Capital model benchmarking automation using technology and data models
At RiskMinds International in November, Olivier Miart, Head of ISDA Analytics, ISDA, will be presenting his work on Capital model benchmarking automation using technology and data models. We caught up with him before the event, the get some insights on what you can expect in Barcelona.
Background
The final parts of the Basel III framework, including the Fundamental Review of the Trading Book (FRTB) and revisions to the credit valuation adjustment (CVA) framework are due for global implementation by January 1, 2023. One of the defining features of these reforms will be far greater use of standardized approaches to the calculation of capital.
Given the planned increase in the use of internal models, the complexity of standardized approaches has been significantly increased so that firms using the standardized approach will no longer be using basic formulae to determine their capital requirements. To avoid the standardized approach becoming a ‘blunt instrument’ that applies a one-size-fits-all approach, the complexity and sensitivity of the new standardized approach is far greater than in previous iterations of the framework.
The transition to the new standardized approach is taking place over a fairly short time period and is a major undertaking. Ensuring implementation is both accurate and consistent around the world is even more challenging. Banks need to interpret the rules to implement the new, more complex standardized model, but they need to do so in a way that is consistent with their peers around the world. Regulators also want to see consistent implementation so that they can accurately compare banks with one another and ensure sufficient capital is being set aside. Banks therefore need a way to benchmark their interpretation and implementation against an industry standard.
ISDA SA Benchmarking
In 2018, ISDA launched the ISDA SA Benchmarking initiative to enable banks to implement standardized approaches accurately and consistently. Starting with a pilot group of 15 UK banks with the support of the Bank of England, ISDA quickly showed the unique value and accuracy of its benchmarking. The initiative is now in its fourth phase and 70 banks around the world have participated to-date, and 17 regulators have used the data to monitor implementation in their jurisdictions and better understand the drivers of any divergence.
ISDA’s Benchmarking programme is underpinned by PerunTM, an ISDA in-house quantitative analysis tool that combines with ISDA’s Common Risk Interchange Format (CRIFTM) to automate much of the analysis and deliver detailed, customized reports to banks and regulators. The benchmarking process includes the hypothetical portfolio exercise which is a detailed, time-consuming process in which banks carry out the end-to-end capital calculation with a set of hypothetical trades and then submit them to ISDA for analysis. A final report is prepared at the end of this process to present the results and explain any variance in capital results against the median benchmark capital for each portfolio. Perun is critical to enable and automate the validation and processing of big data sets submitted by participating firms, as well as the generation of results explaining the sources of divergence in resulting capital requirements.
In recognition of the success of the ISDA SA Benchmarking and its Perun technology platform, ISDA won the Risk.net Innovation in Technology 2022 award.
ISDA's role
The benchmarking initiative has evolved quickly to a global service that is now underpinning the effective implementation of Basel III. The initiative is into its fifth year and ISDA has provided analysis of standardized approaches to a level of detail and accuracy that no other institution, association or regulatory agency has achieved before. This has been made possible by ISDA’s close collaboration with market participants and their willingness to submit confidential data to the benchmarking exercise, coupled with the highly advanced analytics capabilities of its in-house benchmarking technology platform, Perun.
Perun leverages existing standards and enables accurate, automated analysis of complex risk data. Perun was developed from scratch to support the ISDA SA Benchmarking initiative and enable banks and regulators to properly understand the drivers of any divergence in capital requirements. Perun is coded in Python and allows users to zoom in on specific data sets for relevant subsets of banks, portfolio or risk types, helping to identify patterns and trends across the results and to quickly categorize the sources of divergence. Perun also calculates standardised capital based on firms’ CRIF inputs, and allows for easy extraction of charts to populate the reports that are issued as part of the benchmarking process.
Standardized risk data reporting is also critical to the success of benchmarking and this is made possible by ISDA’s CRIF. Use of the CRIF started with the ISDA SIMM and has been extended to support ISDA SA Benchmarking, ensuring that data is submitted to Perun in a standardized format. In the future, the ISDA Common Domain Model initiative (CDM) is expected to accelerate the automation of CRIF file generation by firms and facilitate digital regulatory reporting.
Over the past five years, ISDA has developed a highly advanced analytics function within its risk and capital division. Leveraging these quantitative resources through CRIF and Perun has enabled ISDA to develop a highly unique and valuable offering to support the effective implementation of the final Basel III reforms across the industry. Regulators and market participants around the world have recognized the benefits of the benchmarking initiative and the unique value of the analysis it delivers.
Additional Resources
- Video explanation of PerunTM: https://www.youtube.com/watch?v=BBWgVba8RfY&t=5s
