short trains vs long trains - economics

[David Benton]

Just something i have been pondering awhile , and resurfaced reading an old new zealand rail magazine on the weekend .
The magazine described the then new opearting plan brought into action . ( around 2000 ) .
Basically they found that wagons on average were spending 35 % of their time sitting in marshalling yards , waiting for a long enough train to thier destination . there also was quite a cost involved in shunting the wagons , both finanically and in terms of worker safety .
Thier answer was to run trains to a fixed timetable , on fixed routes . such routes were not defined by the old system of each mainline or branchline been a route , but rather from traffic source to traffic destination , which may cover more than one traditional route . the trains were of fixed lenght , based on the 20 ft container teu , each train would provide say 100 teu capacity , and would run with that wether they were full or not . thus eliminating marshaling etc .
the plan was supposed to reduce marshalling costs , increase wagon and loco utilisation , and thereby increase efficency . wether it did or not is hard to say , as the railway magazine unfortunatley folded shortly after , and its not the kind of info you find in a newspaper .
so my question is , would this scenario or a version of it work in the USA . Would it be more economic to run shrter trains more frequently , say to 1/2 the time spent marshalling , or double wagon utilisation . i read of wagons spending days in a marshalling yard , waiting for a train to reach 100+ cars going there way . Are train crew costs the difference ( nz ones are single manned ) , or is the crew cost outweighed by the capital costs of all tose wagons sitting around ??? would extra business be generated by reducing overall trip time for a consignment ( i.e it doesnt spend days enroute sitting in marshalling yards) ????.

[2nd trick op]

So much ink has been spilled on this one that one hardly knows where to start, but let me throw out a few general observations and memories.

Very long freights, especially in flat country, were a trend that developed in the US around 1960. A railroad president named William Deramus was noted for pioneering the strategy on both the Kanasas City Southern and Chicago Great Western (which was down to operating one daily throught freight, plus a local and a passenger run on each of its major routes).

The local traffic was, by this time, declining rapidly as single-car pickup and delivery fell out of favor. But the lack of long sidings and the secondary effects (greater delays for broken drawbars and air lines, deferred maintenence et al) took such a toll that it was likely a big factor in the fall of the Rock Island.

And about the same time, Rio Grande made a point of its not subscribing to that theory; but Rio Grande operated in mountainous territory, originated a fair amount of traffic on line, and had to keep several partnering lines happy.

Until around 1985, this trend gathered speed on most roads; Mr. Weaver posted some observations on operating 2-mile-long freights on the old NYC Albany-Buffalo-Chicago line a year or two ago, so I hope he'll be heard from.

The crew-size and crew-district reforms of the 1980's also appeared for the time to have interrupted the trend toward longer freights, but i haven't seen any hard evidence of late.

You gentlemen with greater first-hand exposure; please lead on.

[David Benton]

Thank you 2nd trick op , i was hoping to hear from you .With your experience in the freight forwarding industry , how do you see the time premuim affecting the market ??? If railroads offered at least a daily departure to all destinations , would they gain significant market share ?

i see now as the opportune time for the railroads to try these ideas , less traffic , wagons and locos been mothballed .Presumably plenty of crews avaliable .
I did read in trains that Union pacific was trying a twice weekly service for domestic containers from the LA area to Chicago . why not daily ??? . why not to all points east ???

[2nd trick op]

Anything I could add at this point would be strictly second-hand, but I was able to dig up a link to the subject from two years ago. We've added a couple of knowledgeable rail "insiders" since then, so hopefully Mssrs Cowford or QB 52.32 can add a little.

For now: http://www.railroad.net/forums/viewtopi ... g+freights

[QB 52.32]

This is an interesting point of discussion because it goes to the most fundamental economic, operational and marketing issues surrounding railroading, and, really, to all modes of transportation in general. That's because "vehicle" utilization and load factor is such a big determinant of economic success. So, in the most simplistic form, as you add revenue units (cars, trailers, containers) to a train (start) the better the economic bottom line....and there's a big fixed cost component just to decide to run the train so the benefits of adding revenue units are, in economic terms, leveraged. Unfortunately, the realities of technology, infrastructure characteristics, the network-nature of the business and those pesky customers have to be weighed against simply considering maximizing the number of revenue units you run per train and creates a "tension" between operating (costs) and marketing (revenue) with decisions having a big impact on the bottom line. It is a huge consideration and driver of railroading, both tactically and strategically.

Speaking in a general sense, where you'd find the multiple departures would be in the intermodal networks. Multiple departures driven by different types of business with specific service requirements or, more specifically, within the premium segment of the intermodal traffic, where parcel/mail/LTL carriers need multiple departures to fit into their own sorting/distribution networks. Other than that, in intermodal for the domestic or international general consumer freight traffic, a reliable daily departure suffices (of course where there are multiple departures because of premium traffic, then this general freight also can ride different trains). There is no need within this sub-market for multi-departures, though like David's example in NZ, this network is based upon reliability, reasonable speed, hub-to-hub service which bypasses intermediate yards. Auto networks are about daily departures, reliability and since the customers are quality-sensitive, preventing damage. The loose-car network is about aiming for reliability for a marketplace that is more price-oriented than service oriented, but constrained by a production-oriented assembly-line like system where management works to minimize yard handling. I thnk the overall trend in this network during the past 20 years has to become more scheduled and less tonnage-focused, though with single daily departures in origin/destination pairs. Then there's the bulk networks like coal and grain which operates in conveyer-belt mode, to put it simply.

So, to answer David's question about whether railroads should/will look to add train frequency to grow their marketshare, I would say no. Their performance during this economic downturn has been surprisingly positive and it comes from sizing the networks to meet demand (and furlough crews, mothball power). They have other initiatives underway demanding their management focus such as expanding domestic doublestack networks with their ability to provide much higher units/train capacity and, therefore, lower costs and profitable traffic growth. There is complexity and "rigidity", and high capital costs to make changes, in the railroad system so its not so simple just to add trainstarts, particularly in the loose carload network. And, lastly, I think when they look at the transportation marketplace for opportunities where more-frequent service might work, perhaps in the area of perishables, for example, I believe they see other factors that limit their ability to successfully penetrate that traffic, or to grow existing traffic requiring multiple departures, like the premium intermodal business, they see a "mature" business not presenting lots more opportunity. Trying to grow marketshare by simply increasing train starts, given the high operating cost, structural problems without capital expenditure justification, and, shaky probability that marketshare growth would accrue quickly or at all, and with other proven methods to increase business and profitability demanding management's attention, would make this a very risky proposition.

[Cowford]

I'm not too "up" on trends in train length, but thought I'd throw this link out there: another discussion on the same subject.

http://myprogressiverailroading.com/blo ... rains.aspx

[David Benton]

thanks for the informative answers . I'm not so much psoting against long trains , though i see the problems with them listed on that article as relevant .
I 'm thinking more terminal and equipment costs vs train running costs .
Basically running more unit trains , wether theyre full or not .

[QB 52.32]

From my experience the 3 costs (exclusive of revenue consideratons) that have been considered when it comes to tactical decision-making whether to run a train or not is power (ownership, fuel and mtce) and crews vs. equipment (cars). Simply as an example (with plenty of variables affecting this depending upon the particular situation and historic data), as I recall, power and crew costs for a train travelling app. 1,000 miles were, order-of-magnitude, 3 times the cost of 100 general-purpose per diem freight cars (delayed) for one day. In other words, the cost breakeven in deciding to delay the equipment by one day vs run the train would be 300 per diem cars (app. 2 or 3 times train capy.). So, by way of this example, if you, as a manager, delayed 50 per diem cars by one day to save running a 1,000-mile freight train, your cost savings would exceed your additional costs app. 6-fold (perhaps less if you have to consider using less power, but you get the gist).

Furthermore, management organization and fiscal (budgetary) responsibility could play into the decision-making depending upon what costs the decision-maker was responsible for. For example, if the manager making the run/don't run decision for a train was responsible for crew and power costs or number of train-starts, but not equipment cost, well, you know how the decision-making would be considered.

I believe power and crew costs have tended to dominate the decision-making (even if at an aggregated management focus such as how much power is on-line and how many crews are on the payroll vs. number of train starts), and, therefore, drive minimizing the number of train-starts, unless there are marketing considerations or operating/terminal congestion/capacity issues.

During the past 20 years, with the rise of information technology and management-science decision-making techniques, I believe many of the carriers have moved the decision-making to a planning group who closely monitor the network and create operating plans which are implemented by division and line management. And, I believe, to some degree organizational incentives have been restructured to take into account a total -cost approach and realization that all assets (cars and terminals, too, for example) must be accounted for in decision-making. Along with the influences of greater service-sensitivity (reliability), traffic growth and good financial performance, these changes during the past 20 years has shifted the trend toward more-scheduled operations and away from operations based upon holding trains for tonnage or traffic. Of course, that is dependant upon the particular carrier and how they have organized the decision-making, but I think by-and-large that's the case with the Class 1's. And, while this has lead to greater reliability in running scheduled trains, I don't believe it has lead to multiple-frequency trainstarts within particular networks and origin/destination pairs above and beyond those already required by service differentiation or premium customer requirements. This would seem to reinforce that the additional crew and power costs to increase train frequency still outweighs equipment utilization and terminal productivity (which I believe could be negative, anyhow) benefits and cost reduction.

[2nd trick op]

In the days before the development of long-haul trucking, railroading was much more of a "consumer goods" than a "capital goods" industry, which would lead me to believe that there were greater fluctuations in the amount of freight traffic. Also, the operating rules of that day revolved more around rigid schedules and time-table authority for scheduled trains, including freights, which would be more likely to run in multiple sections in the days before multiple-unit Diesel power.

The defensive build-up at the time of the Vietnam War provided a final surge in rail freight traffic before the long-term post-1945 decline resumed; at that time, I was in my mid-teens and living along PRR's Wilkes-Barre line, its principal gateway into New England. Normal traffic consisted of two daily freights each way from RF&P's Potomac yard in Alexandia, Va. plus a third pair orignating or terminating at Conway yard a little west of Pittsburgh. Occasional day-trips to Harrisburg allowed a glimpse of "the big time".

Traffic on those moves had fallen from an average of perhaps 100-120 cars to about 75 during the recession and steel strike of 1958-59, but by 1964, had recovered and surpassed those figures. As a teenager during the late 1960's I saw as many as 160 cars in times of peak traffic, but the only time I ever encountered what in those days were called 'B' or second sections were on occasional unit coal trains. PRR reportedly also tried running a few longer moves via the Wilkes-Barre gateway while experimenting with mid-train "slave"power in the late 1960's, but I've never heard any first-hand accounts.

On the mainlines of the Eastern trunk roads, it was another story, however. Edward DeRouyn's book "The Pennsy in Chicago" lists about 20 daily eastbound departures from Chicagoland, and while only a handful of them had the same ultimate destination, eastbound traffic could be "balanced out" at Conway, Enola, and possibly other intermedite points. Nevertheless, traffic had grown to the point where multiple sections were common on a number of prominent eastbound "arranged freights", and one of them, NW-88, had provisions for routing via either the former Panhandle line via Columbus or the preferred passenger main via Crestline, now underutilized,

The westbound, usually-empty cars generated by these moves usually moved on non-symbolled extras and traffic was heavier between Friday and Sunday as a rule, sometimes even stretching into the traditionally slow Sunday overnight/Monday morning time frames as a last resort, but for some reason, there was also a period of time when Enola-Pittsburgh maid-of-all-work M-9, a last survivor of the dead-freight era, ran in as many as five sections. I suppose it all varied depending on the preferences of division officials.

PRR had not granted time-table authority to freights for many years, so in single-track territory, all freights, symbolled or not, ran as extras, and one fatal wreck during that era, on the Elmira Branch near Stanley, NY, led to a tightening of the rules in manual-block territory. But on the multiple-track main lines, Automatic Block Signal systems usually made train orders redundant.

[QB 52.32]

Very interesting 2nd trick op. Regarding multiple daily train departures then and now, I think things are probably pretty much the same but the equipment and traffic sources have changed. In other words, for example, NS is now generating multiple departures from the Chicago gateway to the east coast via the ex-Pennsylvania RR (as did PRR), but the traffic that rode in boxcars is now in trailers and containers and traffic once produced domestically is now produced internationally. So within today's multiple intermodal departures one would find on the NS's Chicago gateway-east coast lane, for example, some are designed for premium traffic (replacing the LTL boxcar and mail trains), some for general domestic freight, and some for international general freight (both replacing general merchandise boxcar trains or the empty extras). And, with those modern doublestacks handling twice the freight (app.) for the same trainlength as PRR's boxcar trains. Then, as is now, I don't think we see multiple train departures for the same traffic moving in the same o/d pair or lane unless it has to do with customer requirements or demand exceeding train, line or yard capacities, but not for saving equipment or terminal costs due to the costs to run a train.

[10more years]

Is there any kind of math formula that would relate train length/tonnage/power/ crew costs on a per mile basis. ( I have a visual image in my mind of that guy on the Numb3rs tv show doing equations on the wall.) Just seem to me that the costs per car of a train would go down the longer a train is up to a certain length and then start diminishing.

[QB 52.32]

Yes, you could use per mile or per ton-mile costs and no doubt that drives parts of costing models being used. But, since its an average it would have to be used in certain ways depending upon what you wanted to use the cost information for.

You hit the nail on the head about adding cars to a train- the per car cost drops as you add cars up until you reach the maximum number feasible. The costs move in a "step function" because at certain points you have to add power or, in the case of a local, for example, additional crew. But, just to "start" the train has big fixed costs that don't change (that much) like the ownership and maintenance costs of the power and the basic pay and fringe benefits of the crew no matter how many cars. While I can't provide any specifics, I can say that spreading those power ownership and maintenance and basic pay and fringe benefit crew costs over more cars (earning revenue or moving empty for the next load) outweighs the additional costs like fuel. There's a lot of art to railroad costing and how it's used because it has complexity and the interrelationships of a network, and many ways to consider costs, too.

I have always been impressed during my experiences how big a factor in railroading the number of revenue units (cars, trailers, containers) per train affects the economic outcome and, therefore, how much it drives decision-making. But, if we look around its no different in the airline business, container steamship line business, parcel or LTL business, or, I guess, any transportation business where the decision around originating a "vehicle" and load factor drives the business. Of course, on the other side of the equation is the revenue picture....the higher revenue, the less pressure on load factor. Managing both together has become a big underpinning of the airline industry with demand pricing where sophisticed decision-making systems vary rates to maximize load factor, but that hasn't translated to the same degree to other modes of transporation that I am aware of.

[10more years]

Two more points: A) When train crews were first downsized with the loss of the flagman and brakeman, unions and the workers told management that with crews being downsized, that the crews would not be able to do as much work ( a two or three man crew could not do as much work in a 12 hour day as a full four man crew) and our service to the business community would suffer. Management countered with, "Do as much as you can, and we'll get a relief crew to finish what the first crew can't do. We'll have crews and manpower to finish the work." B) For the last several years, labor has been told that management wants to cut back on "job starts". Evidently, the economics of a "job start" are considerable, much more so than just the costs of employee labor.

[QB 52.32]

I've run across a couple of interesting things regarding train size recently. First, Fred Frailey writes about BNSF and UP's strategic efforts in running large trains in the Jan. '10 issue of Trains. Second, I came across information regarding CSX's operating plan, "One Plan", and how one tenet of their strategy at this point is to maximize the size of merchandise trains. Both reflect not necessarily the tactical economics (though I'm sure that's wrapped up in it, too), but the strategic economics of using large trains to maximize capacity, throughput and velocity and therefore minimize capital costs to build additional infrastructure capacity once anticipated volume growth comes into play.

[Cowford]

I'm not paying too much attention to this issue (shame on me), but to QB's point, the railroads are being (a) creative and (b) proactive in capacity optimization ahead of the next, inevitable traffic surge. They deserve a lot of kudos for showing continued behavior away from the old "respond only when the crisis is upon us" mindset.