By Neil McGovern, product strategy director at Sybase History may judge the failure of Lehman Brothers as the event that caused the most changes in financial services institutions in the credit crunch of 2007 to 2009. The failure of an institution that was never meant to fail, especially one such as Lehman, has changed attitudes to counterparty risk. Lehman Brothers seemed especially unlikely to fail, because as one of the Wall Street giants, it was thought to have the risk management systems and balance sheet management practices to weather any storm.
Lehman underwrote huge quantities of credit default swaps (CDSs), leaving many of its counterparties uninformed of the liabilities they would face under certain likely scenarios. Even for those organisations that understood their exposures to the Lehman paper, the lack of liquidity made it difficult to value or price their portfolios, let alone measure the risk associated with them. The Counterparty Risk Management Policy Group (CRMPG) recommends that counterparty risk across asset classes and all counterparties be calculated in hours. With some very high profile organisations coming under pressure and the historically high price of volatility, many organisations are looking to perform intraday risk assessment to ensure they can react more quickly and effectively. In this article, we will identify some of the major technical challenges and explore the available technologies that can help financial services firms improve counterparty risk management. Bringing All the Data Together Most institutions do not have an enterprise view of risk, primarily due to the required data being spread across many different systems. The major divides in the data are: • Business function –front office versus middle and back office • Asset class – portfolios can span multiple asset classes • Geography – for example, New York versus London Data Volume and Velocity The last few years in capital markets have seen exponential growth in the volume and velocity of market data. Since any near real-time counterparty risk analysis system is going to have to monitor this data as it streams into the organisation, new methods and technologies that enable faster event identification and data analysis are required. Model Complexity Counterparty netting agreements are cross asset, so there will be multiple models that comprise the complete risk analysis. These models need to be portfolio-based rather than product-based, and are required to generate expected exposure and match front office pricing. Market Analytics Oriented Data Store The top performing risk analytics engines available today are built on a column-based approach to data storage (as opposed to a row-based architecture common in the top relational databases in the market). While this is not a new technology, the key advantages are just being appreciated, as can be seen from some high profile start-ups in this area, as well as the dramatic increase in shipments from the existing vendors. The appeal is clear: many customers are reporting analytic rates 100 to 1,000 times faster than previously seen. One top financial institution’s risk modelling procedure, which originally required more than five hours, was reduced to 30 seconds when they migrated to an analytics-based solution from a transaction oriented product. This is an order of magnitude advantage for this organisation, allowing them to identify potential crisis quickly and react before the market and their competition. Compressing Market Data The latest market data stores not only store the data in a different architecture, but also use compression techniques when loading the data to ensure that it is stored in 20-30% of the space. Thus, one megabyte of incoming market data is compressed to 0.2 megabytes instead of expanded to 2 to 3 megabytes in traditional architectures. Again this is an order of magnitude advantage. Identifying Market Significant Events in Real-time Given the complexity and resource intensive requirements of intraday or near real-time counterparty risk analysis, an intelligent screening mechanism that can identify key events in the market can significantly improve an institution’s counterparty risk analysis. These technologies, commonly called complex event processing (CEP), are becoming extensively used in trading systems, making the crossover to risk analysis more straightforward. Unstructured Data Analysis With the introduction of tagged news from Dow Jones (Dow Jones Elementised News Feed) and Reuters (Reuters Instrument Codes), the ability to scan and analyse news feeds in real time and determine the impact the news may have on counterparty risk is now a possibility. The importance of understanding and measuring counterparty risk is a top issue for capital markets firms in 2009. Many organisations need to change their counterparty risk methodology, moving to more complex portfolio oriented models that are calculated more frequently throughout the day. The technical challenges are exacerbated by increased volume, velocity and volatility, requiring a new generation of technology designed for real-time analysis to take over from traditional systems that cannot scale to the levels required. Fortunately, new technologies are being introduced and it is essential that risk managers review and understand how these technologies can assist their business goals.