form of usage should not be attributed to railroad responsibility. The 1964 AAR study of tank cars disclosed a substantial number of "no-mile" and "low-mile" cars. Although the responsibility may lie with the railroads for putting some portion of those cars into that category, through derailments and other accidents, for example, the probabilities are that many of the cars were idled by owners or shippers in a way, or for a purpose, inconsistent with freight transportation by railroad.

MCA, in particular, concedes that the use of an appropriate ratio to discount nontransportation use of tank cars would be proper, but it criticizes respondents' method as based on an incorrect assumption. Respondents' method, in effect, discounts the total annual tank car costs by a single percentage factor to reflect purported transportation utilization, e.g., the inflation of actual miles to a single factor mileage divisor.

At the hearing MCA proposed an alternative structure of specific allowances based on a complex formula separating "time-related" and "mileage-related" tank car expenses, and allocating them in accordance with a "transportation use factor" especially derived by MCA for that purpose from a number of related statistical analyses. Respondents criticize the MCA formula for being founded upon a greater number of assumptions and arbitrary judgments than the AAR formula, and thus a less reliable costing tool. API joins the respondents' criticisms in large part, and, further, argues that if the Commission should conclude to make some provision for nontransportation use nothing greater than a 5 percent discount of annual tank car costs is warranted.

Several of the arguments of respondents and the API against the MCA formula are well taken. The basic formula put forward by respondents in the case is essentially in accordance with the Commission's previously stated guidelines, and the areas of disagreement among the affected interests in the application of that formula have been very substantially narrowed. The MCA proposal depends upon a number of new "judgment" statistical samples -such as 1-day studies of chemical tank car movements -to develop operative factors in its allowance formula for the entire tank car fleet. The examiners do not find the MCA formula to be inherently unreliable, particularly as the interim device for which MCA proposes it, but we do find the respondents' basic formula more suitable and reliable for present purposes, with the adjustments thereto discussed below.

On brief MCA presents a new analysis of respondents' cost-allowance formula, and adjusts it in accordance with time-related and mileage-related cost factors. MCA's first step in this analysis is to translate respondents' "inflation" of 1964 actual miles for its mileage divisor into a percentage relationship. At the hearing respondents' cost expert conceded that, for all practical purposes, respondents' process of inflating actual mileage

was the equivalent of applying a transportation use ratio. In other words, the inflation of the 6,068 actual miles recorded for group 4 cars to a 10,000 mile divisor represented a judgment that group 4 cars were utilized only 60.7 percent in transportation functions compensable by allowances. The conversion of respondents' selected mileage divisors into transportation use ratios is shown in the following table:

Actual 1964 mileage 4,876 4,764 5,389 6,068 9,668 10,556

Respondents' proposed 6,000 8,000 9,000 10,000 11,000 12,000 divisor (miles)

Comparable transpor- 81.3% 59.3% 59.9% 60.7% 87.9% 88.0% tation use ratio

All

MCA

Next MCA separated the AAR 1964 study costs into time and mileage allocations, in accordance with the theory that (a) tank car depreciation costs should be apportioned 60 percent on a time basis, and 40 percent on a mileage basis; and (b) operating costs should be apportioned with 50 percent on a time basis and 50 percent mileage. interest costs would be regarded as related to time. maintains that all costs related to mileage should be regarded as transportation use costs and should be included in the cost base for allowances. It concedes on the other hand that time related costs should be discounted by transportation use ratios, such as those computed above. The following table reflects the MCA adjustments of the respondents' allowance formula, including the resulting allowance levels derived from the AAR formula so adjusted.

Transportation use ratio 81.3% 59.3% 59.9% 60.7% 87.9% 88.0%

The examiners conclude that the separation of tank car costs into time-related and mileage-related categories provides a more realistic and reasonable basis for accurate analysis than lumping all tank car costs into one category; and that the evidence in the present record is sufficiently complete to employ such an analysis in the manner MCA proposes. Since the record warrants the use of such an analysis, the respondents' single-factor, inflated mileage divisor is too blunt an instrument to differentiate between costs which should be reimbursable as "transportation related" and those which are not. Although a permanent allowance structure will not be attainable until an adequate utilization study provides a basis for resolution of the controversy on relative responsibility for nontransportation use of tank cars, any temporary allowance structure should be based on the most complete data and most realistic analytical procedures available.

As described previously, uses such as static storage, excessive loading-unloading car detention, intraplant movements, and car holdings or "build-up" for peak season movements are most plausibly the functions which may properly be classified as nontransportation uses. Only a comparatively small proportion of car mileage would seem to be generated by such nontransportation use. Indeed, it is respondents' contention that "no-mile cars" and "lowmile cars" are a primary indication of the existence of nontransportation use. Protestants' argument that all costs attributable to mileage should be reimbursable costs in the allowance structure has substantial merit. On the other hand, respondents maintain, convincingly, that some nontransportation uses of freight cars, such as intraplant switching service, 21 do generate costs normally related to mileage, although there is no data available to measure those costs.

211t is respondents' experience also that large manufacturers at times will attempt to move freight cars within their plants by use of tractors or cranes, and that this kind of use generates substantial car damage normally associated with mileage.

The examiners conclude that the annual tank car costs attributable to interest and to the time-related elements of depreciation and operating costs should be allocated to railroad responsibility (and thus included in the reimbursable cost base) only in the proportions implicit in the transportation use ratio given in respondents' proposed formula; but the annual costs attributable to the mileage-related elements of depreciation and operating costs should be discounted from the reimbursable cost base by a substantially smaller factor. In a somewhat similar situation relating to per diem payments in Chicago, B. & Q. R. Co., v. New York, S. & western R. Co., 332 I.C.C. 176, the Commission recently concluded that the costs of repair and maintenance of per diem freight cars should be apportioned 40 percent to mileage and 60 percent to time. These ratios are generally consistent with the engineering studies underlying related factors used in the Commissior's Rail Form A cost finding formula, although Form A would suggest assigning 70 percent of repair costs on a mileage basis. While MCA would include all mileage-related costs within the tank car cost base for reimbursement, the examiners conclude that pending completion of the AAR utilization study, these costs should be discounted by a judgment factor of 5 percent. The probabilities are that shipper and lessor responsibility for such costs will lie within a close range of 5 percent, and that figure, the examiners conclude, will be a just and reasonable element in the interim allowances prescribed herein.

The following table shows the effect of the 5 percent adjustment the examiners will require for the MCA formulation:

Respondents, as noted earlier, adjusted the computed costs per loaded mile by an additional factor in selecting the allowance levels they ultimately proposed. In applying this factor, 91.91 percent, respondents sought to maintain their total payout under the proposed allowances at the same level (assuming the same level of traffic) as would be payable under the existing 3-level structure of allowances In other words, a redistribution of allowances would be made, rather than a change in the level of allowances. Respondents' explanation for their use of this factor has been uncharacteristically vague. Supportive arguments vary from unelaborated "economic factors" to discussions of the problem of nontransportation use of tank cars, but there is no clear rational basis for the use of the 91.91 percent adjustment. The examiners conclude that use of such an adjustment would be purely arbitrary, and that it has not been shown to a just and reasonable adjustment.

Protestants make the additional argument that respondents unjustly have failed to provide an increase in their proposed allowances to reflect the inflationery pressures and increased costs occurring since the survey of 1964 tank car costs. They ask the Commission to prescribe an adjustment in the allowances to account for these increases. As noted earlier by the examiners in connection with proposed valuation ceilings, the kind of evidence needed to support the prescription of a realistic inflation factor was not developed on this record, and the examiners do not believe such a factor should be prescribed in this case.

Although respondents computed average costs per tank car mile on the basis of actual miles, the payments proposed would be made to car owners on the basis of short tariff miles. While the "actual miles" represents the number of miles a car moves over the actual route of movement, "short tariff miles" represents the shortest mileage between origin, destination, and interchange points over the lines of each individual railroad as published in mileage tariffs filed with this Commission. Respondents' proposal would continue the practice presently followed by all major railroads, in paying the trilevel allowances on tank cars on loaded short tariff miles only. Prior to the 1964-1965 period of transition, tank car allowances were applied on both loaded and empty miles over the actual routes of movement. Protestants contend in this proceeding that the new allowances should be paid on actual miles, or, alternatively, if respondents proposal to pay on short tariff distances is approved, that the allowances be increased by an appropriate factor reflecting the average circuity of actual operations, as compared with short tariff mileage tables. In the prior Tank Car Allowances case the Commission concluded that actual mileages run would constitute the best standard for determining actual costs, and therefore also the allowances based on actual costs. The Commission went on to say, at 329 I.C.C. 483,