enunciated by the Commission. Among other questions in need of clarification and definition of policy are: the extent to which cost and income experience of more efficient firms, as distinguished from "representative" firms, should be the basis of fixing "reasonable" rates; the treatment of commission splitting or "give-ups," when not compensating for services in connection with brokerage transactions on the exchange; the inclusion or exclusion of interest income on customers' debit balances; the concept of "fair return" and, in relation thereto, the appropriate treatment of invested capital and of partners' or stockholders' compensation.

6. The Income and Expense Report form of the NYSE in its present form or as revised from time to time should be required to be filed by all member firms doing a public commission business and, with appropriate modification, by other categories of member firms deriving income from member or nonmember commissions. Similar reporting should be instituted by the other principal exchanges. The Commission should regularly receive copies of such reports (with or without identification of firms) and through this means and otherwise should keep advised of the economics of the securities business insofar as relevant to the fixing of reasonable rates of commission. Since volume of trading has important bearing on profitability and therefore on reasonableness of commission rates, the Commission in conjunction with the exchanges should seek to develop improved standards and procedures to take account of significant changes in volume from time to time. Consideration should also be given to the feasibility of establishing unit costs for various components of the brokerage function and ancillary services, as a further guide in applying the "reasonable" standard on a continuing basis.

7. To place rate review by the Commission on a more orderly and efficient basis, existing procedures should be modified to assure that proposed changes in rates will be submitted to the Commission adequately in advance of their proposed effectiveness; and consideration should be given to the feasibility of providing that, where an increase becomes effective pending the Commission's review, refund or adjustment will be made in respect of any part of the increase not ultimately approved.

8. While the odd-lot differential is theoretically an adjustment of the price charged the customer's broker rather than a commission, in practical effect to the customer it is another element of the cost of effecting a stock exchange transaction. Customers' confirmations of odd-lot transactions should be required to show separately the odd-lot differential and the brokerage commission.

J. AUTOMATION-ITS NEEDS AND POSSIBILITIES

In various places in this chapter and elsewhere in the report, reference has been made to the needs and possibilities for the utilization of modern data processing techniques in various exchange procedures and mechanisms. In this part some additional factual matter is presented and some general observations are made with respect to the present usage of these techniques, and the need for integrated progress in the area.

In viewing accomplishments of the exchange markets in the field of automation, the overall impression is one of unevenness; on the one hand some highly advanced electronic equipment has been installed, while on the other, procedures are employed which have not progressed significantly since the telegraph and telephone came to the floor of the NYSE around the beginning of the century. Thus, the NYSE clearing corporation has introduced important automated techniques. Also, some member firms have taken advantage of computer technology to implement their programs for internal surveillance and supervision, and to integrate transmission of orders to the floor and reports of execution with the back office work necessary to close transactions. Finally, the Exchange has announced plans to automate certain aspects of its quotation and sales information system. Yet, as great as the progress has been, many opportunities for further progress remain, including areas still relatively untouched by technological progress.

1. THE INTERESTS IN AUTOMATION

The actual and potential uses of electronic data processing systems in connection with the exchange markets are of importance in many ways. Member firms and their customers are affected in that automated procedures may have a large impact on the efficiency and costs of transacting business, the quality of supervision, and the extent and reliability of available information. A clear example of the first point is discussed in part E of this chapter in connection with the odd-lot systems; the second is shown by examples in chapter III.B in which broker-dealers' automated surveillance of branch office activity is discussed; and the third is illustrated by the fact that during the hectic markets at the end of May 1962 it was impossible for many investors to obtain fresh and accurate information about either the market or the status of their own orders (see ch. XIII).

The exchanges themselves have an interest in this area not only to aid in their ordinary operations and to facilitate the convenience of their members and the public, but also to discharge effectively their responsibilities for surveillance and enforcement. This point is very clearly illustrated by the New York Stock Exchange's "stock watch" program. As described in chapter XII and elsewhere, this program depends in part on the computers utilized in the stock clearing operation. In a recent case the Exchange was able, through the use of its sophisticated equipment, swiftly to identify and remedy a situation involving an improper purchase of securities by an allied member of the Exchange.

Modern data processing techniques are of particular importance to the Commission in the conduct of its surveillance and enforcement programs and in the discharge of its other responsibilities under the Exchange Act. Only through quick and efficient mechanisms can the necessary amount of disaggregated market information be accumulated and evaluated. The discussion in chapter XIII of the 1962 market decline illustrates the difficulties faced by the Commission under present procedures for the accumulation of needed information in times of severe market stress. Moreover, the slow but highly important currents of change in market patterns that are described in chapter VIII call for more complete and continuous accumulation of data that can be efficiently processed only with computer equipment.

2. THE TRADING PROCESS

One of the most important sectors of the securities markets upon which the impact of modern technology should be assessed is that of trading on the floor of the stock exchanges, in particular the NYSE.622 The Exchange floor is also the area where the problems involved are the most complex, not so much in the technical sense as from the point of view of the ramifications involved in altering the structure of a marketplace whose members and employees' livelihoods are intimately connected with present procedures.

In spite of its importance, the floor of the NYSE has been untouched by most of the technological developments of the 20th century. A critic of the NYSE's progress in technological innovation has said that the basic organization of the Exchange's floor has not changed since the "period in which the institution solidified-slightly before the telephone." 623 While the Special Study should not be understood as espousing the proposals made by this commentator, there is undoubtedly some merit in his analysis. Aside from recent developments in methods of transmitting orders to the floor, noted above, and various innovations and proposed innovations with respect to the reporting of transactions discussed in part E above, there has been no basic change in the methods of executing orders since the NYSE floor took its present form. Except for firms utilizing teletype devices, orders reach the Exchange by telephone and are written down on slips by clerks. From that point, orders are transmitted manually by brokers, or through tubes, to the trading post. Orders given to specialists are again transcribed by hand onto the specialists' books. 624 At present there is no internal means of assuring that quotations announced on the floor of the Exchange are the same as those disseminated to the public. Even after the Exchange automates its off-floor quotation service such assurance will not be provided.625

The mere fact that the Exchange floor has not changed substantially over the years is not of itself proof that the Exchange has been laggard in utilizing the resources of modern technology in the trading process. It is obvious that very fundamental issues, going beyond mere technology, are raised by the suggestion that orders be executed by intricately programed computers, 626 thereby removing all brokerage skill and judgment from the marketplace and possibly even eliminating the role of the specialist. But many intermediate possibilities between the present system and total automation of market facilities can be envisaged. Although in theory all orders on the floor receive personal attention, the floor broker frequently has no choice but to execute a customer's order in a mechanical fashion against the existing quote. The question arises whether it is really necessary that such orders be personally executed by the floor broker at the post rather than from

622 See pt. B of this chapter for a description of the present trading mechanics on the floor of the NYSE. 623 William S. Morris, as quoted in "Electronic Finance?" in Forbes, May 1. 1963, p. 37. 624 See the recommendation contained in ch. VI.D that techniques be devised to preserve the specialist's book, which is at present an ephemeral record.

625 Since there seemingly is no technical problem in connecting post indicators to the automated system, the question arises as to whether there is any justification for permitting the possibility to exist of different quotations being given on the floor of the Exchange and to investors. Thus, the Exchange's proposal Illustrates the Commission's interest in matters such as automation from a regulatory standpoint apart from achieving efficient market mechanisms.

628 See "Electronic Finance?" Forbes, May 1, 1963, p. 36.

the office of a member firm enjoying instantaneous connection with the post by television or other means. Any such system would, of course, have to be limited to situations for which it was clearly appropriate. For other situations, a firm might choose to utilize the services of what would be a limited number of floor members.

This discussion is not a recommendation that any particular kind or degree of change be adopted or even seriously considered for adoption at this time-it may well be that other possibilities yet undeveloped would be superior. Yet modern technological capabilities are such that these matters are no longer in the realm of science fiction. What is pointed out here is the necessity for continuous awareness and exploration of the possibilities of automation-not merely on the part of the NYSE and the other exchanges 627 but also by the Commission itself. The entire question is of particular importance and immediacy in the case of the NYSE in view of its intention to move its quarters to a new building which will include a new trading floor and ancillary facilities.

3. ACCUMULATION OF MARKET DATA

a. Periodic reports

More obvious and immediate possibilities for the introduction of data processing techniques lie in the realm of accumulation of market data, where both needs and potentials seem substantial. One of the sources of market data of different kinds is a variety of reports submitted by members and member firms to the exchanges and by the exchanges to the Commission. At present there are about 20 such reports being filed on a daily, weekly, or other periodic basis with the NYSE.628 They cover a host of matters and contain valuable data which could be utilized to construct a comprehensive picture of market developments. However, each of these reports was developed for a specific end use and is usually analyzed only for that single purpose; e.g., floor traders' reports are used for surveillance of floor trading activity, while reports on underwriting commitments are used primarily as a check on member capital positions. Because the design of these reports and the reporting cycles differ, it is difficult, and in some cases impossible, to integrate the information contain in them. For example, while one report shows short positions in individual stocks as of the 15th of each month, another provides information on daily aggregate short sales of all stocks combined.629 Furthermore, the processing of many of these reports is done laboriously by hand.630 There may be a considerable timelag between the filing and availability of the data even for the particular end use for which the report is designed. Thus, the monthly report of customers' debit and credit balances, as well as the weekly report on members' principal transactions, are available only after a timelag of 2 weeks.

It seems clear that the whole area of periodic reporting by members should be reviewed, first, for the adequacy of each report, and second, with a view of meshing the data contained on each report with other data so as to give a continuing, comprehensive market pic

627 As mentioned below, the NYSE established a department of operational planning and development in 1960, and the Amex is currently studying various aspects of automation. 628 A tabulation of the forms received by the NYSE from members is provided in table VI-91.

029 See pt. H of this chapter.

630 See ch. XII for a discussion of the processing of floor trading data.

ture. Although such a goal could not have been achieved a few years ago because of the sheer mass of information, such data can now be processed expeditiously through the use of modern techniques. This is an area which is particularly appropriate for a cooperative effort by the various exchanges and the Commission to simplify the reporting burden on members while increasing the usefulness of the whole product to all concerned.

b. Daily market information

The foundation of any system for the accumulation of market data for regulatory or informational purposes is an accurate and reliable report on each individual transaction. This might seem to be the simplest information to accumulate, but in fact under present procedures it is extremely difficult to gather. This section does not purport to present a comprehensive assessment of either the problems or the remedies but merely points to certain areas of difficulty and suggests general approaches for further study and action.

Any attempt to study what is happening in the market as a whole or in specific stocks must start with prices, volumes, and an identification, at least by class of investor, of who is buying and selling particular stocks. In the course of the Special Study's investigation, a considerable amount of such data was collected. It has been the study's uniform experience that with present techniques, and even with the fullest cooperation of the exchanges and broker-dealer firms, accurate information is extremely difficult to obtain.

There are two major difficulties in securing accurate market information on a daily basis, aside from the mechanical problems presented by handling masses of data. First, there is no reliable and expeditious means of tracing an order back from the point of its execution to its source, nor of reconstructing the sequence of transactions for any period of time. Any market study or surveillance must consequently rely on varied secondary sources. This difficulty is connected with the second problem-that the data so recovered are not complete, a fact that may not become apparent until the information is used for purposes other than those originally intended.

At present there are only two major sources of market transaction data for individual stocks: the exchange tape and clearinghouse reports. The tape is prepared from the reporter's notation at the post of each transaction he observes or about which he is informed. Clearinghouse data originate from member reports of executions only on transactions which are cleared, and there is no requirement that all transactions go through the clearinghouse.

Tape data, generally known as "reported volume," are not comprehensive for various reasons, such as human error and limitations in reporting requirements.631 Whatever the reasons for such omissions, it seems apparent that reported volume is not a wholly reliable source of market data.

For surveillance or market study purposes there are other problems in the use of reported volume. The reporter's notation does not re

6 Members of the NYSE are not required to report certain kinds of transactions in "stopped" stock. According to the NYSE, such transactions amount to about 5 percent of volume and constitute the major source of omissions. However, on a sample day examined by the Special Study, only 14 percent of the omissions from the tape were accounted for by stopped transactions. See pt. D of this chapter and tables VI-47 and VI-48.