Few investment and pension administration applications nowadays operate entirely on their own – most need to be connected to data sources, or reporting, back office or other systems.
The internet has simplified the connection to external sources. But the piecemeal implementation and connection of internal systems can lead to inefficiency and inflexibility - especially if an organisation is trying to change and grow. Putting time and investment into the design and building of an IT infrastructure that interconnects and supports the applications can pay dividends, and even make the outsourcing of some business processes easier.
Depending on a pension fund’s size and the functions it operates in-house and those that it outsources, it could have a range of systems, from pension administration, to online trading, order management, and performance and risk management. Blue Sky Group, the pension fund of KLM Royal Dutch Airlines, has several systems including a portfolio management system from New Jersey-based Princeton Financial Systems, and has a number of external connections to data suppliers such as Bloomberg and Datastream, as well as to its brokers and its custodian Northern Trust. Most of the connections are now through the internet, says Cor van Vliet, information and communications manager at Blue Sky. This provides a single consistent communications mechanism, simplifying process of linking to the various systems.
But things can get more complicated - and less efficient - where there are multiple links between in-house systems. Very often the systems will have been put in at different times to meet urgent business needs, and will have been connected to each other on an ad hoc basis.
“What happens is that the organisation ends up with lots of point-to-point connections between different applications, with data that is pulled from some systems that are closer to real time like order management systems, and then mixed with data from overnight batch systems,” says Stephen Cheng, business development manager for the HiDataWarehouse product for London-based investment technology supplier DST International. What is more, the data is often inconsistent and in different formats. This makes it difficult to reconcile and consolidate.
There are generally three elements to an organisation’s IT set up - the input, the business applications and the output. The input can include market data, reference data and corporate actions. These are likely to come from different sources and in different formats, such as Bloomberg, Reuters and Datastream. Errors and inefficiency can arise if there are inconsistencies or incompleteness in the data, such as different identifiers for the same entity.
Business applications often have to exchange or share data - for example, an order management system might need to link to brokers and a custodian, as well as pass data through to back office and performance and risk management systems. All the links between the systems must be created and then maintained, with issues arising when systems are upgraded or changed. Furthermore, there can be problems if, say, the performance and risk systems do not get their data from the same source - this could undermine the validity of their results.
The third element is the output - reports or online analytical applications. There can be problems reconciling calculations if they are based on different source data - performance and risk is a good example again. Or if data is held in different formats, it can be difficult to bring together into reports. Or systems may update at different times, and elements of the data for reports might not be in synch.
As the problems with this kind of set-up proliferate, an organisation needs to take a step back and review its IT infrastructure, says Cheng. Many of the problems can be solved by redesigning the systems’ architecture and by making use of new technologies and technology standards, particularly messaging systems for translating and moving data around, and data warehousing for collecting, cleaning and reconciling data and providing a single consistent source of information for analysis and decision making.
A number of investment technology suppliers, such as DST International, Pennsylvania-based SunGard and London-based Misys Asset Management Systems, offer messaging and data warehousing along with their business applications, and there are a number of companies that specialise in generic messaging and data warehousing, such as North Carolina-based SAS and California-based Oracle. But the decision to tackle the infrastructure issue goes beyond realm of IT. “It is not about putting in faster machines or better applications - it implies changing the way events and processes are thought about,” says Cheng.
Instead of seeing the pension administration and investment events, processes and systems each in isolation, they should be seen within the context of an overall architecture, where the links, dependencies and overlaps are clearly visible, and an where infrastructure can be designed to support them that will ensure efficiency, flexibility and reliability.
On the input side, an organisation could implement a data warehouse that will collect all market, reference and corporate actions data, and make sure it is clean, complete and consistent, creating what is known as a “golden copy” of reference and other data, says Cheng. This will then be the reliable source for all business applications that require such data. The same approach can be taken with the business applications themselves, and the output as well, with data warehouses collecting and consolidating data for processing (managing orders and trades, etc.), and for reporting (performance and risk analysis, etc.). To link these components and the individual systems within the infrastructure requires a messaging technology that, depending on the technology used, can do much of the work of extracting, standardising and transporting the data (such tools are known in the computer industry as extract, transform and load - ETL - tools).
In addition to the advantages already mentioned of more consistent and accurate data, and the ability to consolidate data more easily, such an infrastructure can greatly reduce the complexity of a pension fund’s IT set up, making it easier to maintain, upgrade, replace or introduce individual systems. This in turn makes the IT set up more reliable and easier to scale up with growth. It can also makes outsourcing certain functions simpler, says Cheng. If, say, investment accounting or performance and risk were to be outsourced, there will now be single points of connectivity between the pension fund and the services supplier, rather than having to link multiple sources.
Outsourcing can be a way of passing much of the IT infrastructure responsibility to someone else. BNP Paribas Securities Services is in the process of developing an outsourcing service for pension funds that will relieve them of much of the effort of creating and maintaining an IT infrastructure to communicate with their investment managers and having to perform accounting and performance measurement tasks. BNP Paribas’ master custody service will link to all the investment managers that a pension fund appoints, and collect and process investment transaction data on the pension fund’s behalf, consolidating the data across the various investment managers, and undertaking performance measurement if required.
“It will take the burden of accounting and reporting of assets away from the internal mechanism of the pension fund,” says Nick Kirk, managing director of BNP Paribas Securities Services UK. “The pension fund will no longer need to have interfaces to a number of investment managers.” BNP Paribas will perform the accounting calculations and provide an automatic upload to the pension fund’s general ledger, either at a summary level or at a detailed posting level. “The pension fund won’t need more than a standard general ledger package,” says Kirk. BNP Paribas is planning to launch its service next year and it will compete with existing services from some large US custodians.
Whether to outsource functions is a business decision, but such decisions should not ignore the IT infrastructure implications. In tight markets where performance is difficult to achieve, operating costs become more of an issue. In reviewing IT costs, pensions funds should not only look at individual systems and their efficiency, but the entire IT infrastructure that supports the systems, and how this can be best rationalised and optimised. As well as reducing costs and improving operational efficiency, a well architected infrastructure can bring additional benefits such as improved reliability, greater resilience (less downtime), and easier scaling up to accommodate growth.