Algo Collateral Newsletter

In the past year, Algorithmics upgraded its entire user base to Algo Collateral v3. Based on the stability and success of this version, Algorithmics, in the course of its normal support operations, developed its next milestone release:  Version 4 (v4). Algo Collateral v4 was made available to development sponsors in the early part of 2001. It was also used as a demonstration version in April, 2001 at ISDA's Annual General Meeting in Washington. Algo Collateral v4 will be available to the general user base at the end of 3rd quarter 2001, and it is anticipated that most users will upgrade to version 4 in 2002.

Algo Collateral v4 builds on the success and stability of v3 and expands the product's core capabilities to existing modules. Taking into account user input, this new version improves on the product's operation basis and advances its collateral management functionality.

The following document demonstrates how v4 enables institutions to achieve their collateral business goals, as well as fulfill timing and budget requirements. Upgrades will be scheduled on a first-come, first-served basis. Some lead-time is required. It should be estimated that a regular upgrade exercise - without new reports and/or external modules - will require five workdays (three for an Algorithmics technical consultant and two for a business consultant). Please note that this estimate refers to an average installation (e.g. single-site, less than 200 agreements, 1 - 4 APIs). A longer lead time may be required for more complex installations.

In order to assist in quantifying the costs of adding new reports and/or external modules, Algorithmics has devised an implementation process that will formally be introduced at the User Conference in February, 2002. A tailored presentation will also be made available for individual clients. Please contact your Account Representative to discuss budgets and the business and/or technical impact of an upgrade.

The following introduction will provide detail on v4 including descriptions of the new functionality and the associated benefits institutions will receive.

• Metrics - Algo Collateral v4 features functionality to generate collateral program metrics that can be used for statistics, management reporting, and feeding downstream systems such as credit systems. Some client reporting requirements are generalized and are therefore part of the default System Metrics or Agreement Metrics. Client-specific data can be defined through the Extension table and generated via external modules.
• Account Tracking - v4 facilitates Straight-Through Processing ("STP") and provides for a vast number of additional fields for settlement instructions of a collateral transaction. These additional fields can be used to propagate S.W.I.F.T.-based messages for settlement confirmation.
• Cost Allocation - With v4, clients interested in allocating collateral costs at the transaction level may now enter cost factors.
• Additional Agreement Properties - v4 offers support for "non-dispute" band, margining on a gross basis, and more.
• Operation Controls - v4 offers additional operational checks within the workflow including alerts for insufficient collateral on margin calls and substitutions.
• Recon
• The underlying module has been re-developed for performance and other user-friendly features to assist in dispute resolutions. These improvements include processing for large trade volumes (over 5,000 transactions), many-to-many trade assignments, and the flexibility to match on tolerance parameters (for example, on the market-value of a trade). Additionally, this module now supports collateral position reconciliation.

The API Toolkit now supports both import and export capabilities for numerous tables. Algo Collateral has successfully been linked to a variety of commercial and proprietary source/destination systems including transaction, risk, limit and legal systems.

Algo Collateral continues to attract greater use as clients move to implement STP for their collateral operations. Improved support for this module facilitates promotion/reversion of the workflow based upon external events.

Collateral Optimization and Allocation
Algo Collateral now provides a second generation of its rehypothecation/re-use feature that allows for the automatic allocation of collateral to margin calls based on user-defined preferences. The allocation is defaulted from system level preferences and overwritten at the agreement level and/or the margin call level. For example, a firm may never want to re-use 'gilts' because they can make better use of them internally. This policy is then reflected as the system level default. Or, the firm may prefer to post long-dated bonds to lower transaction costs on substitutions or for operational reasons. Such a "policy" preference can be overwritten at the agreement level to accommodate the character of a particular counterparty (for example, a counterparty that is very efficient at processing substitution requests). Finally, at the margin call level, a firm will have complete control to override the previous preference to attend to a particular call's requirement. Also at this level, a firm will be able to see all available-for-delivery eligible collateral for the counterparty's margin call and then be permitted to decide the best manner to satisfy that call. To complement this feature, an institution can implement and integrate a proprietary optimization algorithm to override Algo Collateral default selection (for example, a "cheapest-to-deliver" algorithm).

Recognizing the difficulty to actively manage the composition and opportunities in the flow of a collateral program, Algo Collateral now provides Asset Management support that allows the interrogation of a collateral portfolio. With the Asset Management utility, an institution is able to create insightful views on its collateral program. This will recognize shortages in the future due to internal borrowings and for analysis of its current collateral portfolio. Such analysis can include views for: legal risk (with respect to a counterparty's domicile, jurisdiction and/or location of collateral), concentration risk, regulators (presenting the collateral portfolio based upon regulatory requirements), etc. This utility can be used to complement policy where a repo desk is lending or borrowing securities from the collateral unit by identifying positions held/posted that are current 'specials' in the repo market and that can be substituted and more optimally deployed.

Algorithmics recognizes an institution's concerns for greater control over data changes and has introduced a "4-Eyes Principle" as an optional data entry mode. Under this mode, two data table sets are used to control data entry (three sets, including audit), one table set for 'active' data and a second table set for 'edits.' With the 4-Eyes Principle, data changes are entered into the 'edits' table set and held until approved or removed upon cancellation. Changes held in the 'edits' table set can be viewed and organized for the approval process.

If the changes are approved, the 'edits' table set values are used to update the 'active' table set. All processing in this data entry mode is based on data held in the 'active' table set.

If a firm, or one of its counterparties, is rumored in the market to be downgraded (or upgraded), the collateral management department faces the question, "What effect will this have on our collateral balances and liquidity needs?" With Algo Collateral's initial Analytics module tool, What-If Scenario, users are able to create different credit rating scenarios to investigate the effect of changes in the firm's and/or counterparties' rating.

As with all upgrades, Algo Collateral provides an automated process using SQL scripts to modify the database's table and data definitions and then migrate the data into this new structure. Algorithmics has enhanced this procedure based on previous upgrade experiences. It is the responsibility of the institution to make changes to data and/or procedures to maximize the benefits of Algo Collateral's enhanced abilities.