.jpg)
There are a number of individuals who would consider the above to be a statement of the obvious - there is no HOW, it just does! It is for this reason that we need to be clear on what we are talking about. The question needs to be put into context and any assumptions defined.
The context of the question assumed in this article is the extension of the STP pipe from its traditional domain of the back office to the "rarefied" arena of the dealing desk.
The terms "order routing" and "clearing and settlement" are familiar enough, and represent two activities at either end of the investment management chain. These activities are already linked in a business process, in that one follows on (eventually) from the other, but for the purposes of this article we will assume that we mean linking electronically. Further, linking two activities at each end of the pipe implies that all of the activities in between are also linked. Information captured in the early stages of the trade cycle needs to be managed all of the way down the pipe otherwise we will get "leaks".
In the current situation, most companies are still divided internally into the part that processes the original order and records the subsequent trades and the part that performs all of the tasks necessary to ensure that the trades are complete as contracted. The clear division of labour in a business model that is almost universal strengthens this front office/ back office split. It is further strengthened by the spurious but common belief that the front office is revenue generating whilst the back office is a cost centre.
An operational risk opens up because decisions about new markets and products are unilaterally made without reference to the cost of actually doing the business. Integration means integration not only of systems, but also of different parts of the office, and the removal of this clear division to create an entity that works as a whole to carry out the business of the company from inception to completion, viewing the whole operation as a single revenue generating entity.
This article will explain why this extension of the STP pipe is increasingly becoming a requirement of doing business in global capital markets and how it delivers integration. The drivers that are inextricably leading us down this path fall into five broad categories, which are described below in no particular order of priority.
Risk reduction
All investment opportunities offer a certain return for a certain element of risk. Not all of the risks are a necessary corollary of individual investments. We can divide risks into the acceptable and the unacceptable. The more the unacceptable risks can be reduced, the better a business opportunity the investment becomes, as the risk/return ratio improves. In the current investment environment, acceptable risks are the macro-economic factors and the company performances that affect the price of a security. The unacceptable risks are those which arise after an investment decision has been made, particularly that the execution of that decision may result in a failure.
Competition
Competition on several levels drives the need to link activities more closely. A single company needs to compete against its peers within the same market to win the investment capital being directed into that market. One level up, a particular market is in competition with other markets, be they in different regions (e.g. UK verses Germany) or based upon a different business model (e.g. Electronic Crossing Networks, or ECNs verses traditional stock or bond markets). This exerts pressure on the markets themselves to become more efficient and more attractive to investors. As the markets change in response to the competition, so the pressure turns back upon the participants to adapt to the new conditions.
Globalisation
Investors and money managers are increasingly looking for opportunities to spread their risks geographically to avoid the dependence on local market and economic conditions in a particular location. The volume of cross-border trading is increasing by some estimates at a rate of over 40% per annum. This is associated with a high rate of trade failure, which is proving to be very expensive for the global securities industry. Not only does this pose a significant cost burden for parties involved in the process, but it also creates an enormous amount of operational risk for participants.
Homogenisation
As the distinction between asset classes blurs, and the investor's desire to spread the risks increases, the need arises to deal with all asset classes within a single organisation. This allows a service provider such as an asset manager to offer a fully integrated product to the investor. External integration can only be efficiently offered by the linking up of the processes internally, and may not only be from front to back office but also across asset classes.
Regulatory pressures
Markets, in complying with the requirements of governments to protect their electorate, and the requirements of the participants to work on a level playing field, introduce regulations to ensure that business is being conducted as fairly and efficiently as possible. A good example of this is the current intention to reduce the trade settlement date to become one day after trade date (T+1) in the North American continent. Internally this increases the need for linking all of the processes within a company together, but this will not have the desired effect unless all of the companies within the market co-operate to bring about the desired change by linking more closely together themselves.
Requirements
On the premise that a T+1 settlement cycle is the industry's ultimate goal, what changes are required to ensure that firms are actually in a position to achieve it? STP is a destination and not a journey - only the transport mechanism changes. Therefore, before beginning that journey, one needs to ensure that internal transaction management is sufficient for the task. Continued reliance upon a patchwork of legacy systems and a high degree of manual intervention is not the solution. With trading volumes set to treble to 150 million per annum by 2002 and a T+1 settlement cycle in North America by 2004, it can clearly be seen that action needs to be taken in the short term.
Since the G30 recommendations of 1989 first underlined the need to eliminate, or at least reduce the mountains of paper, on the surface a lot of progress has been made, particularly in the post-trade space. Electronic Trade Confirmation (ETC) applications are now an accepted part of daily trade processing for a number of buy-side and sell-side houses across the globe. After a relatively slow start, there is now no doubt that efficiency has been achieved across the value chain, but unfortunately not end-to-end. STP remains a processing chain, each link being represented by either a separate processing engine or a middleware application. In very few instances are these links welded into one STP pipe. A chain is only as strong as its weakest link. Without full integration and connectivity between all the elements that make up the internal trade process, STP will remain Straight-To-Printer. Indeed, at least two recent independent studies have shown that communications between all parties to the trade is often varied and disparate. The results of one revealed that:
- 28% of communications were via paper and fax
- 26% used proprietary connections
- 27% used SWIFT messages
- 19% by other means
Contemporary studies have also shown that account allocations sent electronically to the broker saved on average between two and six hours per trade between execution and positive affirmation. In addition, 30% of all ETC exceptions were caused by incorrect or amended account allocations, the majority being where allocations continued to be advised by phone or fax. Given that a T+1 environment will effectively mandate that trades are affirmed or matched within six hours of execution to ensure settlement, it is clear that there are going to be a number of challenges along the way.
Fortunately, there are steps that can be taken to smooth the transition path. An effective way to reduce internal errors is to ensure that any reference data is standardised across all databases. Incorrect data upstream of the main data enrichment process is the main contributor to failing trades. In terms of both time and headcount, trade repair is disproportionately expensive and may cost up to three times the charge of confirming the original trade. Secondly, adherence to industry agreed standards and protocols are essential. Examples of these may include the ISO 15022 Data Field Dictionary or FIX. Similarly, processing performance should conform to any industry guidelines such as the ISITC 'ETC Code of Best Practice'. Finally, data ownership between front and back offices should also be considered.
In today's model, the front office is responsible for the pre-trade elements such as portfolio management, investment decision, order routing and execution. The traditional back office and trade support roles are geared towards the management of the fund allocations, confirmations and static data maintenance plus, of course, settlement. However, in order to accommodate shorter settlement cycles, this model must change. T+1 effectively mandates real-time processing and real-time data feeds. For example, fund allocation by the investment manager should be made available to the executing broker either at the point the order is given or, immediately after the NOE is received and matched.
These requirements begin to drive the need for organisational transformation. Roles and responsibilities must change. For example, the front office must become the owner of, and take more responsibility for, trade data; whilst the back office increasingly focuses on exception processing and repair and migrates towards more of a client facing and support organisation.
By definition, pure STP suggests that any executed trade would be handled automatically throughout its lifecycle without manual intervention. Unless checks and balances are included in the 'pipe', STP could mean 'Straight-To-Prison'. In a high volume environment a loose control of trading and settlement activities could allow non-authorised or inadequately verified transactions through without managerial intervention. Therefore, even in any future highly automated STP model, the operations department undertakes the vital role of transaction lifecycle control. With this consideration in mind, a simple definition of STP can be formed.
"The automated handling of transactions without resort to manual intervention or message repair, except for reasons of policy, across the entire value chain."
Within this environment, the back office function becomes further integrated with that of the traditional so-called "revenue-earning" front office. Neither can exist without the other and T+1 will underline this fact. Indeed, a general re-alignment of settlement cycles across the globe will ensure internal system integration moves to the top of the agenda. There is little doubt that 24-hour exception-based monitoring will become standard as cross border trading increases due to continued deregulation and as trading begins to 'follow-the-sun.'
Having established a closer knit STP environment, the next stage is to maintain or improve the level of workflow efficiency. No firm is likely to want to achieve 100% STP across all instruments and processes. Market statistics provided by the quarterly STP Benchmarks publication, for example, consistently show that the optimum level of automation to achieve the majority of the benefits of STP is between 80-90%. There is clearly a trade-off between IT investment and the law of diminishing returns.
The old adage "what can be measured can be managed" is the clue. An understanding of how systems (and people) are performing in line with external expectations is crucial. The establishment of internal performance benchmarking metrics will enable management to closely monitor performance breaks. In a highly automated environment where STP responsibility has expanded upstream the weakest link, aside from trade repair, remains human intervention; whether that is a trader manually entering account allocations or the fund manager verbally passing information to his broker. Any tardy action on the part of any one of these front office data contributors will impede the process downstream. Certainly, verbal inter-action between a broker and his client when instructing trade details is open to error and can cause up to 30% of the confirmation failures within the ETC community. Figures for the non-automated world are far from clear.
As already implied the STP pipe will not be complete without clearing, settlement and reconciliation (the latter an element often ignored by STP commentators). The requirement to match on trade date emphasises the need for accuracy and speed. Custodian banks and settlement agents are the vital entities in the settlement process as, without clear instructions from both parties to the trade, the transaction is unlikely to settle. As has been already highlighted, the settlement parties now need to be in a position to act upon an instruction within hours rather days.
Whilst a majority of settlement instructions are received over the SWIFT network e.g. MT52x, MT53x or MT54x message series, all too many require routing through a proprietary interface to the relevant bank. Although the latter provides obvious benefits to the recipient bank, for the investment manager it can necessitate access to a multitude of disparate systems. This in turn can impair efficiency where internal infrastructure is based upon internal legacy platforms or applications. There are many firms who will find that the major investment required to re-engineer their STP process, coupled with the pressures of time-to-market, precludes them from pursuing this route. Tactical outsourcing (either IT infrastructure or business process) could be an alternative and viable solution in these situations. Although this may not be the ultimate panacea, and a decision not to be taken lightly, it is now beginning to find favour. With approximately 30% of buy-side costs being attributed to back office operational and accounting costs and 20% on IT, there can be benefits to this approach.
As mentioned earlier, there is a need to link internal and external processes earlier in the value chain. For example, in order to integrate both buy-side and sell-side, it becomes vital to automate the upstream trading data exchange. This may either be by generating an order directly from a real time price feed, an Indication-of-Interest (IOI) or by connecting the participant's order management system to the appropriate trading counterparty. In today's trading world this tends to be through point-to-point interfaces using the FIX protocol.
In the future this may well be complemented or replaced by a connection to a central hub to which other participants/counterparties are also connected. In this way the inconsistencies of use and implementation of FIX between various parties can be 'smoothed' out or mediated. Such a hub could be provided and maintained by a neutral independent party such as Reuters. Certainly, a large proportion of sell-side IT spend will be focussed on external connectivity during the next two years; that of the buy-side on more internal STP issues (up to 26% according to a recent Tower Group report).
It becomes clear that linking order routing to clearing & settlement and all that it implies moves us inextricably towards integration of the trade cycle as a whole (not just those areas already automated). A pre-requisite of linking these two activities is capturing the necessary data that describes and defines the trades, as early as possible. The benefits of integration are becoming more clearly defined as the business strives for greater efficiency, in order to compete. The primary reason for this is to reduce the costs of processing transactions, as these costs become more evident and important as a differentiator.
How
Based on the above we are now able to draw some conclusions. The integration of order routing with settlement and clearance is all about extending the STP pipe or STP capabilities of the company. We have seen how this implies all the steps in between, some of which have been discussed and we have looked at some of the requirements to achieve this extension. It becomes clear how the integration referred to in the title of this article is delivered as a direct result of the need to comply with regulatory pressure and other market drivers and that this integration occurs on two levels - both systemic and organisational.
The magnitude and cost of the change should not be underestimated! Although participants in the market do not have any choice in this matter, there are a number of alternative approaches that can be adopted to reduce or assuage the cost of integration.
Any required organisational changes, whilst complex (and in some cases painful), should be clear and reasonably obvious. They are inextricably linked with the system and infrastructure changes, which is where the largest proportion of the cost is and where future success will be determined.
It is our view that these infrastructure and system changes will be based on two generic constituents. The first is well known, the second less so:
- Application integration sometimes referred to as EAI (Enterprise Application Integration)
- Process management and workflow
It is inevitable that the vast majority of financial institutions will be using a variety of applications to support the various parts of the process from order routing through to settlement and clearance. Oftentimes this is a legacy of previous IT decisions, but increasingly it is because of the desire to use the "best of breed" application for each requirement. In any event it is highly unlikely that we will be dealing with a homogenous environment. Given the requirements of extending the STP pipe it becomes necessary to implement an environment that provides for these applications to work in a very integrated manner.
EAI is not middleware, nor is it workflow or data transformation (as found in data warehousing products). Each of these types of product offers a solution to a specific piece of the general business problem that is EAI. Rather, EAI is a combination of the technologies employed in these different kinds of product. A complete EAI solution employs the connectivity services provided in middleware products, the data transformation services provided in data warehousing or ETL (extract, transform and load) products, and the process management services provided in workflow products.
The area of process management and workflow is increasingly referred to as e-process, whose objective is to provide cross-enterprise process management. Everything is focused on that objective, and as a result some of the more sophisticated workflow facilities offered by independent workflow vendors are absent. Currently, therefore, these e-process tools will need to integrate with other software that can provide the missing functionality.
There is some overlap between EAI and e-process, and in the future there is likely to be more. However, the two currently remain relatively distinct. EAI manages the transfer of data between systems; e-process automates business processes. There still needs to be more clarity in the area of deciding which tool is suitable for specific circumstances. The ideal situation is where EAI tools become a back-end technology that runs behind e-process systems. Currently, we have a situation where technologists favour EAI and business managers favour e-process.
The extent to which you succeed in implementing the correct solution in this area will determine ultimately how successful you are and how much flexibility you have in the types of solution you implement. If you can succeed in building an infrastructure intrinsically based on process automation, management and workflow, you have not only the opportunity to determine your overall processes based on the way in which the institution likes to do business, but also to take advantage of outsourced capabilities where appropriate. That is to say, building architecture based on the participant's business processes not the generic ones that some supplier has determined should be used! This provides the additional benefit of allowing organisation structures to be created that best reflect current and future business processes. An example would be to utilise the changes in roles and responsibilities generated from the work required to extend the 'pipe' to create groups which are more customer service oriented and organised without reference to asset class.
Because of the inherent capabilities of process automation and management, the utilisation of outsourced capabilities can still be presented in a seamless manner. This will allow those institutions who have unique and marketable services, but not the necessary capital to make the investments necessary to meet all of the requirements of extending the STP pipe, to continue offering services through the judicious use of outsourced capabilities.
The views expressed in this article are those of the authors and not necessarily those of Reuters Group.
