of a district, region or state band together in a sample study, the ground covered and purposes served may be quite different from that of the waybill material. The Summary of Utilization attached indicates that rail-carrier organizations in various parts of the United States have from time to time utilized the waybill data. The replies indicate that motor and water carriers also make use of the material.

The need for traffic information, whether as to volume, traffic distribution, revenues or other factors is not limited to formal proceedings. Most rate issues require planning periods during which traffic research is conducted as a basis for policy decisions and action. Many rate issues between shippers and carriers are resolved without recourse to the Commission. The existence of continuous, current traffic and revenue statistics bearing on cases at hand clearly serves to facilitate rate agreements or other resolution of the subject matter.

State regulatory commissions are commonly lacking in data with which to portray the pattern of either their own intrastate traffic consist or the competitive interstate rate patterns. Comparisons between states or between intrastate and interstate traffic are not commonly available except as may be occasionally assembled by the carriers for specific purposes.

Over and above the foregoing, waybill data fill many gaps in the material available for economic research over a wide range including equipment, circuity of haul, cross-hauling and the general distribution of goods. There is evidence that such data may be particularly helpful in connection with national defense.

The response to the letter of inquiry indicates that in the opinion of the preponderant portion of those answering Question 2 there exists a public interest in the waybill statistics. In the mind of the committee a proper concept of the public interest in transportation implicitly requires a regulatory agency to be possessed of all reliable facts pertinent to the discharge of its duties. The waybill study contributes to this end.

Based on the Summary of Utilization, the responses to the letter of inquiry and the purposes and uses of the waybill studies, the committee is of the opinion that there is a basic need for information of the character contemplated by the waybill statistics.

(2) Are The Needs Referred To Above Being Adequately Met By The Current Waybill Program And Is The Cost Justified?

The committee believes that the basic needs, as set out above, for more informative traffic and revenue data can be furnished by a waybill sampling program. Furthermore, it believes that the present program meets in a practical way such need, in spite of certain inadequacies referred to herein.

The facts available on the present utilization of the waybill releases indicate a workable degree of confidence in the material for a wide range of purposes. This applies to a significant proportion of the users in nearly all of the groups contacted in the committee's inquiry. It is

also brought out by the list of utilization made over the last several years. The absence of the waybill data would leave a tremendous void in the present-day knowledge of territorial and state-to-state traffic flow, lengths of haul, comparative rate levels by commodities and territories, rate indexes, working margins between prevailing rate levels and direct costs, the distribution of the transportation burden among commodities and other factors. No suggestions have been received as to how this knowledge, much of which is now taken for granted, could be obtained on a practical basis in the absence of the waybill studies.

The committee believes, however, that the current use of the waybill statistics falls substantially short of their full potential. While the data are by no means a cure-all in the matter of traffic and rate information, they do provide useful and unique traffic information over a very wide sector of the transportation economy.

The failure to make full use of the material appears to be caused by a number of factors. As pointed out in several responses to the inquiry, there appears to have been something less than an effective program of demonstrating the purpose, benefits and utility of the material, although special releases and lectures on the uses of the waybill data are now accomplishing much in this direction. Commission utilization today comes largely through the statistics introduced as evidence by parties appearing before it. In addition, a substantial volume of cost-revenue comparisons are provided by the Cost Section to the Board of Suspension and others on the Commission staff. Such material is largely dependent upon data made available from the waybill studies.

There is no doubt that the unrealized potentialities of the study spring in substantial part from concern over the adequacy of the one per cent sample. The committee believes that there has been insufficient exploration of the adequacy of the waybill material in all its sectors and groupings. Where many subdivisions of the material are made, it is inevitable that the sample will become thin in spots. Even a 10 per cent sample might well prove inadequate on some of the breakdowns.

The degree of confidence necessary depends on the use to be made of the material. Substantial accuracy may be demanded in some instances whereas broad approximations, to be used in conjunction with other indicators, may suffice in other instances. The degree of error permissible will also depend on whether one is interested in traffic flow, rate, equipment or other data. Also, thin samples in any one year may now be combined with the results of a continuous period running back some six years. A combination of the results for several years would be acceptable for some purposes.

Staff defense of the reliability of the waybill studies in the past has largely rested upon reference to sampling theory and doctrine. However, tests are currently made with the Commission's commodity statistics regularly furnished by the railroads and these support the over-all reliability of the waybill studies.

Also, the record of commodity originations and terminations by states (Q 550) maintained through 1950 provided reference material with which numerous tests have been made generally confirming the reliability of the waybill statistics. No doubt the carriers, for their own reasons, have extensive traffic flow data for some of the more important commodities with which comparisons can be made.

road has rather vigorously asserted that its own checks show that the waybill data are not reliable but what comparisons were made and how such conclusions were reached are not shown, although a request for such information has been made by this committee.

Rate and revenue patterns may be test-checked against tariffs where there is familiarity with rates which move the traffic. No doubt traffic evidence based on extended test periods may be available. Comparisons of waybill data between years can be revealing as to consistency.

It is to be expected that such tests will develop wide sectors of reliability in the waybill data. It is also to be expected that extensive areas will be revealed where the sample will appear erratic or otherwise of doubtful validity. Such may be the case where the observations are scattered, where transit services apply, or where commodity classes consist of groupings of varying items brought together for reporting purposes.

In the interest of building up familiarity and confidence in the waybill statistics, it is just as essential that these weak areas be charted, explained and tested against available data as to identify the strong

areas.

It is true that the waybill releases carry their own qualifications, including a mathematical formula to determine possible sampling error, and warning in the use of material where transit is involved, but judging from many responses received, this subject needs much more extensive and special treatment.

As an additional step the services of one or more recognized sampling experts might be called upon either separately or as an integral part of the foregoing appraisal. The objective would be to furnish an independent review of the results of the sampling procedures in the light of practical and convincing controls and tests. It might here be observed that while it is essential to have discussions relating the waybill sampling procedures to the laws of probability the more pressing need now seems to be for demonstrations of the general representativeness of the material based on comparisons with known controls. The latter would be the more convincing in resolving the areas of doubt that have been raised.

With respect to the cost to the railroads of furnishing the waybill data to the Commission, the railroad associations make no estimate as to the amount, although contending that the expense is not justified. The manner in which the information is obtained on a continuous sample basis makes the ascertainment of such cost difficult, since much of it is a part of routine operations.

The Commission has estimated the cost to it of the waybill studies to be approximately $190,000 per year. There is some evidence that

by increased mechanical aids this cost can be somewhat reduced. The possibility of reducing costs by calling for the waybill data in alternate years does not appear practicable, because of the breaks in routine which would follow with resulting losses in efficiency and dependability in the furnishing of such data. Taking into account the cost to the Commission of the waybill studies and the demonstrated need thereof, the committee is of the opinion that continuation of such studies by the Commission is justified.

(3) What Changes Might Advantageously Be Made In The Waybill Program?

A wide range of suggestions and recommendations were offered in response to the committee's inquiry as to how the present waybill procedures and programs might be made more effective. These are set out at some length in the Digest attached. The major items are briefly summarized in the attached numerical analysis of the replies to the questionnaire, pages 2-4. Review and comment on the principal items follow below.

Inadequacy of Sample-Suggestions were made that related commodity classes be combined where the sample was "thin," and that "controlled" or "stratified" sampling be adopted which would adjust the size of the sample to the volume of the movement or other traffic characteristics. A few references were made to the possible use of sampling on a cluster or selected-day basis.

How thin a sample can be and still not yield useful information is a question difficult to resolve. A much wider degree of approximation may be acceptable in one situation than in another. Also, as previously indicated, thin samples may sometimes be combined for a period of years to improve coverage, depending on the information sought. It is difficult to draw a line where all usefulness ceases. It is suggested that serious staff consideration be given to grouping together related commodity classes, the size of which commonly prevents the data being given any significant weight. Special runs of the smaller items could be made on occasion if conditions warrant.

As to "controlled" or so-called "stratified" samples by commodity classes or other breakdowns, these involve complications and expense which the committee feels would make their use prohibitive. The committee also is of the opinion that for any given number of waybills included in a sample the continuous basis is preferable to the use of selected days.

It is recognized that an enlargement of the sample would no doubt help solve some of the criticisms. However, even though the sample were raised to 2 per cent, a large number of the finer breakdowns on the smaller commodity classes would still be "thin." Also, such enlargement would increase the costs, something which is to be avoided. Enlargement of the sample is not recommended.

Release Waybill Data More Quickly-Several suggestions were made that the time lag in the publication of the waybill statistics should

be reduced. Needless to say, every effort should be made to keep such information current and both the railroads and the Commission appear to be cognizant of this problem and to be working toward improved results.

The Waybill Statistics Need Better Explanation-Many appear to feel that the waybill statistics have been inadequately presented and analyzed; that better promotion in the way of explanations of purpose, benefits and use is desirable.

The committee believes that a great deal has been accomplished by the Commission staff in developing and disseminating the waybill statistics. On the other hand releases outlining the purposes, character of use and general adequacy of the material have lagged. A general stock-taking with emphasis on the purposes served and benefits to be derived from the program, is desirable as many writers suggest. It is quite evident that even among many receiving the releases widespread ignorance exists as to their utility and purpose.

Provision Of Additional Breakdowns-The suggestions here made go to a finer subdivision of the sampling data as to type of rate, commodity classification and metropolitan area; also to export and import, water transshipment and local vs. interline traffic. The committee believes that this subject might be explored and breakdowns furnished as special runs where the information is at hand. It is skeptical, however, as to any general expansion of the breakdowns now made in the studies.

Reorganization and Renumbering of the Waybill Statements-It is stated that some statements overlap and might be discontinued; that some might profitably be broken up into smaller releases or consolidated with others; also that user mailing lists should be checked for continued interest in given releases. These matters require a detailed familiarity with each waybill release, the demand for it, and the extent to which a reorganization of the statements could be effected in the interest of simplification, economy, or utilization. The committee believes that such a task can best be undertaken by the Commission's staff.

References are made to the effect that the titling and numerical identification of the waybill releases are of very limited assistance to users in filing the material. Suggestions are made that improved numbering systems be adopted. The committee believes that such suggestions warrant serious consideration by the staff.

Expansion of the Study to Other Agencies-Twenty-seven replies to the committee's inquiry suggested the expansion of the waybill statistics to other agencies, primarily motor carriers. In most instances the inclusion of truckload or volume shipments (starting at either 5,000 or 10,000 pounds) was considered although a few would expand the coverage to less truckload traffic. In most cases no distinction was made between common and contract carriers.

Some railroad representatives argue that it is unfair to impose upon the railroads the burden of supplying the waybill data and exposing them to the adverse use thereof when the other transportation agencies are not required to furnish waybill data.