With the Fundamental Review of the Trading Book (FRTB) planned for phased rollout beginning in 2022, firms that can master the ability to apply accurate and relevant market data could unlock future business benefits under the new regime and strengthen the value-at-risk (VaR) framework for capital adequacy.

The new approach utilizes an alternative method of measuring market risk using expected shortfall (ES) as the primary exposure measure instead of VaR, while also ensuring alignment of the front office desk models with market risk calculations for their internal models.

FRTB represents the next generation of market risk regulatory capital rules for large, international financial institutions. It has also encouraged vendor solutions to develop alternative offerings that can help firms implement new risk management capabilities.

Among the many ambiguities of implementing the new regulation, internal modelling of market risk exposure is a well-known challenge across the industry. Without access to useful data, non-modellable risk factors (NMRFs) can result in suboptimal capital alignment with the underlying risks, which can undermine the viability of desk level internal models. In markets where data is scarce, sourcing quality market data will play a pivotal role in the difficulty of mitigating against non-modellable risk.

Firms should conduct a current-state assessment of market data capabilities and infrastructure to design and implement a robust market data solution. This requires a thorough understanding of key requirements, technology architecture and, operational effectiveness. From a current state perspective, firms must consider whether internally stored data satisfies the Risk Factor Eligibility Test (RFET), as well as stressed expected shortfall requirements.

To address the prioritized requirements, each asset class (e.g., credit) needs to be assessed in its component products (e.g., corporate bonds) and then sub-products (e.g., high yield versus investment grade), which are mapped to risk factors. Additionally, a risk factor can be broken down into a more granular level of risk drivers (points on a curve/surface). These buckets provide a mapping schema for RFET observations to be used as criteria in the sourcing decision.

Banks can choose to address data gaps wholesale by designating a centralized source that accommodates FRTB alongside other market and credit risk data requirements. Ultimately, a firm can benefit significantly by investing in a centralized market data source that aggregates data from internal systems of record alongside external vendor data. This approach’s benefits include centralized governance and oversight structures to promote data quality and efficient delivery to multiple systems, models, and users. Specialized ‘risk security masters’ can be a benefit to firms, serving as golden sources of data with 24/7 data availability, enhanced data sets, and calculations.

Enhancing market data internally

For firms implementing the new regulation, improving a firm’s FRTB pricing observations strategy is a key concern.

The key drivers for data-related change in the RFET are twofold. Firstly, the new requirements have a stricter criterion for what constitutes a real price − banks can no longer rely on price observations such as trader marks. Secondly, FRTB prescribes a strict bucketing criterion for price observations so firms must ensure that they have the appropriate depth and granularity of data for relevant price observations. Banks can review coverage with a risk factor catalog that has all risk factors from pricing models mapped to products and real prices. Firms are likely to find that their quarterly VaR completeness exercise can be a good starting point to achieve the risk factor catalog.

Accessing market data externally

To establish a vendor as an ongoing data provider which meets the data requirements for FRTB,  a formal RFI/RFP approach is recommended where the companies are compared for service, data quality, and pricing.

For firms that are considering collaboration with a vendor or multiple vendors to address gaps, there are several factors to consider:

  • Availability: Firms must assess whether the vendor can supply the required data. For this exercise, it may be helpful to utilize the product to sub-product to risk factor mapping prepared as part of the current state assessment.
  • Existing vs. new package: The required data may be a net new buy or a feed that the firm gets currently through an existing package, but it is not fully utilizing.
  • Technology: If vendors can provide data through existing technology connections or data feeds, it may be possible to avoid or reduce costs associated with technology builds.
  • Cost: For smaller firms, it may be more cost-effective to commission a team to create internal market data rather than buying the data continuously from a vendor.
  • Best of breed vs. single source: Firms may consider one or multiple vendors for an analytics platform that supports FRTB components such as PLA or modellability and choose to supplement that platform with niche data from other specialized vendors.

Firms can also expect improvements in overall efficiency due to technology upgrades and cost savings resulting from redundancy reductions. Tracking these benefits over time can help justify spend and create a strong case for a firm to invest in a robust market data solution.

It’s essential for an effective business decision to understand both the one-time and ongoing costs associated with building out the market data function for FRTB. One-time costs include documenting requirements and scope, customizing the platform, establishing vendor data feeds and STP for business processes as well as one-time purchases of historical data to backfill missing time series. Ongoing costs include data purchases to satisfy change requests, subscription costs (monthly) to data storage and ongoing data governance/quality control functions.

The selection of risk market data solutions often requires additional, customized and expert-driven assessments to get into the detail and solve specific scenarios.

Considering the problems holistically, as a comprehensive data operating vision, the needs for the data span from risk mitigation through to cost control, increased governance, and ongoing accuracy with suitable availability, these multi-directional business cases require a careful and detail-oriented approach. To make a vision become a reality, practical advice and support are required to enable change.

Was it worth reading? Let us know.