, and a 28P20, 28P21, and a 28P22 from Portland or something like that, run as distinct numbers into Spokane and each two section set combined to Chicago. Total trainsets 6, possibly 8 on the rails at once, total train numbers, twelve, possibly 14 at once.
This is how it might have been done on one particular RR years and years ago, and whether this system is followed today, I have no idea. For all I know, they name the trains #7 Pluto, # 7 Mars, and #7 Raquel Welch to identify them and distinguish them while they are all in the system at the same time.
let us break down load factor into several blocks.
1. the europeans have a higher load factor beecause of multiple frequencies covering a route with much non-stop runs. i noted this especially in France.
2. Amtrak needs to get their labor contrac in order to not be required to have an assistant ccnductor for more than six (?) revenue cars in the consist.
3. Switching fees are very high when conducted by private RR's.
4. this being said there are locations where Amtrak switching could be productive.
A. Washiington DC. if the southern routes to Miami/Savannah, Atlanta, and Cincinnati/Chicago would add and subtract sleepers and coaches as the RF&P, SOU, and C&O did it would free up 1 for 1 car to fill out the consist and lack of load north of Washington. During holidays SOU would sometimes run extra sections of the Cresent between Atlanta and Washington and the ACL SAL would likewise their trains from Washington and Florida.
B. The station capacity in revenue cars is severely limited at many locations so a train longer than the station platform causes double stops unless major changes are made. My last trip to France had the main routes receiving platform extensions including all stations in Paris to accommodate two complete 11 car TGV's plus their four locos. Sometimes these trains would be split down line ie(Paris-Lyon) this has the added benefit of no more train starts but additional capacity on each train. The Freight roads would appreciate this.
Michael:
Do you have a source on the Hiawatha OR or DOC numbers? Our Wisconsin DOT guys have been complaining that Amtrak is remarkable opaque about the actual cost numbers on that train that Wisconsin contributes several million per year in operating subsidy -- WisDOT going before Amtrak is like you or me going to the car dealership where they write a whole bunch of numbers on what your monthly car payment on a piece of paper and whether you are getting a reasonable deal or not is anyone's guess.
I am not challenging or disputing any numbers, it is just that if we could see sources for numbers, that would help out a bunch of us in Wisconsin a lot. The states contributing subsidy money to Amtrak runs sure could use an Edmunds or a Blue Book to get independent numbers on what the subsidy contribution is covering.
Paul Milenkovic wrote: And, there's nothing special about the Northeast. The Hiawatha Service has a load factor 38.6%. OR of 40.3% Do you have a source on the Hiawatha OR or DOC numbers?
Do you have a source on the Hiawatha OR or DOC numbers?
Paul, the most recent monthly performance report for Amtrak contains this information. For the operating ratio by dividing direct operating expenses by revenue, and for the load factor by dividing the number of passenger miles by the number of seat miles.
http://www.amtrak.com/pdf/0709monthly.pdf
I agree on the calculation of the operating ratio, and I agree on how to calculate the average load factor for the system. But I don't see the direct seat and passenger miles for individual trains on the routes shown in the Route Performance Report. However, you can back into them, I think, but it does not give one the information necessary to calculate the load factors by segments.
The ability to look at segments, which I presume Amtrak management can do, is important to see which segments are generating the revenue. I ride the Eagle from Texas to Chicago and back three or four times a year. The passenger load that I have observed south of St. Louis in the coaches is considerably lighter than the load between St. Louis and Chicago both ways. The load in the sleeper, as expected, tends to be higher south of St. Louis.
Average load factors are potentially deceiving. They don't indicate how well a train is performing on a segment, e.g. Dallas to Cleburne, Austin to San Antonio, etc., or on what days they have a relative heavy or light load. Medians are usually more informative because they tend to dampen the variance around the mean by giving less emphasis to the highs and lows that can distort averages. I have been tracking the on-time performance for trains 21, 22, 1, 2, 821, and 822 at key Texas stations since August 1, 2007. In every instance, the median number of minutes the trains have been late is less than the average.
Does anyone know the number of saleable seats, on average, in a Superliner coach and sleeping car? Thanks.
I had in mind the days, usually during Thanksgiving and Christmas, when the Broadway Limited, as an example, ran in two sections. The second Number 29 usually ran about ten minutes behind Number 29. The Pullmans on the second section were not nearly as nice as those on the first section.
I grew up in Altoona, PA, which was a crew change point on for the Middle and Pittsburgh divisions. I would go down to the station around 10:00 p.m. to watch the flag trains stop in Altoon for a crew change. On several occasions, in the early 50s, I recall seeing the Broadway in two or more sections over the Christmas holidays. In 1957, the year that I graduate from high school, there were 38 passenger train movements a day through Altoona. Today there are two.
Samantha wrote: I agree on the calculation of the operating ratio, and I agree on how to calculate the average load factor for the system. But I don't see the direct seat and passenger miles for individual trains on the routes shown in the Route Performance Report.
I agree on the calculation of the operating ratio, and I agree on how to calculate the average load factor for the system. But I don't see the direct seat and passenger miles for individual trains on the routes shown in the Route Performance Report.
You're right. I thought it was there. Oops, not sure where I saw that. However, the numbers can be arrived at by dividing contribution per passenger by contribution per seat, which are contained in the report at page C-1.
Train Load Factor
NEC
Acela/Metroliner 59%Regional 44%Average 48%
Short Distance
Ethan Allen Express 39%Vermonter 38%Albany-Niagara Falls-Toronto 46%The Downeaster 36%New Haven - Springfield 43%Keystone Service 32%Empire Service 34%Chicago-St.Louis 40%Hiawathas 39%Wolverines 52%Illini 44%Illinois Zephyr 33% Heartland Flyer 39%Pacific Surfliner 35%Cascades 46%Capitols 26%San Joaquins 35%Adirondack 66%Blue Water 67%Washington-Newport News 47%Hoosier State 35%Kansas City-St.Louis 29%Pennsylvanian 63%Pere Marquette 63%Carolinian 64%Piedmont 40%Non NEC Special Trains 93%Average 40%
Long Distance
Silver Star 57%Cardinal 52%Silver Meteor 60%Empire Builder 61%Capitol Limited 60%California Zephyr 51%Southwest Chief 65%City of New Orleans 58%Texas Eagle 51%Sunset Limited 52%Coast Starlight 58%Lake Shore Limited 60%Palmetto 45%Crescent 47%Auto Train 60%Average 57%
"However, the numbers can be arrived at by dividing contribution per passenger by contribution per seat, which are contained in the report at page C-1."
That's what I did to test my thinking. Obviously, you calculated the load factor for each route shown. Did you copy the numbers from the report and paste them into an Excel spreadsheet, or did you calculate each factor separately?
I don't see enough information in the route performance report, or anywhere else, to calculate the load by segment. I see that the load factor for the Texas Eagle is 51 per cent. Based on my experience, the load factor north of St. Louis is higher, perhaps much higher, then the load factor south of St. Louis. Do you know whether Amtrak would provide that information if requested? I suppose one could exercise the Freedom of Information Act if Amtrak did not provide the numbers, although I am not sure if Amtrak falls under the FIA.
On another point, some readers claim that Amtrak or the airlines juggle their passenger load figures. It is possible but not probable. The financial statements of Amtrak and the airlines are audited by independent auditors, usually from one of the big four accounting firms, e.g. KMPG, PWC, etc. Although their focus is on the veracity of the financial statements, including the internal controls associated with the accounting systems that generate the data to produce them, they look at the operational statistics because of their impact on the financials. Amtrak and the airlines could fudge the figures, but if they were caught by the auditors, they would have a reportable item on their hands. And management, in this post Enron era, does not want to go there.
Samantha wrote: "However, the numbers can be arrived at by dividing contribution per passenger by contribution per seat, which are contained in the report at page C-1." That's what I did to test my thinking. Obviously, you calculated the load factor for each route shown. Did you copy the numbers from the report and paste them into an Excel spreadsheet, or did you calculate each factor separately?I don't see enough information in the route performance report, or anywhere else, to calculate the load by segment.
I don't see enough information in the route performance report, or anywhere else, to calculate the load by segment.
I used Excel; I hardly ever do anything without Excel, unless its on the back of an envelope. Over the years I have seen Amtrak station loading numbers in various state agency reports, consulting reports, and other publications. Since Amtrak isn't an investment vehicle, I don't pay much attention to it, but I assume from seeing those numbers here and there that they are generally available somewhere.
My secretary taught me "one of those lessons" last week regarding an STB question as I was fuming and fussing around the office about some document or another. After a little while, she pops in with it. Naturally, I asked "where did you get that?"
"I called the STB and they faxed it."
"Oh".
Perhaps a call to Amtrak might generate an answer or a direction to go.
Our community is FREE to join. To participate you must either login or register for an account.