be coordination of several locally owned and managed regional depositories. Another could be concern about the safety of certificates immobilized in vast quantities. A third could be the accuracy of records of ownership and changes therein, and prompt and accurate reports thereof.

The comfort that I would give to those with these concerns is that by far the most effective of all potential sources of discipline exerted on the CSDS will be that of its depositors. These include sophisticated financial people. Their own self-interest and that of their clients rides with the depository. If a CSDS makes an error, has a delay, or incurs an undue risk, its people will hear about it, not from a quarterly or annual inspection by an authority, but from a depositor in a day-if not an hour.

In general, enlightened self-interest will take care of coordination among regional depositories. It does in Germany. If New York wishes to get away from the physical delivery of securities to California or Chicago, and substitute a book-entry transfer, it cannot do so unless California and Chicago are equipped to handle the other side of the book-entry. The same goes for other financial centers. These interstate entries will have to be handled accurately, promptly, and safely. No governmental authority, in my opinion, will come close to the effectiveness of depositors in making sure that this is done.

Safety of certificates in the depository will also be a matter of intense concern to the owner-manager-depositors. However, it is believed that this can be reduced to a minimal risk. In a fully operational national CSDS, security holdings can be consolidated into huge "jumbo" certificates-millions of shares each. They could even be stamped on their face to be non-negotiable and non-pledgable. Such certificates would be about as immune from illegal conversion after theft or loss as a piece of paper evidencing value could be. The only thing that could be done with them would be to send them to transfer to be split.

Accuracy, timeliness, and completeness of the CSDS's records as to securities ownership and changes therein is a subject, above all, as to which depositors can be depended upon to exert discipline with a reaction time of 24 hours, if not

sooner.

(g) What specific problems do bank trust departments and other fiduciaries, such as pension funds, have in participating in depositories? How can such problems be best overcome? Is there a need for a Federal law in this area? The only known legal problem of bank trust departments in depositing the securities they hold in CSDS is the restriction in Estates, Powers and Trust laws, already referred to.

Non-bank fiduciaries and pension funds, not under the regulation or supervision of a federal or state authority, would not be eligible to be direct depositors in the CSDS. The securities they hold would have to come into the CSDS via depositors. It is believed that most of the pension fund and large fiduciary holdings of certificates are already lodged with banks, and could come into the CSDS with the change in EPTL.

II. DEVEOPMENT OF A MACHINE-READABLE CERTIFICATE AND ACCOMPANYING

DOCUMENTS

(a) If one or more depositories are developed (see I above), will it also be desirable to develop a machine-readable stock certificate?

I do not believe so.

(b) If so, how should such a machine-readable stock certificate be developedie. what should be its form?

See the answer to (a).

(c) What specific problems are likely to be encountered in the development of a machine-readable stock certificate?

Answered in my Statement and the appendices.

(d) What would be the anticipated cost, and would the benefits to be achieved by the system justify the cost, including the anticipated time and effort necessary to put the system into operation?

Also answered, particularly in Exhibits C, D, and E.

(e) What, if any, problems of state law are likely to be involved?

No state law is known to interfere with making the certificate machinereadable.

(f) What, if any, other regulatory problems (i.c., problems of state or Federal regulatory authorities or self-regulatory bodies, such as the New York Stock Exchange or National Association of Securities Dealers) are likely to be involved?

Some exchanges have prescribed precise specifications as to the certificate. At least some of these regulations would have to be altered to provide for the machine-readable specifications. The Department of the Treasury and the FRB determine the specifications for government issues. No central authority exists, so far as I know, for the specifications of state and municipal bonds.

(g) Would a Federal law which would preempt the field of regulation, possibly to specify that all or most stock certificates must be in machine-readable form, be feasible and desirable? If so, what general form should such a law take?

In view of my conclusion, expressed earlier, that the machine-readable certificate does not now promise a satisfactory cost/benefit ratio, this question becomes academic to me.

If ever it were decided to convert to machine-readable form whatever certificates were circulating outside the depository system, then all issuers should convert to a certificate meeting exact specifications in paper and placement of the machine-readable characters. All exchanges and the NASD, acting in concert, or the SEC, might be able to do this under existing authority. I would not know.

(h) Are there special problems which bank trust departments and other fiduciaries, such as pension funds, would have in connection with the use of machine-readable stock certificates? If so, how should such problems be best resolved?

Some banks have stated that handling and counting card sized certificates would present problems. No special problems would attach to page sized machine-readable certificates, so far as is known.

(i) Are there special problems which transfer agents would have in connection with machine-readable stock certificates? If so, how should such problems be best resolved?

Transfer agents would no doubt be required to place certain machine-readable information on certificates (as would banknote companies). The agents would do the imprinting by computer or typewriter but, as is discussed in Exhibit D. the quality control here permits little tolerance. Substantial system changes would be needed.

I have been informed that there are about 850 transfer agents in this country acting for more than one issuer, and some 4-5,000 issuers doing their own transfer work or having it done locally such as by their lawyer. An industry-wide system based upon satisfactory machine-reading of characters on certificates would require strict adherence to standards by all of these, or transfer work would have to be consolidated in the hands of a sophisticated few.

(j) In what manner could the development of a machine-readable certificate be best related to the development of one or more depositories?

The two are almost mutually exclusive alternatives. With an effective depository system in operation, the pressure would be to immobilize all possible certificates which would then not need to be "read" on broker-dealer and bank premises by either man or machine.

(k) Would significant problems arise in connection with preserving the physical security of machine-readable certificates and insuring that they would not be tampered with in some fashion?

Advocates of machine-readable certificates point out that the ability to run large inventories of them through reading equipment would permit much more frequent vault or box counts than is practicable now. This would be a deterrent to theft, because the fact that certificates were missing would be detected more quickly than at present.

Others caution against relying too heavily on machine counts, which might or would not detect raised quantities, or a fake certificate with the right OCR characters on it or punched holes in it.

It is worth repeating that it is the large quantities of fungible securities now in vaults or boxes that are the prime candidates for immobilization in depositories.

(1) Would significant problems arise in connection with any transition which might be made from use of the present form of certificate to use of a machinereadable certificate?

The transition from a system based upon human inspection of certificates to one of machine-reading them would be difficult. The machine-reading system could not be effective until, say, half or more of the certificates in hand were of the new type. After that, an exception procedure would be required as long as old certificates came in.

Exhibit B, in Table VI on page 1843, shows the velocity of certificate turnaround, thus indicating how long dual systems would be required for banks or brokers, depending upon the mix of their customers.

(m) What problems have arisen or are likely to arise in the development and implementation of uniform or machine-readable accompanying documents, such as transfer instructions, delivery tickets, and forms for comparison and reclamation?

BASIC has developed and exposed for comment four uniform forms. A copy of this publication is attached as Exhibit F. Comments received to date object to a wide variety of details on these forms; only a few object to standardizing one or more of the forms at all. The comments are running about as expected. As to making accompanying documents machine-readable, the same elements of difficulty would exist as with the certificate, but magnified in some respects. Control of the quality of imprinted OCR characters would now be in the hands of thousands of typists and computers. The readability of OCR on multi-copy forms would add further problems.

Modern communication technology has developed fast. There is every reason to believe that it can "leap frog" the machine-readable document in providing a means of communicating information as to, and effecting, securities transactions among at least the high-volume financial organizations to whom machinereadable documents would be attractive.

(n) What problems have arisen or are likely to arise in connection with the development of a system of uniform numbering for the purpose of positive identification of differing issues of securities (i.e., CUSIP)? What, if anything, can be done to implement this system in a more rapid fashion?

Through BASIC's recommendation, the CUSIP identifying code now appears on most stock certificates, and is to be made mandatory on input to the clearing corporations of the New York and American Stock Exchanges by April 1, 1972. Three of the four proposed uniform forms call for the CUSIP number.

Many broker/dealers and banks are converting their internal securities codes to CUSIP. Most, however, do not see this conversion as in their short-term selfinterest. We are trying to convince those of this frame of mind that it is in their long-term self-interest to do this quickly-to have everyone using the same standards.

III. DEVELOPMENT OF A MORE EFFICIENT SYSTEM FOR CLEARANCE, SETTLEMENT, DELIVERY AND TRANSFER OF SECURITIES

BASIC's scope is confined to interindustry problems. This being the case, it considers matters of trading, comparison, and netting of broker/dealers beyond its field.

BASIC believes that the CSDS can interface with any of the several clearance systems now in existence, or known to be planned. In other words, it can effect book-entry delivery of securities upon appropriate authorizations, whether such authorizations are trade-by-trade, the result of netting transactions daily, or the result of continuous netting.

It is planned that the NYCSDS will equip itself to net and settle daily the cash side of deliveries-against-payment among its depositors. However, it could handle any "free" delivery of securities by book-entry, with the cash side of the transaction (if there is one) handled outside the system by the depositor. The fact that NYCSDS handles cash within its system would not require any other regional CSDS to do the same, if some other arrangement better suits its needs. As to the transfer of securities, BASIC's approach is to obviate a large proportion of the transfers by substituting book-entry deliveries (which, of course, require no physical transfer).

A study of 6,151 cancelled stock certificates of 27 companies turned in for transfer, about one-third each during the months of April, May, and June 1971, indicates that:

1. The 6,151 certificates were involved in 22,796 physical movements nationwide during their periods of circulation. This is an average of 3.7 moves per certificate. Of the total movements, 9,573 were by New York brokers and banks. 2. Of the 9,573 movements, 54% would be eliminated-book-entry substitutedunder a fully operational New York Comprehensive Securities Depository System (NYCSDS). With a fully operational national Comprehensive Securities Depository System (national CSDS), the New York movements would be reduced by 72%.

3. For all brokers and banks in the U.S., a national CSDS would reduce physical movements by 74%.

4. Part of the reduction of broker and bank movements is offset by an increase in movements handled by CSDS-from an actual of 2,677 to 4,518 under a NYCSDS and 3,186 under a national CSDS. If one of the two handlings of the same certificates by CSDS and transfer agents were eliminated, for members of the New York financial community a NYCSDS would change a decrease of 27% to a decrease of 64%. For a national CSDS, a decrease of 52% would be changed to a decrease of 78%.

5. On the assumption that all depositors will immediately deposit in CSDS all certificates received from outside the system, the handlings of certificates in the names of individuals by the financial community will actually increase somewhat under CSDS. Moreover, as most of the other certificates are immobilized in a CSDS, individuals' certificates will come to represent most of the certificate handling problem of brokers and banks. These developments will undoubtedly increase the pressures and incentives to attract individuals' holdings into the system.

6. Movements of certificates in the names of foreign financial institutions and their U.S. agencies are not inconsiderable. Accommodating these in CSDS in due course is suggested.

7. Recognizing that a bias toward New York holdings exists because the study included no certificates cancelled by transfer agents in other states: of all movements of certificates handled by brokers and banks in the U.S., 80.9% were of those located or clearing in New York City; as to movements handled by non-New York brokers and banks, 53% were accounted for by 5 states and 75% by ten states.

8. In terms of effect on transfer volume, a NYCSDS would eliminate some 40% of certificates issued, and a national CSDS would eliminate two-thirds. This is under conditions existing when the examined cancelled certificates were originally issued, and may be substantially altered by the recent trend to "jumbos". 9. The average age of the cancelled certificates examined was 23.5 months, but this is by no means a helpful average. For Cede certificates, 83% were returned for transfer within 6 months of issuance. At the other end of the scale, one individual's certificate was returned after having been outstanding 33 years and 2

months.

10. In examining the denominations of the 6,000-odd certificates, the 100-share certificate dominates. Many of these undoubtedly could have been consolidated into more-than-100-share certificates without sacrificing efficiency or economy.

These findings, including any limitations regarding them, are discussed below. The sample; research method

Twenty-seven common stocks were chosen for the study: 18 NYSE listings, 5 AMEX listings and 4 OTC. Approximately 200 cancelled certificates were chosen for most issues; in a few cases some 600 were selected. No more than three months' transfer activity was examined as a general rule. The certificates examined were all received and cancelled by transfer agents in New York City. To the extent that certificates turned in for transfer in other parts of the country would show a different pattern, there is some bias in the sample. However, it is believed that this bias is not so great as to disturb the usefulness of the results. A summary of the certificates examined, classified by the names in which they were registered, is contained in Exhibit 1.

The movement of each certificate was traced from the registered owner to other parties by the various endorsements, guarantees, stamps, signatures, etc., on the back of the certificate. In some cases, judgment had to be used as to sequence of the steps in the route a certificate took. Certain paths were known to be impossible (e.g., there could be no movements directly from an individual to CCS, even though the record on the back of the certificate so suggested). It should be noted also that not all those through whose hands a certificate passes evidence this fact by written notations on the back. However, it is believed that these omissions are not sufficient to invalidate the order of magnitude of the statistics derived.

A file card was prepared for each certificate examined, containing the actual or nominee names of those who handled the certificate. Later, each name was classified into one of eight groups: transfer agents, CCS, New York brokers, New York banks, non-New York brokers, non-New York banks, individuals, and others. Most statistics were gathered by these groups.

Exhibit I is the set of detailed instructions used to gather the data.

Assumptions regarding the methods of operation of the NYCSDS and of the National CSDS

In order to compare the movements which actually took place with those that would have taken place had the NYCSDS and national CSDS been in existence, several assumptions were made, as follows:

(1) New certificates registered in the names of individuals do not show whether they were obtained for that individual by a broker or bank, or directly from the transfer agent by the individual himself. The latter procedure is said by transfer agents and others to be minority practice. Wherever the back of the certificate showed that an individual subsequently disposed of the certificate through a broker or bank, it was assumed that he had received the certificate from the same source and this was recorded as an actual movement. This assumption was carried through for projected movements.

(2) In the NYCSDS, it was assumed that all banks and brokers in New York City except foreign agencies would by participants; in the national CSDS, all banks and brokers in all states would be participants. Counted as a New York broker was any broker having an office or clearing in that city.

(3) It was assumed that any time a participant received stock from a nonparticipant in either system, he would deposit that stock in the depository who in turn would transfer it into its nominee. Any time stock was delivered by a participant to a non-participant, the participant would withdraw the stock from the depository in order to make the delivery. This withdrawal would be by transfer, and the new certificate would be registered in the name of the non-participant. (4) Any delivery between two participants was assumed to be made by bookentry within the depository.

Using these assumptions, a hypothetical path was constructed for each certificate to follow under either system using the actual path followed to identify those making deliveries of the security involved.

The movements of the certificates are summarized by the eight groups of holders and handlers in Exhibit 2, are given in greater detail according to the name on the certificate in Exhibit 3, and are shown in complete detail in Exhibit 4. Those examining the data extensively will note a few figures that do not cross-check with others by one or two movements. These discrepancies were detected late in the summarization process. They were not adjusted because the large amount of work required would result in insignificant changes.