Ness and a Leading Market-Infrastructure Provider Partner to Develop FRTB Data Service

Case Study

Ness and a Leading Market-Infrastructure Provider Partner to Develop FRTB Data Service

Executive Summary

A post-trade infrastructure provider for the financial industry embarked on a plan to provide a new value-added service for FRTB regulations. The real-price observation service would enable the provider’s customers to optimize balance-sheet capital while demonstrating the modellability of risk factors and allowing positions in their trading book to be capitalized using internal models. The infrastructure provider asked Ness to partner with them to co-develop a front-to-back service solution using scalable aggregation and an agile, DevOps-first approach.

About The Client

The client is a premier provider of post-trade market infrastructure for the financial industry. It enables the automation, centralization, and standardization of financial-transaction processing to mitigate risk, increase transparency, and drive efficiencies across market participants.

The Challenge

The Fundamental Review of the Trading Book (FRTB), released by the Basel Committee on Banking Supervision (BCBS), updates the minimum capital requirements for market risk to address shortcomings of the current Basel III market risk capital framework. FRTB requires banks to evidence enough market liquidity for the positions in their trading book that are capitalized using an approved internal model approach (IMA). Without empirical trade data and tools to interact with it, banks struggle to understand the impact of IMA on capital charges and its quantitative benefits. Failure to secure robust, high-quality price observation data to assist with non-modellable risk factor (NMRF) analysis can subject firms to significant capital inefficiencies, which could threaten the viability of trading in certain asset classes and the withdrawal of liquidity from the market. The client is uniquely positioned to assist the industry with its FRTB data needs for OTC derivatives and cash instruments to test modellability of illiquid instrument classes.

Why The Client Chose Ness

The client previously engaged Ness because of the firm’s deep risk domain knowledge, its commitment to innovation, and its capacity to meet timelines with high quality and predictability. Ness’s impressive delivery track record with the client, combined with a forward-thinking, value-driven approach, positioned the firm as the partner of choice for the implementation.

The Solution

Ness worked with the client’s business and technology leadership to design and implement a cloud-based scalable platform in AWS for real price observations. One of the key features of the solution was the de-duplication of 1.2 billion transactions across all asset classes. The data-transformation layer processed 18 months of anonymized transaction data including validation of lifecycle events. This data was then made available for consumption by the data-aggregation layer.

ActivePivot™ by ActiveViam, an in-memory OLAP aggregation engine, was used as the aggregation layer to process 175 million transactions. Ness’s subject-matter experts built a platform for multi-dimensional analysis with an average query response time of 5 seconds across more than 150 concurrent users. The platform allowed for business rules–based filtering and drill-down capabilities, allowing regulators to analyze observations results across asset classes. Alerting capability was built into the solution to notify of risk factor eligibility changes.

Results and Benefits

Ness enabled the client to deliver a scalable service that measures modellability and capital impacts, thereby allowing firms to maximize risk capital charge efficiencies. The solution was built with extendibility in mind, which allowed the client to migrate other legacy platforms to the new architecture and provided new revenue-generating opportunities. In addition, the client is also able to conduct new ideation workshops using AI/ML to provide advanced analytics around pricing models and PnL attribution.