In its latest paper, Algorithmics sets out the case for an integrated data-oriented approach to risk management, and outlines the practical steps risk IT professionals can take to achieve this. Based on Algorithmicsâ innovative risk technology and decades of experience working with financial institutions around the world, the paper provides insight into risk-specific data issues, such as the use of shared architectural foundations for marshalling data into risk analytics systems for end-to-end risk management.
The paper, âData management for risk management: the importance of data-oriented systemsâ, outlines how centralizing risk data collection facilitates the most efficient validation and normalization, enabling risk managers to do more with the data that already exists across the organization. Only such an approach makes it feasible for calculation results and stress tests to accurately reflect the interdependence between different risk types.
Neil Bartlett, Chief Technology Officer, Algorithmics, commented: âContinuing to manage data in a siloed risk environment is simply not a feasible option; using an integrated, data-oriented system is fundamental to enterprise risk management and tackles financial institutionsâ other important business needs, including the minimization of operational risk. In such a system, all calculations are driven from a centralized data source, providing consistency, timeliness, minimization of workloads, and elimination of the errors that manual processes typically produce. The benefit is less time spent on data management, and more time spent on adding value to the business across all aspects of enterprise risk management.â
In Algorithmicsâ view, a data-oriented risk management system with integrated data architecture is characterized by the ability to leverage a single data intake across multiple risk disciplines, and includes the following features that are explored in the paper:
â¢ Configurable data modeling and management: Simplified, configurable data interfaces that facilitate data capture from multiple sources, and an integrated end-to-end data workflow from data input through to risk reporting. This makes for easier stress testing by capturing the interactions and potential compounding of market and credit risks.
â¢ High levels of data accuracy: Data needs to be current, accurate, complete and traceable, and encapsulate risk domain-specific knowledge.
â¢ Embedded data transformation: Functionality that services multiple risk data consumers that may have different data requirements, such as a risk analytics engine or reporting tool.
â¢ Risk-focused, high volume data handling capabilities: Services designed to contribute to the risk management process, such as high volume data pooling and de-pooling required by ALM and regulatory capital management to reduce simulation times.
â¢ Maximized data reuse and consistency: Using outputs in one area as inputs elsewhere reduces redundant calculations. For example, automotive propagation of counterparty risk/CVA between market and credit risk in an auditable way.
â¢ Operational data services: Services designed to simplify data interaction and support operational needs across the enterprise, including support for large data volumes, as well as fault tolerance and failover mechanisms.
â¢ Robust data security and governance: Features that define what specific users can or cannot do with data.
Neil Bartlett, CTO, concluded: âLessons learned from the financial crisis have been driving firms to recognize the importance of a comprehensive and consistent approach to the data that drives their risk analytics. Data-oriented risk management facilitates consistent data modeling, quality, security, transformation, pooling and de-pooling, and workflow, making it an approach that allows risk managers to do more with the existing data that they already have.â