The Bank of England Task Force set up in the wake of the failure of Taurus has recommended a new electronic settlement system, CREST, which it aims to introduce by 1996. CREST, it maintains, is the simple solution that will give the London market the settlement facilities it needs to remain competitive into the next century.
The Press and City have, so far, bought the vision. Comment has generally been favourable, if muted.
But the Task Force's recommendation is brief. Whilst CREST appears attractive and may form the basis of a solution, the report leaves many questions unanswered, some of which it should perhaps have addressed. The desire to move forward from the Taurus debacle should not obscure the lessons learnt from it; there is a danger that in answering these questions at more detailed planning stages, the objective of creating a cheap, effective settlement system may be lost.
The Task Force decided against simply adapting Talisman, the Stock Exchange's existing computerised settlement system, as a replacement for Taurus because it would have been more expensive and inflexible in the long term. It also narrowly (and rightly) decided against introducing a central nominee that holds title to all stock in the system, or a central registrar for UK equities.
Instead, the Task Force has opted for a system based on electronic book entry transfer between participants. Unlike Taurus, this system is directed at achieving two main objectives: daily updating of the register and delivery versus payment (DVP).
The Task Force states that stock will be held in the system in separate participants' accounts and legal ownership will remain with the account holders. Participants will have to satisfy participation criteria. Many institutional shareholders may well become participants (assuming they meet the criteria) and so will be able to hold stock directly in the system in their own name.
Equity trades between parties holding stock within the system can be settled against Sterling payments on the basis of electronic messages (book entry transfers) transmitted between participants. After the trade, messages will be sent to the registrar and legal ownership of the stock will change only after registration.
This is one of the major differences with Taurus; in Taurus, ownership relied upon system records, meaning that it was critical to ensure that the legal and system status of the computerised records was watertight.
It is envisaged by the Task Force that CREST will be able to handle the vast majority of stock events, such as rights issues and scrip dividends, currently handled in Talisman. But the Report contains no details of how CREST will deal with the various complexities arising from these corporate finance events.
Payments under the system would be based on existing banking relationships operating through CHAPS (The Clearing House Automated Payment System). A bank acting for CREST (possibly the Bank of England) would be at the centre of the payment system and have real time accounts for each of the settlement banks, who would be direct members of the CHAPS network. Under this system a delivery versus payment arrangement would eventually be possible, removing counterparty risk. This was not a feature of Taurus.
Because legal ownership still depends on registration and because the system is optional, with participation only open to those meeting defined criteria, it is a great deal simpler than Taurus. If companies wish shareholders to have the opportunity to settle trades through the CREST system they will have to ensure that their registrars are members and meet appropriate service standards, providing a rapid response to messages of trades. The Task Force envisages cost penalties for those failing to meet these standards.
The Task Force considers that the structure for ownership and transfer of securities under the CREST system is relatively straightforward so that fewer changes will be required to existing law. They think that what is required can probably be achieved through secondary legislation under the broad enabling powers intended for Taurus in the Companies Act 1989 (section 207). Ultimately the Task Force considers that ownership of CREST may be vested in a consortium of outside investors.
All shareholders of a company that joined Taurus would have had to have participated in the system. This caused many problems. The interests and objectives of small private shareholders are often very different from big institutions and much energy was expended trying to meet their sometimes contrasting demands.
CREST focuses upon the wholesale market, which is probably the correct choice; it may be the price which has to be paid for reducing counterparty risk and bringing London into line with other international markets.
But it does mean that there may be a tendency for retail shareholders to have to hold through nominees.
The SIB is looking at a proposal that the Financial Services Act should be amended to bring nominees within the net of SFA regulations, and this may reassure trustees that their interests are protected where nominees are used.
Under the CREST system there would be the option to retain share certificates, which may be attractive to private investors. The system will provide a range of functions for investors holding certificates but the full benefits of electronic settlement will only be available to those investors using a CREST participant to act on their behalf. The Task Force has suggested that, for shareholders who only occasionally buy and sell, the certificated route will still be the cheapest way of handling shares despite the fact that, particularly on the buying side, charges for certificated settlement may be relatively higher than a book entry transaction because processing certificated transfers is relatively costly. It is still not entirely clear how the two systems will be integrated - the Task Force envisages paper movement being replaced with electronic messages wherever possible.
The Task Force made it clear that it views the extent to which a company is able to identify the beneficial owners of its equity stock as very important.
In the current system visibility is limited because of nominee holders and delay in updating registers. Although under the CREST system company registers would be updated on a daily basis, the nominee problem is likely to be exacerbated at the retail end of the market. This means that in order to identify the actual shareholder the company may have to issue section 212 notices (under the Companies Act 1985) requiring the nominee to reveal for whom it is holding the company's stock.
The Task Force set out in its report a number of the ideas aimed at addressing this problem, which it considers should be dealt with in the next stage of development. It recommended that the Bank of England should assume leadership of the next stage of the project, taking advice from a steering committee drawn from the market. It may be that some of these ideas will be pursued in that phase.
On the face of it CREST seems to be a simpler system than Taurus and it has some attractive features. If implemented properly it is likely to be easier to integrate with traditional company law procedures than a central nominee system, where legal ownership through the share register by participants would effectively be ignored. But the core proposal is still in outline form and does not address some important issues:
The nominee problem is, of course, not new. It has tended to be a feature of trading at the institutional level; Talisman, for example, is based on a central nominee. CREST is unlikely, in itself, to make this position worse.
But the current use of nominees is in itself a legal quick-fix, designed to reconcile the old paper based systems with new demands. Companies are creatures of statute. The Companies Act has evolved from rules carefully drawn up earlier this century to govern transfers of shares and the relationship between shareholders and company. It bases many of its rights on the stock transfer form and share certificate, and on the assumption that the registered shareholder owns the shares; it does not contemplate large numbers of shares being held and transferred by nominees.
Taurus confronted many of such issues head on. Its structure was designed to deal with many of the questions that arise when you seek to legislate for a system that will replace years of legal practice. The concept of entitlement tried to define the legal effect of entries made and deleted in a computer system, thus providing a mechanism for book entry transfer, and it in part defined the role of controllers (essentially the nominees holding entitlements for shareholders). The creation of an interlocking web of operators and controllers sought to introduce security measures to prevent fraud.
It is clear that CREST will not be as all embracing as Taurus. It may well be that compromises have to be made to the ideal legal solution if the new system is to avoid the technical complexities that were the downfall of Taurus. We may have to accept that those holding through nominees have different legal rights from those that do not. But these compromises need to be identified and accepted now.
The team at the Bank of England is likely to report on the detailed specifications of CREST by April 1994. One of its first aims is to lay down a clear template with a computer system and legal regime CREST will follow. If they do not do this effectively, there is a real danger that this simple system will become more complicated as additions are made to try to address some of these issues on a piecemeal basis. CREST will begin to expand in the same way as Taurus. Without firm control this new calf could become another uncontrollable bull. SMF/RJD
Based on a comparison in the Task Force's Report
One of the major reasons for reviewing the settlement system in the first place was the need to shorten the period between trade and settlement to reduce the risk of a failed trade or insolvency of a counterparty. The "Group of Thirty" recommended in 1989 that settlement should be reduced to 3 days by 1992.
As a first step, the Task Force has recommended a move, by July 1994, to 10 day rolling settlement, instead of the current account system where trades are settled in batches every two weeks. Under the current system, a deal which takes place toward the end of an accounting period will of course be settled in less than 10 days but a move to a 10 day system will have the benefit of uniformity. The Task Force wants five day rolling settlement by the beginning of 1995. Whether this is a realistic objective under the existing paper-based system will probably depend upon the cost effectiveness of widening the availability of book entry transfer facilities under Talisman before the introduction of CREST.
The following diagram is also available in the PDF version of this article: