Why T+1 may never come to pass
23 November 2001 | 42 views | 0
During an investigation into the business requirements for settlement on trade date plus one, Paul Pickup, director of consultancy Trading Technology Ltd concluded that T+1 is neither desirable nor attainable.
Trading Technology recently concluded a study of leading global brokerages into the risks involved in reducing settlement cycles to T+1. The conclusions show that while everyone is caught up in the stampede for shorter settlement cycles, few have thought through the risks that this introduces, and even fewer have identified any benefit. Trading Technology argues that the aim of the regulators and central facilitators should be to reduce the complexity involved in the settlement process, and allow straight-through processing (STP), rather than merely tightening the settlement window.
The mantra of the central regulators is that delayed settlement is a hang-over from the days of deal tickets, and that in this electronic age there is no need for any delay at all. Therefore, shorter is better. We have seen the reduction in settlement periods down to T+3 in most markets and the DTCC has decreed that all settlement will be T+1 in the US in 2005. Other countries are huddling together to try and match or even beat the US, in the assumption that they are right, but is this particular Emperor really wearing any clothes?
Everyone is missing the point. They are labouring under the belief that shorter settlement cycles are better. However, it is not the latency that needs to be addressed; it is the efficiency and complexity of the overall process.
The central facilitators and regulators have a responsibility to help brokers streamline the entire settlement process, both between the buy and sell side and for cross-border transactions. The central facilitators should support the GSTPA and Omgeo initiatives, which are at the mercy of the whims of the market participants. Todd Eyler’s recent Forrester report entitled “The real Benefits of T+1” go so far as to predict the bankruptcy of GSTPA in the absence of any real impetus from brokers to use the service.
Are the timescales too long?
The duration of this particular project, now extended to 2005 in the US, is itself a risk. It is possible to calculate the probability of a project failing by its sheer longevity, ie the longer it takes, the greater the risk. T+1 is already a four year-plus project in the US, and thus has a high risk of being delayed indefinitely.
T+1 is a poor relative to T+0
The shortening of the settlement cycle to T+1 will put intolerable strain on the established bureaucratic settlement processes, without delivering any notable business benefit. Instead, there will be problems of cash flow management, margining, stock lending and risk management. These are made worse under T+1 as there is less time to identify risks and manage them. Only the utopia of real-time settlement will actually bring any benefits.
The Forrester report “The Real Benefits of T+1” accurately predicts many savings in the brokerage community as a result of T+1, but apart from the risk of cash or stocks being held in suspense for one-day rather than three days, the benefits are all from straight-through processing rather than from T+1.
While it may be perfectly possible for national business to be settled in one day, it will be very different in global markets. If the client is in Hong Kong and instructs trades on the US markets, the concept of a trading day is lost. By the time the executions come back from the US market, the HK client will have gone home. Who will confirm the allocations?
This time difference really means T+1 becomes T+1/2. The situation becomes worse when orders are passed back the other way, requiring settlement to occur the same day.
Although the GSTPA solution goes some way to solving the allocation process, it still requires both parties to be able to respond to post-trading events. Therefore, there will either be a need for pre-allocation standards and algorithms to be established, or for brokers to man 24-hour operations to confirm executions and allocations as they happen around the world.
Client cash collection
Private clients will either need to deposit cash up front for executions, or brokers will need to stretch credit limits in order to settle in time. Again this is due to the delay caused by central banks in facilitating payments. This increases the risks for the broker, in addition to the cost of systems re-engineering.
Banks not geared up for Continuous Linked Settlement
While the pace of technology quickens for central depositories, the central banks are far from ready for real-time payment. Many still operate on an overnight batch basis, which given the time zones that are operated on by the global banks, will cause latency in the settlement process. The problem is far greater where FX transactions are involved, being currently a two-day settlement process.
System costs – the need for the re-architecture of settlement processing
In a recent survey of brokerage firms, only 40% were thinking about re-architecting their operations systems to be resilient to failure. While the current settlement process is three days, a four hour system outage is an inconvenience, but with T+1 it will be a disaster. Brokers are concentrating on the need to process transactions in online and away from batch mode, but they are ignoring or unable to deal with the risks involved in system failure.
The cost for brokers to re-architect their applications is estimated to be $3.3bn in the US alone, with a further $2.9bn spent on standardisation. Yet there will be no discernible benefit for the brokers or their customers.
Rather than reducing risk by having shorter settlement periods, it is likely to be increased. If all your business goes “straight through”, there is a danger that this will apply to risk too, and it will be undetected. Nick Leeson managed to hide futures positions in back-books for months, irrespective of the settlement period. The most important points are “the accounting points” in the process, and that the change needs to be led from this viewpoint in addition to a technical one.
As far as we are aware, Trading Technology is the only organisation to have developed a re-engineering programme along these lines.
There is little point in racing towards T+1 without the central facilitators and industry bodies taking a step back to eliminate the existing bureaucratic processing in securities settlement.
The processes that induce latency of central banking payment, netting and order allocation need to be re-engineered in order to move towards near real-time settlement.
In this respect, the central facilitators should concentrate their efforts into ensuring the success of the streamlining process, such as those offered by GSTPA and Omgeo, rather than turning the screws on an already tight settlement cycle.
Trading Technology is a specialist consultancy in technology issues for central markets, offering project management, business analysis, IT marketing and strategy advice.