In “An Innovation Lens to Risk Data Aggregation and Risk Reporting,” Avinash Singh and Celia Wanderley highlight how banks can tackle persistent RDARR compliance challenges by leveraging GenAI. This transformative approach not only enhances data quality and governance but accelerates compliance, offering banks a faster path to meet regulatory demands and improve risk management.
Background
One of the key revelations arising from the Global Financial Crisis in 2008 was the inadequacy of bank technology infrastructure and data architecture to support timely risk reporting and hence effective risk decision making. In response the Basel Committee on Banking Supervision published BCBS 238 – Principles of Effective Risk Data Aggregation and Risk Reporting (RDARR) in 2013. Despite the implementation of the principles being underway for approximately 10 years, a number of Global Systemically Important Banks (G-SIBS) are struggling to meet the requirements for full compliance as reported by the Basel Committee on Banking Supervision (BCBS) in their latest progress update. If there was ever any doubt about the importance of timely risk data for decision making, recent stress events such as the pandemic have reinforced its importance particularly reliable, granular, high frequency data, ad-hoc reporting and new regulatory requests. Due to the slow progress underway to achieving full compliance regulatory authorities are being encouraged by global standard setters to ratchet up intensity of supervision including measures to enforce compliance such as capital add-ons, restrictions on capital distributions and business activities as well as fines and penalties.
Results from 2022 progress update
The assessment of 31 G-SIBS were conducted using a 4 point scale across the 11 principles. The rating scale was as follows: Rating 4 – fully compliant, Rating 3- largely compliant, Rating 2 – materially non-compliant, Rating 4 – not adopted.
Source – https://www.bis.org/bcbs/publ/
The results reveal that whilst the majority of banks have achieved a rating of largely compliant and fully compliant across most of the principles, a significant proportion remain materially non-compliant. The aggregated rating trend across principles is also very insightful, revealing a deterioration in the trend across 6 principles during the progress update periods between 2019 and 2022. These include Principles 1 (governance), Principle 5 (timeliness), 7 (accuracy of risk reports while aggregated ratings for Principles 4 (completeness), principles 6 (adaptability) and principle 9 (clarity) have remained largely stagnant.
Key insights from the results
- The overall pace of implementation by banks is occurring at a much slower pace than envisaged driven largely by complex data architecture, legacy IT environments and manual processes
- Banks still lack a common data taxonomy, data lineage which impacts harmonization efforts and the ability to detect data defects
- While new technologies like AI hold great promise, banks are yet to benefit from them in RDARR, largely due to data quality which is a critical prerequisite
- There is insufficient oversight and ownership from Bank board and senior management on data quality issues and therefore allocation of adequate resources, budgets and accountability for delivery of RDARR initiatives
- Data management is not a once-off event and requires continuous investment and improvement to maintain agility with changing business strategies, operating models, external stress events (e.g. the pandemic is a case in point), new technologies and regulatory requirements.
Regulator recommendations
- Bank boards and senior management to intensify oversight and accountability for data governance and its implementation which includes RDARR
- Ownership for data quality should be implemented enterprise-wide with clear roles and responsibilities across the three lines of defence
- New technologies capable of digitizing and automating data and accompanying workflows must be preceded by a focus on quality. Data lakes also need to be scalable and agile, capable of adapting to environmental changes and needs of stakeholders
- The RDARR principles should be rolled out and implemented more broadly in organizations covering all risk data, financial and regulatory reporting in all business units, legal entities and jurisdictions using a risk-based approach and not limited to material risk types only.
Accelerating the Path Forward with GenAI
Addressing the regulator recommendations effectively, however, remains an ongoing challenge. We understand that fragmented data repositories across businesses, legal entities and countries, lead to duplication of data and make it harder to meet different regulatory requirements. IT infrastructure needs to be agile to meet ad-hoc requirements (regulator and or market disclosure) and new and emerging risks. Key challenges have included vendor technology complexity, coordination of work effort across organizational boundaries, depth of data requirements and specialized skills.
To address the challenges articulated, it has long been understood that banks must modernize legacy technology, implement robust data governance frameworks, improve data quality, and establish clear data lineage. These foundational elements are critical to ensuring compliance with regulatory standards such as RDARR, while also laying the groundwork for enhanced risk management capabilities.
It is also well understood and accepted that fully leveraging advanced analytics and artificial intelligence (AI) to achieve business outcomes also requires strong data architecture, infrastructure, engineering and all those same foundational elements firmly in place. That often leads banks to a perception of linear progression: first, solve the data infrastructure and engineering challenges, and only then will we be equipped to start using AI.
But what if we could turn this thinking on its head? What if we could use emerging technology, particularly Generative AI, to significantly accelerate the process of building those very same data foundations? What if doing that could allow banks to achieve two to ten times the speed in improving data governance, data quality, and data lineage? Now AI becomes not just a tool to be used at the end state but also a tool that allows us to get to the end state in a faster, more automated, more cost-efficient manner.
GenAI offers an opportunity to shift the paradigm. By automating various aspects of data engineering, from detecting and correcting data anomalies to acquiring missing data and harmonizing data sources, these technologies can enable banks to overcome barriers that have traditionally slowed down progress. The volume, velocity, and importance of data is only going to grow, and organizations must address these challenges faster than ever to manage risk, remain competitive and comply with regulations.
The opportunity at hand is profound. By using technology that is now readily available, banks can meet the regulatory and operational demands of data quality and governance at a pace that was once unthinkable because most of the work had to be done manually and solely depended on the precious time of those same subject matter experts who are responsible for keeping the lights on and continuing to operate and grow the business. Generative AI, coupled with predictive machine learning (ML) capabilities, can streamline and even automate a portion of the manual work associated with managing vast and fragmented data ecosystems.
We are seeing significant interest from clients across industries in this approach, particularly because it enables them to manage data risks more effectively while also unlocking new business opportunities. For banks, where data management has long been an impediment to innovation, the potential to leapfrog these challenges through advanced AI solutions is especially compelling. This is not only about compliance but also about positioning banks for future growth, resilience, and market leadership.
How can Ulysses Partners & Bits in Glass help banks comply with expectations of RDARR
- Data Readiness Assessment for RDARR – we can conduct a comprehensive assessment of data governance against regulatory expectations to identify gaps that need to be closed as it relates to data risk frameworks; policies, standards, procedures; data management operating model including roles and responsibilities across the 3 lines of defence and reporting templates and pathways from operating committees to the executive and board level.
- Data Foundations Accelerator – leveraging the results of the Data Readiness Assessment for RDARR, we can use GenAI to harvest metadata about your data and implement GenAI-powered data governance solutions. This will not only assist with point-in-time data lineage and governance implementation but will also help you keep the information updated in a semi-automated fashion, giving data owners and stewards the opportunity to review, refine and confirm but providing them with first drafts and automatically mapping out lineage. Leveraging leading edge technologies like GenAI and ML to identify data quality challenges can improve accuracy and completeness and allow for timely aggregation. Automating detective and preventive controls to identify data errors and anomalies, through reconciliation and variance analysis of material risk types and metrics can increase compliance.
- Independent Validation and Remediation Management – we can implement automated workflows to enable RDARR initiative owners to trigger independent validation from 2nd and 3rd line of defence teams, tracking remediation activities and maintaining an audit trail throughout. This can then be integrated with the overarching banking regulatory reporting.
Ultimately, addressing the key pain points of data architecture, manual processes, common data lineage, common data taxonomy and data quality does require stronger governance and prioritized effort. Improving compliance in complex environments like banks can only be achieved using technology and automation to supplement the manual effort from subject matter experts. The good news is that the developments in technology, particularly in the last few years, has become our greatest ally in getting there.
About the Authors:
Avinash Singh | Risk and Compliance Partner | Ulysses Partners
Celia Wanderley | Chief Innovation Officer | Bits In Glass
