It has been a hectic start to the new millennium for the securities
industry. Hot on the heels of Y2K preparation comes something even more challenging: T+1.
Currently, securities trades in Canada are settled in three days (T+3). With the March release of a T+1 white paper by the Canadian Capital Markets Association (CCMA) , the Canadian securities industry has confirmed its intention to settle trades in just one day (T+1). The target date for this transition is June 2004, with testing in 2003, to coincide with a parallel move in the U.S. The US industry estimates that the cost of these changes will amount to US$8 billion – more than it spent to prepare for Y2K. To be ready in time, CIOs will need to be involved early.
In the late 1980s, when trades were settling on T+5, the U.S. securities industry announced that it would move to T+3, a move that was accomplished in 1995. To avoid an outflow of business to the U.S., the Canadian industry moved to T+3 settlement on the same day. Essentially, the T+5 process was left intact — activities were just completed sooner, typically by adding additional people to the back office and more batch automation.
“There is significant manual intervention by all parties to institutional trades,” says Kenton Warren, Vice President and Director of National Securities Operations Division at RBC Dominion Securities Inc. “This is primarily due to a lack of technological interoperability among dealers, money managers and custodians. This intervention results in error rates of 15 to 20% at various points throughout the process.”
In the decade since the move to T+3 was announced, North America’s capital markets have seen substantial increases in trading volumes and volatility. These circumstances have kept the concern about the risk of unsettled trades at the forefront, justifying an even shorter settlement cycle. The planned move to T+1 is a response to this situation.
The CCMA’s Institutional Trade Processing T+1 White Paper suggests the move to T+1 will require substantially more changes than the move to T+3. The latter transition did not involve significant changes to the post-trade processing model. Co-authored by IBM Global Services and Katamax Solutions, the white paper argues that the current processing model will have to be replaced and most activities will have to be automated to settle trades by T+1. IT departments will need to understand the business requirements in order to build a technological infrastructure to execute straight-through processing (STP) required for T+1.
Obstacles to T+1
Current practices. According to the white paper, current practices present a key obstacle to T+1 settlement. On one hand, broker/dealers tend to wait for an entire order to be filled before notifying money managers that a trade has been executed. This wait often means that trades are not communicated until late on trade day (T+0), T+1 or even T+2, leaving money managers little time to calculate the client allocations underlying the trade block. On the other hand, money managers often send allocations and settlement instructions to broker/dealers and custodians late anyway, waiting to first rebalance their portfolios at the end-of-day.
Worse still, some money managers do not act upon the informal notice of execution, and prefer instead to wait for formal confirmation of the trade. These confirmations are typically received on T+2 or later, depending on whether they were sent via fax, computer, or post. In fact, due to these kind of processing delays,
approximately 40% of trades are not ready to settle until T+3 itself. Most of these manual activities will need to be automated to ensure that all trades are settled by T+1.
IT departments must consider that some of these processes are internal to the organization whereas others will require linking to external business partner’s systems.
Technology Infrastructure. The industry experts identified the existing technology infrastructure as another leading obstacle to T+1 settlement. Industry
participants in Canada use as many as 30 different platforms, which are generally not interoperable. Also, systems typically run in batch rather than real-time, meaning that communications among parties to a trade are slow. To circumvent the lack of interoperability and real-time transmissions, parties to trades commonly use telephones and fax machines. This requires recipients to manually enter the information into their own systems, which is very risky. Any single ‘typo’ can lead to a failed trade that will be costly to repair. A manual process also consumes a lot of time – something that will be in short supply with a T+1 deadline.
To settle trades by the end of T+1, existing systems will need to be upgraded or replaced to enable interoperable, real-time straight-through processing; and manual processes will need to be automated.
Incorrect Data. A third leading obstacle to T+1 settlement is incorrect data. Money managers typically break down trades, with allocations to numerous institutional client accounts, each of which may use a different custodian. The broker/dealer must cross-reference the money manager’s identifier with the broker/dealer’s own identifier to find the custodial account number. When trades appear in the clearing system, custodians clear a trade for settlement if the details match those received directly from the money manager. However, due to the very cumbersome procedures that all participants must go through, many trades do not match. This delays settlement and requires the parties to collaborate to repair failed trades.
According to SWIFT, a global industry protocol group, the cost of repairing a trade on T+3 is quintuple the cost of processing a matched trade. Automated procedures with standardized data would not only enable T+1 compliance, but would reduce operational costs as well.
Five Alternative
Processing Models
To overcome these obstacles, the CCMA white paper sets out five alternative T+1 processing models and outlines their relative advantages and disadvantages. Two of the models are based on a solution proposed by the Global Straight-Through Processing Association (GSTPA). Two others are based on vendor proposals, while the fifth is the “New Design”, a theoretical institutional trade-processing model for Canada.
The CCMA has requested feedback from the industry-at-large on these models and the New Design. The New Design will serve as a benchmark to judge the merits of these other models. The decision on which model or models will be adopted in Canada will be made, hopefully, this summer.
The New Design model “is focussed on addressing the concerns of the Canadian marketplace”, says Allan Cooper, Chair of the CCMA. If the New Design were adopted, all key industry participants -money managers, brokers/dealers, custodians and the depository – would be linked for the first time via an interoperable, common, post-trade matching system. Electronic notices of exe-cution from brokers/dealers and allocations from money managers would feed this system. Money managers would attach custodial client account data to the allocations, which could be submitted either before or after the notice of execution. The new system would match the execution and allocation details and forward them to the custodian, via the settlement system, in an already affirmed status. With real-time processing, trades would settle by T+1.
The New Design would offer several advantages. First, it would specifically address the problems with existing technology and systems. The new common system would improve communications among all the parties by eliminating the need for faxes and phone calls, making manual processing the exception rather than the rule, replacing batch runs with real-time transmissions, and connecting systems seamlessly. In fact, the common system would be designed to inter-operate with other models. Secondly, it would eliminate redundant steps by allowing each party to enrich trade data in an orderly sequence. Money managers would not have to worry about addressing as many failed trades, broker/dealers would not have to worry about cross-referencing critical data, and custodians would not have to do any matching. T+1 settlement would be achieved in a way that would benefit all segments of the industry.
Implications for CIOs
Under any T+1 solution, IT managers will need to renegotiate their service level agreements. Internal SLAs will reflect the new business level requirements for systems availability, disaster recovery processes, and real-time straight-through processing. External agreements will need to address the new standardised interfaces and interoperability with other partners. As well, outsourced operations will also need to be evaluated in light of the new business model.
IT departments will need to lay the groundwork internally. Uninterrupted electronic processing across and throughout enterprises will be essential for T+1 settlement irrespective of the model that is eventually adopted. Accordingly, firms need to develop a comprehensive, focused assessment of their processes, technology, systems, and organizational effectiveness.
Since current processes are a legacy of T+5 and will be modified to comply with T+1, they should be catalogued in detail. Legacy systems also should be assessed to determine if they will have to be upgraded or replaced, to enable real-time trade processing and the automated flow of information between the front, middle, and back office. A review of the current state of the organization may also be desirable to ensure that the staff is motivated for a smooth transformation, notwithstanding the reality that many of their jobs will likely change significantly or become redundant. Furthermore, firms should develop an internal vision to work towards, and evaluate the business and technical infrastructure initiatives that they will require to achieve that vision.
By the time industry consensus emerges on a T+1 model, IT departments should be ready with an enterprise-wide blueprint for straight-through processing development and implementation.
Need for Further Investment
Once it becomes clear what the T+1 processing model and solution will be, firms will have to make even more substantial investments. Linkages to the new solution, for all types of securities, will need to be developed. Internal processes will need to be re-engineered and continuously modified while technology and systems are designed and rolled-out. To accommodate new data standards and industry protocols, enterprise reference data and data ownership may have to be completely reorganised, and systems harmonised accordingly. New and upgraded systems will have to be designed, built, and tested. Staff will have to be reorganised and retrained.
CIOs have their work cut out for them now in order to kick-start their firm’s T+1 projects. In particular, they will need to assess their current environments to identify STP gaps, work closely with the business leaders to jointly choose a T+1 solution, and get involved with industry T+1 working groups (such as the CCMA) to share their IT strategy solution challenges. As the industry learned from the Y2K
experience, every day between now and the target date will count. This will be
especially true of a transformation as dramatic as T+1.
Nick Covelli and Greg Lee are management consultants for the financial services sector with Business Innovation Services, IBM Global Services. They co-wrote the Institutional Trade Processing T+1 White Paper for the CCMA. Special thanks to Doug Young, also a management consultant of IBM Global Services, for his assistance in this article. E-mail: ncovelli@ca.ibm.com, greglee@ca.ibm.com; dyoung@ca.ibm.com Web: www.ibm.com/industries/finan-cialservices/stp