Urban Freight Economics: A New Rail Paradigm For Large Lots.

Abstract

The generalization that trucks are more efficient for short-haul freight than railcars does not pertain where shipment size and complementary traffic levels are sufficiently large. In particular, there appear to be enormous opportunities for the competitive movement of truckload- and larger-sized lots by rail within urban areas, which have a sufficient concentration of freight activity to justify minimum right-of-way maintenance and reliable switching service on industrial branch lines. The recognition of this urban rail opportunity, in an age of escalating road degradation and congestion, would have farreaching implications for regulators, urban planners, and rail management, all of whom have heretofore assumed that railroads' urban future lies only in disgorging trucks onto the urban road network from intermodal terminals.

How will freight move in tomorrow's cities? If you look at futuristic Walt Disney models and Star Wars movies, through vacuum tubes and over hovercraft highways. If you read contemporary transportation articles, over conventional roads.

This article will suggest that the emergent realities of twenty-first century congestion will highlight another alternative that is often already superior for large-lot shipments--conventional railroad. A hub-and-spoke system of railroad freight pickup and delivery should be more economical than truck for large lots in urban areas, and could have sufficient reliability and velocity to meet most large-lot shippers' needs. Widespread understanding of urban rail freight economics could alter the course of evolution in the urban landscape.

U.S rail freight revenues of $35 billion in 1998 were about one tenth of the expenditures on truckload lots, which approached $300 billion in 1998. [1] It is the hypothesis of this article that the competitiveness of rail vis-a-vis truck is a function of shipment size and accessibility to rail infrastructure and non-circuitous service, not a function of distance. Therefore, because a large portion of the $300 [+ or -] billion worth of ground transportation now moving in truckload volumes (even a large portion of the $100 [+ or -] billion truckload moving less than fifty miles) is moving from and to urban areas where there is sufficient concentration of transportation demand to support minimum right-of-way maintenance and daily switching service, a huge truckload market would be vulnerable to rail-direct if rational service and rates were offered.

What are the economics of urban freight movement? Comprehensive statistics do not exist, so historically we have believed what seems reasonable. In his 1877 paper, "The Fixation of Belief," [2] Charles Peirce, the great American logician and philosopher, enumerated four ways we can come to accept a belief. Two methods are pertinent here: (1) because a belief is agreeable to reason, or the "a priori method," and (2) because a belief is agreeable to experience, or the "method of science." Peirce clarified in 1893 what he meant by #1 as "merely accepting without question a belief as soon as it is shown to please a great many people very much." By #2 he meant "the method of scientific investigation."

This article will suggest that the current belief about short-haul rail economics--that trucking is inherently less expensive than rail for shorter hauls--may seem reasonable, but that upon investigation this belief needs revision.

THE OLD PARADIGM [3]

"Trucks serve the low density and short-haul freight markets more efficiently than rail..." That is the old paradigm as stated in Improving Railroad Productivity, [4] a government report prepared for President Nixon's Council of Economic Advisers in 1973. The report committee was chaired by John R. Meyer of Harvard University.

As much as any other individual, Meyer articulated the current belief about urban freight transportation economics. In 1959, he co-authored The Economics of Competition in the Transportation Industries, [5] which concluded that piggyback should replace rail-direct merchandise service on low-density spur lines and industrial sidings. [6] In 1971, Meyer co-authored The Role of Transportation in Regional Economic Development, which flatly stated:

The railroads' role in carrying manufactured commodities is limited by high terminal cost, large capacity, and low speed. The line haul costs of rail transport are lower than those of trucking, but rail terminal costs are much higher....In particular, piggyback decreases rail terminal costs (although piggyback terminal costs are higher than truck terminal costs) and decreases the railroads' disadvantage in pick-up and delivery service. [7]

In Chapter V of Improving Railroad Productivity, Meyer prophesied:

...[A]s the railroads continue to specialize as carriers of a limited number of bulk commodities and of long hauls of manufactures, the prospect is for continuing concentration of rail movements over a more limited rail network.

This outcome is based on the premise that road use is less expensive than rail use for short-hauls. Taken at face, this premise may seem sound; indeed, it is a common explanation for railroads' short-haul market loss. However, there is reason to question the premise. This article will investigate its validity for freight movements originating and terminating in urban areas.

COMPARING URBAN TRUCK AND RAIL COSTS

Historical Cost Studies

A bad place to start an inquiry into the economics of urban freight would be the railroads' own switching cost studies. The Interstate Commerce Commission's Rail Form F methodology for switching costs was almost never used because it took too much time and effort for results that were too specific for broad application. Railroads invariably created switching cost studies that served their commercial purpose of setting high reciprocal switching fees, making it difficult for other railroads to reach their on-line traffic.

A good place to start is John Meyer's 1959 book. For urban truck data, he used a 1954 ICC report which estimated that the terminal cost for truck pickup or delivery of general freight was 84[cts.] per ton. [8] That is, using data from conventional accounting methods and not including cross-dock, billing, collection, detention, etc., the out-of-pocket truck operating cost of picking up a dry van trailer in the origin terminal area or delivering it in the destination terminal area was 84[cts.] per ton for a truckload.

Meyer made no attempt to calculate the comparable variable cost of urban boxcar pickup or delivery. All his rail costs were estimated as system-wide averages, using the statistical costing method of multiple regression. However, one regression result is of particular interest -- one which estimated that the variable cost of rail terminal operations would have been $17.83 per yard engine hour. [9] That is, the variable rail pickup or delivery comparable to dry van pickup or delivery depends on how much engine time was required to classify and pull or spot a ton of boxcar freight. In Meyer's words, "It is difficult to generalize about carload terminal expenses since they vary greatly depending on the volume of transportation requirements, the distance of a plant from a classification yard, and the concentration of plants in one area." [10]

However, this variable expense of $17.83 per yard engine hour gives us a good clue. Table 1 shows permutations of variable rail cost given different average tons of lading per boxcar and different numbers of cars classified and handled outbound or inbound by an eight-hour crew. Notice that all combinations to the right of the stair-step line generate a rail cost less than the truck cost of 84[cts.] per ton. For example, if local crews could classify and spot or pull six boxcars in eight hours, and if the average lading weight were forty-five tons per car, then the rail variable terminal cost in 1954-55 would have been 53[cts.] per ton (see boxed number in Table 1), which is 31[cts.] per ton under the truck cost of 84[cts.] That 31[cts.] would have been available for fixed costs like track maintenance, a lower rate, and profit. The point is: Dr. Meyer's conclusion does not follow from his own data and sensibilities. In mid-century, rail was not the unambiguous loser in short-haul economics.

1999 Cost Study: Data

Would a different and updated approach yield the same result? Instead of regression analysis [11] for rail, both urban truck and urban rail will be compared using accounting costs from direct observation. First, this complex problem is simplified by eliminating all costs with little intrinsic difference between rail and truck; thus excluded from this analysis as irrelevant to the comparison will be sales, administrative, billing, and collection costs. [12] Next, service differences must be addressed. Cass Information Systems estimates that non-transportation logistical costs, like inventory carrying costs and warehousing (which are highly dependent on service speed and reliablity), represented $334 billion, or 37 percent, of total U.S. logistics costs of $898 billion in l998. [13] Therefore, for a fair comparison without consideration of non-transportation costs, all rail costs are here predicated upon the normative expenses of daily and dependable rail switching service, which is often not now available for institutional reasons.

Pickup and delivery service can fall into the following categories:

either 1. drop and hook (where the tractor or engine uncouples from the brought equipment and leaves empty or with a different piece of rolling stock),

or 2. "live" (where the driver or crew remains in attendance while equipment is loaded/unloaded), and

either 1. with loading/unloading assisted by transportation personnel,

or 2. with loading/unloading unassisted by transportation personnel.

Truck service can be any combination, [14] although tank truck deliveries are usually "live" since the cost of stationary silos is so much less than the cost of tank trailers. Rail service is invariably drop and hook with unassisted loading/unloading. [15] The difference in truck and rail service offerings makes comparison problematic. To create parity, this article assumes drop and hook with unassisted loading/unloading for all traffic, with built-in equipment ownership opportunity costs of one day for truck versus a variable number of days for rail equal to the ratio of railcar-to-truck capacity for the equipment type in question (as calculated in Table 2, line C). Another simplifying assumption will be made initially: that shipment lot size can fill a rail car. This assumption will be relaxed for some concluding observations about potential market size.

Urban Trucking Costs

Normative urban truck costs in 1999 for pickup or delivery are estimated in Table 2 for various equipment types. The ratio of railcar capacity to truck capacity is calculated; the pickup or delivery cost for one truck is calculated using variable operating expenses; and the capacity ratio is multiplied times the pickup or delivery cost to obtain the equivalent cost per railcar.

A 1999 survey by Martin Labbe Associates found that the average time elapsed for a drop and hook was 2.83 hours at origin and 2.95 hours at destination. [17] Accordingly, the underlined row in Table 2, which allows three hours for a ten-mile urban trip in, drop and hook time, plus a ten-mile urban trip out, will be used for comparison to rail costs. This row also most closely comports with proprietary data collected by the author when he was in charge of Conrail's Flexi-Flo [R] transloading terminals between 1990 and 1999, and with the Transportation & Inventory Costing model of the Federal Highway Administration's Office of Policy and Development, which assumes an urban intermodal drop and hook charge of $100 plus $1.51 per mile. The weighted average in the three-hour row for the six equipment types, based on the prevalence of each equipment type, [18] is $334. As previously stated, this railcar-equivalent out-of-pocket cost does not include overhead items such as sales, administrative, billing, and collecti on costs. [19]

Urban Rail Costs

The rail costs fixed for each eight-hour local crew, fixed for each locomotive day, and variable with car usage are estimated in Table 3. Each crew's costs of $668 will be spread over the number of cars it handles. Daily engine ownership of $125 will be spread over one, two, or three crews. Car ownership will be $42.66, the weighted average ratio of car capacity to truck capacity (2.89) times the weighted average per diem ($16.16). Car maintenance will be miles traveled times 8.71 [cts.] the weighted average.

To parallel the truck calculation, rail overhead is not considered for sales, administrative, billing, and collection costs. However, "fixed" costs for fight-of-way maintenance need to be considered for rail. Take the worst-case scenario, where the right-of-way is not used for through trains, so that the entire cost of track must be covered by local traffic. [21] The cost to maintain this track, not build track, is the relevant cost in this case because the decision to build new track would not be fixed but discretionary based on special circumstances. Industrial rail lasts indefinitely if properly maintained, and normalized annual maintenance for Class I track (10 mph maximum) is only $5,000 per mile. [22] Therefore, track ownership and maintenance is a minor consideration for urban rail (assuming that opportunity costs for existing track structure components and right-of-way are insignificant).

Comparison: Truck versus Rail

Urban rail costs discussed earlier produce the carload cost permutations in Table 4. These are comparable to the urban truck trip-sensitive costs calculated earlier. Everything depends on how many miles of track must be maintained and traversed and how many carloads can be handled per crew shift.

The permutations below the line in Table 4 show the area where the average urban rail cost is less than the weighted average urban rail-equivalent truck cost of $334. That is, where rail volume exceeds four carloads for a crew on twenty miles of industrial track, rail would be expected to have an economic advantage. This is despite the conservative nature of rail assumptions, such as one crew handling only up to ten cars at a time, and the use of union rates for rail and average rates for truck. Rail is even superior where it is least competitive -- general traffic limited by cubic capacity (where the rail-equivalent truck cost in the three-hour row in Table 2 was $169) over a large range of rail cost permutations. [23]

Sensitivity analysis with various cost factors shows the rail urban advantage to be very robust, despite variations that could alter a particular analysis, such as the following:

* those that might decrease truck costs, such as when the trailer is not dropped off and picked up later, but the driver stays during "live" loading/unloading;

* those that might increase truck costs, such as detention for the wait during "live" loads (which is often imbedded in a linehaul quote and not explicity stated);

* those that might decrease rail costs, such as when the shipper and receiver are served by the same rail crew (i.e., same "spoke" from the "hub"), or when rail motive power could be utilized during multiple shifts, or when labor is non-union; and

* those that could increase rail costs, such as when there is excessive rail circuity, new track construction, or the move is intra-terminal (all in one metropolitan area) so that an interchange between railroads or different "hubs" is required.

These and innumerable other move-specific factors change the economic complexion of each particular move, but do not disturb the magnitude of rail's range of expected advantage for carload lots. It should be a common occurrence to find cases where capital projects can be justified and the fixed rail plant ratcheted outward.

Comparison: Rail Intermodal versus Rail Direct

The rail-direct cost advantage is even more pronounced versus intermodal traffic. Urban intermodal costs would exceed urban truck-direct costs as shown in Table 5. Intermodal costs exceed urban truck costs by row C in Table 5, making rail-direct just that much more economical when compared to intermodal. It is the available speed, reliability, and third-party service options of intermodal which make it currently attractive, not its inherent cost advantage over rail direct.

URBAN RAIL SERVICE POTENTIAL

A hub-and-spoke system like the one Federal Express invented for air freight would provide next-day or better service for rail moves within an urban area. If all local crews worked out of one yard for each urban area, if each customer were served at least once a day, and if all cars were classified each night, intra-urban rail service would be no worse than next-day. Table 4 on Pickup or Delivery Cost per Railcar anticipates the costs for up to three switches per day on each "spoke" in order to provide daily or better frequency of service.

URBAN DEVELOPMENT ISSUES

As urban areas sprawl outward, road congestion grows geometrically. A 1997 Federal Highway Administration study estimated that 60 percent of urban interstate paving is in "fair" condition. [25] The Department of Transportation estimates that "[b]etween 1985 and 1993, delays jumped 41 percent in the Washington, DC area, 39 percent for Greater New York City, 30 percent for Metropolitan Chicago, 21.5 percent for San Francisco, and 16 percent for Los Angeles." [26] More road use is not being met with sufficient road funding. The same DOT report says, "Industry experts estimate that there is a $300+ billion investment shortfall for highways and bridges alone.... Given current budget and tax pressures, indications are there will be no new public investment program to close these investment gaps." [27] A separate DOT report estimated that 1996-2015 annual highway investment of $80 billion would be optimal, with $43 billion of annual investment necessary just to maintain 1995 conditions. [28] If TEA-21 is fully fund ed, it will provide an average of only $26 billion in annual highway funding from 1999 to 2005. [29]

We are entering a new era of potholes, rough pavement, extended-hour gridlock, bridge weight restrictions, no-truck lanes and zones, and truck curfews. Expanded car and small truck use and restricted public budgets combine to spell cost increases for the future users of urban roads. Conversely, railroads control their own right-of-way.

THE NEW PARADIGM

We have arrived at a new freight paradigm, which will become more and more pronounced:

Railroads serve short-haul freight markets more inexpensively than trucks when lots are sufficiently large (such as truckload quantities), when there is sufficient density of business (such as in urban areas), and when origin and destination are reached by rail right-of-way.

Sanity Check

The new paradigm immediately begs the question: How can these mathematics be right, given the seeming reality of unchangable current belief? [30] This is a question better answered by behaviorists and students of human nature than by mathematicians and economists. Returning to Charles Peirce, we are told that perhaps the most pragmatic force of all is the power of prior belief:

...[T]he instinctive dislike of an undecided state of mind, exaggerated into a vague dread of doubt, makes men cling spasmodically to the views they already take. [31]

It is not hard to understand why truckers are only too happy to view their mode as superior for shorter distances. Shippers have no reason to question this, since they are not solicited for rail short-hauls, not provided rail service for short-hauls, and not afforded a staff to research administratively complex projects like building or reactivating sidetrack infrastructure (particularly when all the initial benefit usually accrues to the shipper who pays the freight). Shippers are often left without the option of using rail even if they want to, a solvable problem obscured by the old paradigm.

This leaves one primary culprit--the railroads themselves. They have locked in [32] a belief which closes them out of much of a $300[+ or -] billion truckload market. This bears repetition. Railroads could be economically competitive with trucks for a market that is perhaps ten times their current total revenues of $35 billion, but they have largely chosen not to compete. Why? The answer may have to do with the inherent complexity of railroading and the organizational structure which locked in as a result.

Complexity and Diversity

A complex adaptive system is a network of agents that constantly rearrange themselves into systems that are more and more complex, where each agent faces the perpetual novelty of an essentially infinite space of possibilities. Examples are ecologies, economies, and individual industries. The higher the complexity, the more important diversity is for survival. As Doyne Farmer, physicist and pioneer in the "connectionism" of complexity theory, expressed it, "[E]volution thrives in systems with a bottom-up organization, which gives rise to flexibility." [33] Brian Arthur (see note 32) elaborated, "So the question is how you maneuver in a world like that. And the answer is that you want to keep as many options open as possible." [34]

The complexity of running railroads instigated our current economic era. In the 1880s, U.S. railroads were instrumental in giving rise to the line-and-staff corporate structure, modern trade unions, divisional budgets, standardized accounting, investment banking, incorporation law, anti-trust law, labor law, government regulation, group retirement plans, time zones, etc. These conventions are ways to deal with business complexity, but the solutions worked out are not final, and sometimes they give a false sense of security or get in the way of necessary change.

Before there was truck competition, railroad organizations evolved by trying to divide and conquer complexity. Railroads were split into operating and commercial responsibilities, which subdivided again into the traditional "stovepipes" of functional railroad departments. This compartmentalization at first helped, but now allows the intermediation of suboptimizing.[35] decision makers. That is, specialized individuals who are not responsible for profit set artificial day-to-day incentives that can reduce profit.

For example, railway commercial officers often mis-specify their goal as increasing revenue per car trip. This is simpler to measure than the financial goal as increasing profit per unit of asset-time. Trying to increase revenue per trip creates a bias against short hauls, although it is better to have four round trips of $500 revenue with $250 contribution each than it is to have one round trip of $1,500 revenue with $750 contribution over the same time frame.

The reality is that railroads could compete. The natural outcome of this reality would be a different urban landscape from the one we have previously envisioned--a landscape with fewer large truck trailers, with industrial growth clustering along existing rail rights-of-way, and with private enterprise paying for the maintenance of congestion-free freight access.

Similarly, railway operating officers often mis-specify their goal as increasing efficiency. This is simpler to manage than increasing profits. Short-haul traffic is less efficient because it uses more assets per unit of production (a handicap more than offset by short-haul's greater revenue-producing capability). John Meyer said himself in Improving Railroad Productivity that railroads use "cost-effectiveness" instead of "cost-benefit" to measure themselves, and thus end up focusing on things like average length or weight of train, instead of profit per asset-unit-of-time. [36]

Most large railroads have created profiles of what traffic they want, based on productivity measurements instead of profitability. [37] John W. Barriger, who was a senior officer of six different shorter Class I railroads in the 1940s-to-1970s, defined the productivity of inputs as "technical efficiency" and the profitability of outputs as "commercial efficiency." [38] The confusion of these two concepts would lead to an irrational bias against short hauls when they are lucratively competitive with truck but more costly than long hauls per unit of measurement.

The inertia of existing belief underlies the ideas of two pioneers in the art of advertising, Al Ries and Jack Trout. They coined the word "positioning" to stand for the pre-eminence of our first association with some product or service. "The easy way to get into a person's mind is to be first." [39] "Not only does the human mind reject information which does not match its prior knowledge or experience, it doesn't have much prior knowledge or experience to work with. In our overcommunicated society, the human mind is a totally inadequate container." [40]

Therefore, it is not surprising that railroads have been their own worst enemy in promoting the old paradigm. In the nineteenth century they were rewarded for making their daily task efficiency; short-haul statistics are not as efficient as long-haul statistics; ergo, railroads in due course fell into the belief that they are not competitive in the short-haul a priori.

CONCLUSION

The belief that trucks are more efficient than rail for low-density and short-haul freight is half-truth covering up half-falsehood. Truck is superior for low-density moves, but rail is superior for short hauls where density (lot size and complementary volume) is sufficient. That sufficiency can usually be met in urban areas, so that lot size may not have to be a full railcar or even a full truckload in order to warrant rail movement there. In fact, there should be a niche for lower-purchase-price, lower-capacity rail equipment such as the bygone 40-foot boxcar of 3,700-cubic-foot capacity.

The great impediment for railroads in the truckload short-haul market is their failure to see the opportunity and offer service. Abandoned urban right-of-way, infrequent and unreliable switching service, and nonexistent solictation are the legacies of a self-fulfilling prophecy about their own short-haul competitiveness.

Mr. Erickson is principal, Rail Cents Enterprises, Inc., P.O. Box 235, Wallingford, Pennsylvania 19086.

ENDNOTES

(1.) 1998 truckload revenues (i.e., not including less-than-truckload) are estimated at $290 billion as follows:

U.S. Census Bureau, 1997 Census of Transportation, Table 1 Eno Transportation Foundation, 1998 est. General freight trucking for-hire local truckload $8.46 billion local truck: $144 billion

-- = 69.1% [right arrow] x 69.l%

General freight trucking, for-hire total local $12.25 billion $100 billion

(local truckload)

General freight trucking, for-hire, intercity T.A. $51.14 billion intercity truck $283 billion

-- = 67.2% [right arrow] x 67.2%

General freight trucking, for-hire, total intercity $76.15 billion $190 billion

(intercity truckload)

Total 1998 Truckload: $290 billion

(2.) Charles Sanders Peirce, "First Paper--The Fixation of Belief' from "Illustrations of the Logic of Science," Popular Science Monthly (November, 1877): pp. 1-15.

(3.) "...[A] paradigm is 'the basic way of perceiving, thinking, valuing, and doing associated with a particular vision of reality. A dominant paradigm is seldom if ever stated explicitly; it exists as unquestioned, tacit understanding that is transmitted through culture....'" Willis Harmon's definition in An Incomplete Guide to the Future, as quoted by Joel Barker in Paradigms, the Business of Discovering the Future (New York, N.Y: HarperCollins Publishers, 1992), pp. 31-32.

(4.) John R. Meyer, Chairman, Task Force on Railroad Productivity, "Improving Railroad Productivity, Final Report" to The National Commission on Productivity and The Council of Economic Advisers (November, 1973), p. 159.

(5.) John Meyer, Morton Peck, John Stenason, and Charles Zwick, The Economics of Competition in the Transportation Industries (Cambridge, MA: Harvard University Press, 1959).

(6.) Conversely, Meyer's studies of urban passenger economics were not generally accepted. In 1965 he co-authored with John Kain and Martin Wohl The Urban Transportation Problem, which promoted exclusive-use busways, of which only ten miles in Pittsburgh and eight miles in Miami have been built.

(7.) John Meyer, Gerald Kraft, and Jean-Paul Valette, The Role of Transportation in Regional Economic Development, a Charles River Associates Research Study (Lexington Books, D.C. Health & Co.) p. 41.

(8.) Interstate Commerce Commission, Cost Finding Section, "Cost Study of Class I Motor Carriers of General Freight in the Middlewest Territory-Year 1953" (Washington, D.C., 1954), p. 7, Table 2, 1/2 col. 2.

(9.) The Economics of Competition in the Transportation Industries, Table 1, page 46, from data reported by Class I railroads to the ICC for 1954-55. In 1989, the ICC adopted the costing methodology used in this book, explaining it in a footnote at 5 ICC2nd 894, 896: "Regression analysis is a method for establishing a statistical relationship between the mean of a 'dependent' variable and the values of set 'independent' variables. In URCS [Uniform Rail Costing System] regression analysis is used to attribute railroad expenses to capacity and output variables. The portions of expenses thus dichotomized are referred to as 'fixed' and 'variable' expenses, respectively."

(10.) The Economics of Competition in the Transportation Industries, p. 151.

(11.) Regression analysis has the biases of measuring current practice, not potential, and of assuming linearity, although rail operations appear to be irregular stepwise functions. Whether linearity or discontinuity is more appropriate is perhaps the oldest argument in mathematics, starting with the four paradoxes of Zeno of Elea in the Fifth Century B.C. See Men of Mathematics, by E. T. Bell, Simon & Schuster, New York, 1937, pp. 23-25.

(12.) It could be argued that larger equipment capacity makes rail more efficient for billing and collection, or that greater good will makes truck sales and administration more efficient, but these nuances are beyond the scope here.

(13.) Cass Information Systems, as referenced by John Schulz in Traffic World, June 14, 1999, p. 20.

(14.) The "1999 Dry Van Drivers Survey" prepared for the Truckload Carriers Association by Martin Labbe Associates of Ormond Beach, Florida found that 42 percent of deliveries to survey respondents were drop-and-hook (p. 12).

(15.) This limitation is not innate, but institutional--a legacy of work-rules, self-image, and decision makers in cost centers driven by arbitrary efficiency measurements, instead of in profit centers driven by customer utility.

(16.) Sources:

A. car capacities: small end of various railcar types from The Official Railway Equipment Register (i.e., standard 60' boxcars actually average about 6,300 cubes, high-roof 60' boxcars about 7,300 cubes, versus 6,150 cubes used)

B. truck capacities: 53' dry van cubic capacity--1999 production of Wabash National, Lafayette, Indiana; all other truck capacities--80,000 lbs. less common cab & trailer tare weights

D. driver hourly rates: from the ATA's "1997 Driver Compensation Study," page 14 of 1999 American Trucking Trends, assuming 2,500 hours worked and inflation 1997 to 1999 of 3.8 percent per CPI-U--rates corroborated by $15.10/hr. average truckload driver wage per Transportation Technical Service's (TTS) 1999 The Motor Carrier industry in Transition, which is below the Teamsters' 1998 National Master Freight Agreement of $18.46 per hour

E. labor fringe rate estimated as 31 percent per TTS's 1999 The Motor Carrier industry in Transition, p. 153

F. tractor values: from interviews with Peterbuilt and Mack Truck sales and leasing managers

G. registration: from ATA's 1999 American Trucking Trends, p.21, $1,567 average divided by 3,120 hours; insurance: estimate of $5,000 per year from truck line interviews divided by 3,120 hours

H. special services: refrigerated trailer fuel of 82[cts.]/hr. from Merchants Despatch Transportation (MDT) service experience of 3/4 gallon per hour @ mid-1999 diesel price of $1.09/gal. from Traffic World, July 5, 1999, p. 50; tank truck blower costs of $1.01 from $9,500 investment divided by three years of 261 twelve-hour days, or 9,360

I. trailer values: U.S. Dept. of Commerce, Current industrial Reports--Truck Trailers, 1998, plus inflation 1998 to 1999 of 2.2 percent per CPI-U, plus refrigeration units for reefers of $16,786 per MDT records, @ 25 percent lease/year

J. fuel & repairs: assuming twenty-mile urban round trip @ type-specific miles per gallon according to TTS/A.T. Kearney 1995 research ("Private Fleet Benchmarks of Quality and Productivity") for the National Private Truck Council, 1995, pp.47, 56, 59, and 62, all @ $1.09 per gallon per H above; assuming repairs at two times fuel cost

K. matrix cell calculation: C x [((drop & book hours) x (D+E+F+G+H) + I + J]

(17.) "Martin Labbe Associates, op. cit., bar graph p. 12.

(18.) TTS's 1999 The Motor Carrier industry in Transition, chart page 55 from 1992 Truck Inventory/Use Survey, numbers tractor-trailer units at 431,700 vans; 250,400 flatbeds; 115,400 reefers; and 93,000 tank trucks.

(19.) Historically, the ICC assumed 90 percent of truck costs to be variable, 10 percent fixed. The TTS study cited found 9 percent of truck expenses to be overhead (p.153). It is likely that the out-of-pocket variable costs understate the actual cost to society. The U.S. DOT's 1997 Federal Highway Cost Allocation Study estimated that trucks actually pay tax covering only 80 percent of the road construction, wear, and operating costs they occasion (Table ES-5, p. ES-9).

(20.) Sources:

engineer: Brotherhood of Locomotive Engineers, 1 July 99 daily rate for yard service, five-day week, for less than 500,000 lbs. on drivers, including differential for no fireman

conductor: United Transportation Union, 1 July 99 daily rate for local freight less than 100 miles

supervision: Association of American Railroads, "Analysis of Class I Railroads, 1988," page 13, Line 234 (Total Transportation Labor) as a markup over Line 234 minus Line 228 (Administrative Support of Operations Labor)--to cover dispatching, training, and managing

engine ownership: survey of typical lease rates for medium maintenance GP-9 or GP-38

engine maintenance: survey of typical maintenance costs for medium maintenance GP-9 or GP-38

engine fuel: 55.5[cts.] per gallon per Daniel L. Keen of Policy, Legislation and Economics Department, AAR (less than truck due to buying power of larger, longer-term contracts, and no 20[cts.] user charge)

car maintenance: per mile and per day rates estimated by data available--average rates experienced by Conrail under deprescription contracts in August 1996 for foreign cars, using 60' equipped box rate, refrigerator rate, 100 ton [less than] 61 ft. low-side gondola rate, equipped flat rate and [greater than] 5000 cu. ft. covered hopper rate

(21.) Although the overhead burden to the railroad for track ownership and maintenance is spread farther when a side track is located on a line used for through freight, passenger, or commuter operations, there is an increased opportunity cost to the customer because the time windows when the customer can receive local service are restricted. Therefore the worst-case fixed-cost scenario to the railroad is the potential best-case opportunity-cost scenario to the customer because with exclusive use for local freight the customer could negotiate whatever service is required.

(22.) The $5,000 per mile estimate is from the unpublished data of Randolph R. Resor, at Zeta-Tech Associates, Inc. of Cherry Hill, N.J. Property taxes are not a relevant expense of industrial right-of-way in most states, and they are not considered here. Forty states use the "unit valuation" or "unitary assessment" method of railroad taxation, where the state levies a tax based on a somewhat arbitrary valuation of the entire railroad, and then distributes these taxes to local jurisdictions based on their local percentage of track mileage in state. Five states have a gross receipts tax or income tax instead of a going-concern valuation--CT, IN, LA, ME, and MD. This means that in forty-five states operating railroad property is exempted from local assessment and the original taxes have nothing to do with local mileages. Only four states have local valuation and taxation based on operating property--MT. NY, RI and TX (plus two cities in PA), but these taxes are restricted by the Railroad Revitalization and Regulatory Reform Act of 1976. One state, HI, has no freight railroads.

(23.) Rail rates need not, indeed should not, be a uniform markup over average cost. Rail rates should depend on the utility to the specific customer, and must only exceed incremental (including any solely-related capital) costs in order to benefit the railroad. Therefore, moves priced under the average for the appropriate matrix cell may still contribute to an irreducible lump of cost (such as for a crew, for a locomotive, for a branch line) and be desirable, if they can be balanced by cars over the average by the same amount. In an article in Economic Journal in March 1927, Frank Ramsey demonstrated that when a fixed amount must be raised, the socially optimal prices are inversely proportional to the customer's elasticities of demand.

(24.) Sources:

A. lift or transload cost in terminal: survey of charges typical to various terminal types, with reefer charge including $18 (per Merchants Despatch Transportation data) and with average terminal rates per Conrail data

B. ratio of trucks to car capacity: Table 2, row C, except for dry van limited by cubes, where 48' length with 3,550 cubic foot capacity (per The Official Intermodal Equipment Register) is substituted for 53' length with 4,100 cubic feet, changing ratio from 1.5 to 1.7

C. truck cost: Table 2, row E, except for dry van limited by cubes as explained for B above

(25.) U.S. DOT, Federal Highway Administration and Federal Transit Administration, "1997 Status of the Nation's Surface Transportation System: Condition & Performance, A Summary," (Washington, D.C., 1997), chart p. 3.

(26.) U.S. DOT, Federal Highway Administration, "U.S. Freight: Economy in Motion, 1998" (Washington, D.C., 1998), p. 59 quote from National Transportation Statistics 1996, report of the Research and Special Programs Administration and Bureau of Transportation Statistics, (Washington, D.C., 1997).

(27.) "Ibid., pp. 61-62.

(28.) U.S. DOT, Federal Highway Administration and Federal Transit Administration, 1997 Status of the Nation's Surface Transportation System: Condition & Performance, A Summary, (Washington, D.C., 1997), p.7.

(29.) "From a Frank Wilner report in Traffic World, June 15, 1998, p. 15, concerning the signing of the Transportation Equity Act for the 21st Century.

(30.) Interestingly, the Transportation Technical Services' 1999 report previously cited has a table on page 17 showing, without explanation, that rail dominates across all distances for shipments in excess of 90,000 pounds. The FHA's 1998 Economy in Motion report, also without explanation, copies this table on page 24.

(31.) "Op. cit., Charles Peirce, p. 7.

(32.) "Lock-in" is an economic principle propounded by the Santa Fe Institute, and described by its originator, W. Brian Arthur, in his article "Positive Feedbacks in the Economy," in Scientific American, February 1990, p. 92. "...once random economic events select a particular path, the choice may become locked-in regardless of the advantages of the alternatives."

(33.) "J. Doyne Farmer, as quoted by M. Mitchell Waldrop in Complexity, (NY: Simon & Schuster, 1992), p. 294.

(34.) Brian Arthur, as quoted by M. Mitchell Waldrop in Complexity, p. 333.

(35.) "The term "sub-optimization" was coined by C.J. Hitch for an article in the Journal of the Operations Research Society in 1953. It means creating rules for local optimization which preempt global optimization.

(36.) Meyer, op cit., p. 37. Service quality measures get scrambled as well as profit measures. To a customer, service quality means things like dock-to-dock transit time, frequency, standard deviation, and the availability of clean equipment. To a railroad, service quality means productivity measures like dwell time in yards, bad order ratios, and loss and damage claims. As the General Accounting Office stated in its April 1999 report Railroad Regulation, "... the railroad industry has been reluctant to develop specific service measures....In reaction to widespread criticism of rail service, however, railroads have developed four performance indicators... [which] are more an evaluation of oprating efficiency than of quality of service." (p. 66).

(37.) Nature has no such dictum. The only productivity in biology is the end-game of producing progeny. In her details, nature seems very wasteful. For example, a large percentage of the very essence of life, DNA-- probably over 50 percent of DNA in complex organisms--seems to serve no useful purpose (except perhaps as grist for favorable future mutations). The co-discoverer of the structure of DNA, Francis H. C. Crick, was surprised by the high incidence of useless or "selfish" DNA. He wrote that "The spread of selfish DNA sequences within the genome can be compared to the spread of a not- too-harmful parasite within its host." (Nature, Vol. 284, p. 605, 17 Apr 1980, L. E. Orgel & F. H. C. Crick) Conversely, Charles Darwin noted the curiously profligate "economy" of nature. "The principle of the economy of growth...by which the materials forming any part, if not useful to the possessor, are saved as far as possible, will perhaps come into play in rendering a useless part rudimentary." (The Origin of Species, Chapter XIV).

(38.) John Walker Barriger, Super-Railroads, (NY: Simmons-Boardman Publishing Corp., 1956), p.12.

(39.) Al Ries and Jack Trout, Positioning: The Battle for Your Mind (NY: McGraw Hill Book Co., 1981), p. 21.

(40.) Ibid., p. 35. Table 1. Variable Terminal Cost per Ton of Lading for Various Car Weights & Cars/Crew at $17.83 per Hour for an Eight-hour Crew No. of Cars/Crew

1 2 3 4 5 6 7 Lading Tons/Car

30 $4.75 $2.38 $1.58 $1.19 $0.95 $0.79 $0.68

35 4.08 2.04 1.36 1.02 0.82 0.68 0.58

40 3.57 1.78 1.19 0.89 0.71 0.59 0.51

45 3.17 1.58 1.06 0.79 0.63 0.53 0.45

50 2.85 1.43 0.95 0.71 0.57 0.48 0.41

55 2.59 1.30 0.86 0.65 0.52 0.43 0.37

60 2.38 1.19 0.79 0.59 0.48 0.40 0.34 No. of Cars/Crew

8 9 10 Lading Tons/Car

30 $0.59 $0.53 $0.48

35 0.51 0.45 0.41

40 0.45 0.40 0.36

45 0.40 0.35 0.32

50 0.36 0.32 0.29

55 0.32 0.29 0.26

60 0.30 0.26 0.24 Table 2. [16] Urban Truck Costs Per Railcar-Equivalent

Std. 60 ft. Std. 60 ft. Std. 50 ft. Car type Boxcar Boxcar Reefer Commodity general general perishable Trailer type dry van dry van reefer van Limited by cubes pounds pounds A. Car capacity 6,150 140,000 130,000 B. Truck capacity 4,100 48,000 45,000 C. Ratio (A[divided by]B) 1.5 2.9 2.9 D. Driver hourly rate $14.80 $14.80 $14.47 E. Fringes/hr. @ 31% 4.59 4.59 4.49 F. Tractor value/hr. 9.08 9.08 9.08 G. Regis. + insur./hr. 2.10 2.10 2.10 H. Special svc./hr. 0.82 I. Trailer value/day 11.60 11.60 23.35 J. Fuel + repairs/trip 9.38 9.38 9.52 K. P/U or delivery if 1 hour drop & hook $77 $150 $185 2 hours " 123 238 275 3 hours " 169 327 365 4 hours " 215 416 454 5 hours " 261 504 544 6 hours " 307 593 634

Std. 52 ft. Bulkhead Cov. Hop. Car type Gondola Flatcar or Tank Commodity steel lumber bulk Trailer type flatbed flatbed tank Limited by pounds pounds pounds A. Car capacity 175,000 190,000 192,000 B. Truck capacity 50,000 50,000 48,000 C. Ratio (A[divided by]B) 3.5 3.8 4.0 D. Driver hourly rate $13.97 $13.97 $14.26 E. Fringes/hr. @ 31% 4.33 4.33 4.42 F. Tractor value/hr. 8.55 8.55 9.08 G. Regis. + insur./hr. 2.10 2.10 2.10 H. Special svc./hr. 1.01 I. Trailer value/day 11.54 11.54 29.27 J. Fuel + repairs/trip 10.34 10.34 9.84 K. P/U or delivery if 1 hour drop & hook $178 $193 $280 2 hours " 279 303 404 3 hours " 381 413 527 4 hours " 482 523 651 5 hours " 583 633 774 6 hours " 685 743 898 Table 3. [20] Estimated Rail Costs One BLE engineer, eight hours @ $160 + 50% fringes $240 One UTU conductor, eight hours @ $136 + 50% fringes 204 Supervision @ 9.3% markup over engineer + conductor 41 Motive power maintenance, 4-axle GP-9 or GP-38, 8 hours 100 Fuel, eight hours, 150 gallons

@ 55.5[cts.] sbtotal of rail costs 83 fixed for each eight-hour crew $668 Daily engine ownership cost for a 4-axle GP-9 or GP-38 $125 Rail costs variable with car usage:

Std. 60 ft. Std. 60 ft. Std. 50 ft. Std. 52 ft. Car type: Boxcar Boxcar Reefer Gondola Per mile: 7.53 [cts.] 7.53 [cts.] 12.79 [cts.] 13.07 [cts.] Per diem: $16.23 $16.23 $15.80 $15.13

Bulkhead Cov. Hop. Car type: Flatcar or Tank Per mile: 5.75 [cts.] 7.21 [cts.] Per diem: $15.89 $18.01 Table 4. Pickup or Delivery Cost per Railcar (Classification, Spotting or Pulling)

One-way miles

2 4 6 8 10 12 14 16 crews/da cars/da

$

1 2 458 478 497 516 536 555 574 594

1 4 251 260 270 280 290 299 309 319

1 6 181 188 195 201 208 214 221 227

1 8 147 152 157 162 167 172 177 181

1 10 126 130 134 138 142 146 150 154

2 12 168 171 175 178 181 185 188 191

2 14 150 153 156 159 162 164 167 170

2 16 137 139 142 144 147 149 152 155

2 18 126 128 131 133 135 138 140 142

2 20 118 120 122 124 126 128 130 132

3 22 141 143 145 147 149 151 153 155

3 24 133 135 137 138 140 142 144 146

3 26 126 128 129 131 133 134 136 138

3 28 120 122 123 125 126 128 129 131

3 30 115 117 118 119 121 122 124 125

18 20 crews/da

1 613 632

1 329 338

1 234 240

1 186 191

1 158 162

2 195 198

2 173 176

2 157 160

2 145 147

2 135 137

3 157 159

3 147 149

3 139 141

3 133 134

3 127 128 Table 5. Intermodal Costs per Railcar Equivalent 24

Std. 60 ft. Std. 60 ft. Std. 50 ft. Car Type Boxcar Boxcar Reefer Commodity general general perishable Trailer type dry van dry van reefer van Limited by cubes weight weight Term. type piggyback piggyback piggyback A. Lift or xload $30 $30 $48 B. Ratio trk:car 1.7 2.9 2.9 C. Term$ (AxB) $51 $87 $139 D. Trk$ (Table 2, 3 hr.) 169 327 365 E. Imd1$ (C+D) $220 $414 $504

Std. 52 ft. Bulkhead Cov. Hop. Car Type Gondola Flatcar or Tank Commodity steel lumber bulk Trailer type flatbed flatbed tank Limited by weight weight weight Term. type reload reload transload A. Lift or xload $100 $150 $100 B. Ratio trk:car 3.5 3.8 4.0 C. Term$ (AxB) $350 $570 $400 D. Trk$ (Table 2, 3 hr.) 381 413 527 E. Imd1$ (C+D) $731 $983 $927