Wall Street & Technology is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.


08:11 AM
Cristina McEachern
Cristina McEachern
Connect Directly

Reconciling Reference Data

Enhancing security and improving counter-party identification are the keys to cleaning up reference data in 2003.

Enhancing security and improving counter-party identification are the keys to cleaning up reference data in 2003.

Reference data - static information such as data on a security-master file, information on counter-parties, data in account-master files, data on customers and clients, as well as data on currency and country codes - has long been stored in multiple, disparate databases across the enterprise.

Now, as automation becomes a priority, financial-services firms are finding they must standardize the reference data pouring into their shops if straight-through processing is ever to become a reality.

Security numbering and security identification, as well as counter-party identification, are the main areas of focus that financial-services firms should be concentrating on in the coming year, says Rich Robinson, assistant vice president of global equities IT at Deutsche Bank AG.

"Reference data and securities numbering are key to straight-through processing and if you don't have proper security identification, it's going to fail STP in your own system and possibly fail in the marketplace if you're identifying wrong," he says.

Robinson says the problem with security identification is the multiple numbers currently being used to identify securities - including ISIN numbers, CEDOL, CUSIP, local-clearing numbers, etc.

He points out that with global expansion in the financial-services industry, the numbering systems have compounded even more. "As securities are multi-listed across different marketplaces, it's more critical that they can uniquely identify what instrument they're pricing, reporting on, etc," says Robinson. He takes the example of Daimler Chrysler stock, which is listed on both German and U.S. exchanges.

While they use the same ISIN number, the stock is priced differently depending on the exchange, either in U.S. dollars or in Euros. "If you hold the U.S. issue, it won't have the perfect price correlation to the German issue and, with currency-fluctuation risk, you would have different valuations," he explains. "It's the same instrument, but you can't readily buy one and sell it the next day on the German market."

The Financial Information Services Division of the Software and Information Industry Association has tapped industry players, such as Robinson, and other data devotees to create a working group that will examine the main problems surrounding security identification.

"It would be nice to have one code to satisfy everything up and down the trade lifecycle, but I don't think it's going to happen in the next year or two," says Robinson. "I do see proper cross-referencing being done for the different kinds of codes for different purposes." By that he explains that cross-referencing would standardize the codes to ensure that they match up across the board, so even when different number systems are used, the meaning would be the same.

It would, however, require help from standards bodies and the number issuers in order to standardize the numbering scheme, not necessarily the numbers themselves.

Robinson says one important task for firms is to look closely at their securities-master databases, if they have not already.

More specifically, he says that firms should take a look at the primary keys and main drivers of their database - for example, ISIN or CEDOL, and what relationships they can set up between different numbers for this "cross referencing."

He adds that firms should be looking at what is already being done internally because there is no point in reinventing the wheel. Firms should not be coming up with their own solutions, like new numbering schemes, or their own vendor-specific codes to use.

But there is an opportunity for vendors to come in and provide value-added solutions to help certain firms. He adds that any solution has to be flexible and will be dependent on the firm's size and focus of activity.

Counter-party Identifications
Another important aspect of reference-data work will be done in the area of counter-party identification over the next year, says Robinson. "There are a lot of identifiers out there like Crest IDs, BIC, etc, but there hasn't really been any major effort to cross reference all of that information to make sure (for example) this BIC equals that ID and so forth," he says.

For this area as well, he envisions some sort of cross-referencing standardization around counter-party identification. He points out that firms have many other things they are working on to solve the reference-data problem, but that these other initiatives could help with the counter-party identification drive.

"Right now, everybody has a lot on their plate and now we have to finish things like security numbering before we tackle the others," Robinson says. "But it could also serve as a template for counter-party identification, the structure that could be applied and the way it is approached."

When it comes down to transmitting reference data, Robinson says that it is critical to embrace standards and get away from vendor-specific formats. "Standards lower the cost for the whole industry and help vendors by letting them compete on more important, value-added services as opposed to having to maintain their own formats and transmission standards," says Robinson.


The Cost of Reference-Data Management
Over 80 percent of respondents to a TowerGroup survey reported they are considering the development of new systems to help standardize/synchronize reference data across the enterprise. But finding the funds to do so may be easier said than done.

Almost 40 percent of respondents indicated that project plans for developing and funding data-management projects had been approved, with 32 percent saying the 2003 budget cycle would include funding for reference-data projects and 23 percent saying the funding would come in 2004.

TowerGroup estimates that projects to synchronize and standardize enterprise-wide reference data would be about $4.2 million, including technology and operational processes.

The pricetag for these types of data-management projects is broken down further by firm type: custodians surveyed say their cost would exceed $10 million, brokers say the average cost would be around $5 million, but over 20 percent of them estimated that number could also exceed $10 million, and buy-side firms estimate an average cost of $2.4 million for these projects, with 10 percent of them expecting to pay more than $10 million.

For projects of this scope and cost, return on investment, in terms of key benefits that will result from this type of investment, is vital.

TowerGroup found that the main focus continues to be improving automation rates and reducing manual processing. In addition, firms see key benefits in the following areas: improving data consistency across applications, centralizing data management, reducing reconciliation expenses, reducing redundancy of systems, deploying data standards across the enterprise, improving response-time-to-market opportunities, integrating more sources of data and accelerating application development.


Advice on Cleaning up Reference Data
John Bottega - director of global markets and investment banking, enterprise data standards, Merrill Lynch - advises that firms focus on four main components in their attempts to achieve clean reference data: information architecture, technology, process and organization.

The information-architecture layer should include standards covering common names, values and formats that will enable various databases to communicate and share information. In addition, the information architecture should include a data dictionary with common definitions of data elements and common symbology for the consistent identification of data.

For the technology component, Bottega says that as vendor feeds come in, it's vital that feeds are parsed to a staging area and rules engine, and then are loaded to "golden copy" for use across the enterprise. Rounding out the technology component should be workflow tools, a database-management system and query layer.

For the process component, Bottega says that centralized workflow and error correction is essential for consistent and complete reference data. All of these efforts should be supported by the organization for total success of the reference-data program. Overall, reference-data strategies should be a joint effort between business and technology personnel and departments as a top down and bottom up initiative.

Register for Wall Street & Technology Newsletters
Exclusive: Inside the GETCO Execution Services Trading Floor
Exclusive: Inside the GETCO Execution Services Trading Floor
Advanced Trading takes you on an exclusive tour of the New York trading floor of GETCO Execution Services, the solutions arm of GETCO.