The revenue in excess of the computed costs for the 14 railroads was 26.27 percent. For all other class I railroads, it was 10.56 percent, and for the United States as a whole, 23.33 percent.

Protestant did not consider that the cost computed as described above recognized sufficiently the characteristics of the coal traffic. To reflect these characteristics certain adjustments were made in switching-minutes per loaded car and in the number of train-miles required to handle the coal traffic. Protestant made a study at some 200 mine locations for the purpose of determining the total time after the arrival of the road train at the point opposite the mine track, in order to set out and spot empties or to pick up loads, until the train crew, after performing such work, was ready to leave. The number of loaded and empty cars set out or picked up at each mine location was also counted and reported as part of the study. An average of 2.7 minutes per loaded car was developed, but when the minutes per loaded car were weighted for the 1956 production in each of the various production groups, the average became 2.98 minutes per loaded Respondents pointed out that the minutes per loaded car excluded any consideration of the nonproductive time of the train crews doing the switching at the mines. Protestant then increased the weighted average of 2.98 minutes to 3.36 minutes per loaded car at the origin to include an allowance for nonproductive time.

car.

Protestant conducted a study at destination representing 58 consignees. The actual time per loaded car at the industry was 5.1 minutes. Respondents called attention to the fact that this study covered only the switching time on the industrial tracks of the consignees. By reference to material published by our cost finding section, additional time was added to include an estimate for classification and transfer switching, the resulting total per loaded car at destination being 20.44 minutes.

The time studies both at origin mines and at destination industries are subject to the usual tests for representativeness. While each study includes a large number of individual observations, there is no information of record to show that mine production or industrial receipts of coal represent average performance on an annual basis.

As to switching costs, respondents and protestant disagree as to the procedures set forth in Rail Form A. This formula assigns or apportions freight operating expenses, rents, and taxes to the various services, i. e., line-haul or running, switching (train and yard), station (clerical and platform), general overhead, and services unrelated to line-haul transportation. The total freight switching expenses resulting from these assignments and apportionments are related to the actual switching locomotive time in yards and to the train switching time performed by way freight trains as reported by the railroads to this Commission. There is also a relationship between the total switching time and the actual number of cars originated and terminated as reported to this Commission. While Rail Form A does provide for a distribution of the switching expenses on the basis of a car count which has been equated for the productive minutes of switching locomotive time per car, it is necessary to recognize that any reduction in time assigned to one type of switching service, for example, origination, will require an increase in the minutes per loaded car of another type of switching service, for example, intertrain or intratrain switching; otherwise, all of the switching time will not be included in the study. Protestant contends that its study at some 200 mines in 1958 is more representative of actual mine

switching than the studies used in developing the equating factors used in Rail Form A and, therefore, it reduces the switching time in origin service. Similarly, it makes a lesser reduction for destination service based on its study. At the same time it alleges that the intertrain and intratrain switching expenses developed in Rail Form A cover the remainder of the switching service which is given coal traffic between the origin and destination points.

Respondents claim this is a misconception of the formula, Rail Form A, which follows the principle that a certain amount of origin and destination switching service includes classification work in train, yard, transfer to an industrial area and spotting or pulling the loaded or empty car on the industrial track. Time studies covering only the switching service performed on mine tracks would not include an equivalent amount of switching work, being comparable in service to the pulling of the loaded or empty car on the industrial track.

The procedure followed in pricing out the coal traffic in Statement 5-57 is to reduce the switching locomotive-minutes at origin for all coal territorial movements except within the southern region, to the number of minutes per loaded car required in the Pocahontas region, the heaviest coal originating area. The minutes per loaded car in the southern region are approximately the same as in the Pocahontas region. This recognizes a favorable characteristic of the coal traffic. The reduction in minutes assigned to coal is then added to the switching costs of all other traffic.

The adjustment in line-haul costs was accomplished by estimating the trainmiles required to handle the coal traffic under certain assumptions. Protestant had mine personnel observe and report the number of loaded cars in the train at the time of departure from the mine. The results of this study showed an estimate of 44.3 loaded cars of coal per train. Protestant also computed the net tons per car for the United States by dividing the total tons of coal carried by the total cars of coal carried, which for the year 1956 was 59.39 tons per car. The tare weight per car was 23.3 tons which is the weight used by our cost finding section. To determine the gross-ton-miles per train-mile in way-freight train service at origin including the movement of empty cars, an average weight of car and contents of 53 tons multiplied by the loaded cars per train was used, producing 2,348 tons per way train-mile. This figure compares with 1,153 tons for all traffic handled in way-train service and was computed by weighting the cost finding section's tons per way train-mile by the total gross-ton-miles in the four territories. A figure of 1,123 tons per way train-mile was also computed by protestant as being the weight of train which the cost finding section would use in costing coal traffic.

Protestant then developed an average haul per ton of 322.6 miles for the coal traffic in the United States. This figure was computed from data in the waybill studies published by this Commission. This average haul was separated between the miles the coal was handled in way trains and the miles in through trains. The average haul in way trains computed by protestants as being the cost finding section's basis, was 57.5 miles. This figure is in error, as the weighted average of way train-miles as computed by the cost finding section for the four territories is 47.0 actual miles. The difference between 322.6 miles and 57.5 miles is 265.1 miles or the distance that coal moved in through trains. Comparable data computed under protestant's assumptions were 29.75 miles for way-train movements and 292.85 miles for through trains.

304 I. C. C.

The average gross weight of all through freights in the United States of 3,723 tons was developed by protestants from the cost finding section's application of Rail Form A. This figure was adjusted to 4,383 tons to reflect the weight of coal trains by weighting the unadjusted through train weights in each territorial movement, as shown in waybill studies, by the net-ton-miles of coal from the waybill statistics. The weight of the average train, i. e., way and through for handling traffic was computed by adding the gross-ton-miles handled in way trains and the gross-ton-miles handled in through trains and dividing by the average haul per ton of coal. This average weight of all coal trains was divided into the coal gross-ton-miles computed for the year to develop the estimated number of freight-train-miles required to handle the coal traffic as a whole. The estimated coal gross-ton-miles were computed by dividing the coal gross revenue by the average revenue per ton-mile shown in the Commission's waybill studies to obtain an estimate of the net-ton-miles of coal traffic which were increased for tare weight of cars and empty return movement of cars. On the basis of coal being handled in trains of the same weight as all traffic, the coal train-miles were computed to be 71,665,120. With allowances for the adjustment in train weights on the so-called cost finding section's basis, the trainmiles were estimated to be 61,543,034. On the basis of protestant's assumptions, the figure was shown as 55,887,619. However, computations of the difference in costs were based on a figure of 55,777,501 train-miles. The excess train-miles over protestant's estimated figure were 15,887,619 on the first basis mentioned above and 5,765,534 on the second basis. The difference between the trainmiles of coal traffic handled in trains of the average weight for all traffic and in trains of the adjusted weight represented to be the cost finding section's method is 10,122,086. The differences in train-miles were multiplied by a cost per train-mile of $2.433.142, computed by protestant from Rail Form A, to obtain the amounts of adjustment to be made in car-mile expenses to reflect the lower cost of handling coal traffic as compared with the average of all traffic. The reduction credited by protestant to the cost finding section's basis was $24,628,473 and to protestant's basis, $38,365,833. The adjustment in costs related to differences in train-miles rests on assumptions developed with data covering portions of the coal movement with favorable characteristics and the resulting factors applied to the entire movement.

Another adjustment made by protestant related to station clerical and loss and damage clerical expense applicable to coal traffic. Protestant used the weighted average clerical expense of The Chesapeake and Ohio Railway Company, The Norfolk and Western Railway Company, and The Virginian Railway Company, which amounted to $2.65 per shipment as being applicable to all coal traffic in the United States. This assumption produced a cost of $36,647,000 which compared with a cost of $76,044,000 computed by the protestant as being the basis of the cost finding section. The station clerical expense of three originating railroads cannot be said to represent necessarily the entire station clerical cost of coal including the termination as well as the origination of the traffic.

A comparison of the costs submitted by the protestant is shown in table 1, as follows:

304 I. C. C.

It will be noted from table 1 that protestant apparently did not include other station expenses of $16,063,000 in its original application of unit costs developed for the 15 railroads. If this amount is added to the total costs as shown, the new total becomes $1,022,496,000 which is $14,674,000 less than protestant's figure on the basis of the cost finding section. This difference amounts to 1.4 percent. The adjustments in switching expense, station clerical expense, and trainmile expenses, having been made on the basis of factors developed for the more favorable characteristics of the coal traffic and then applied to the entire coal movement, produce costs at a level which is below that for the coal movement as a whole. Since the difference in cost under the protestant's method of using unit costs for individual railroads and under protestant's basis designated as that of the cost finding section is relatively insignificant, it is concluded that the protestant's latter basis produces reasonable costs of handling the coal traffic on a total United States basis.

Protestant also developed the out-of-pocket portion of the costs on the basis of 80 percent of the freight operating expenses, rents and taxes other than Federal income taxes as shown above in table 1, columns 6 and 7. The out-of-pocket portion of an allowance for return was obtained from the Rail Form A applications of the cost finding section and assigned to coal traffic on the basis of service units. It represents a 4 percent return after payment of Federal income taxes. The ratio of revenue to out-of-pocket cost was 132 percent on the basis which protestant described as that of the cost finding section. The latter figure compares with a ratio of 126 percent for the year 1955 as shown in Statement 5-57.

Protestant discusses at considerable length various bases for distributing the transportation burden. Protestant states that, if it is assumed that the transportation burden is in the nature of a joint or indirect cost which must be borne by freight traffic, then it is reasonable that it should be distributed in accordance with the transportation benefits which the several freight traffics derive from railroad service. To the protestant there are three possible measures of these benefits of the service to a given traffic: (1) out-of-pocket cost of providing that service; (2) value of the goods transported; and (3) traffic volume transported.

The concept of out-of-pocket and constant costs as used by our cost finding section does not consider the transportation burden as a joint or indirect cost. Joint cost has been described in a narrow sense as a cost which results from a single operation. Classic examples are the production of beef and hides or cotton and cottonseed. In transportation the line-haul costs associated with the movement of equipment are a joint cost because the production of transportation in one direction makes possible a loaded or empty movement in the opposite direction. Indirect costs are in the nature of a general overhead expense, and in that sense they may be considered for pricing purposes as a constant cost. It may also help to understand the principle involved to think of the constant costs as that portion of a railroad's total costs which would be required to establish the plant and equipment in readiness for business. In other words, some portion of each expense item would have its constant portion. Protestant argues that the indirect costs should be borne in proportion to the direct costs. This position assumes that railroad costs are 100 percent variable with changes in volume of traffic and that a 10-percent change in ton-miles or cars switched will be followed by a 10-percent change in the corresponding operating expenses. No data of record show that this condition exists. In fact, protestant has accepted as fact that for a 10-percent change in