EBA report on big data and advanced analytics

European Banking Agency ("EBA") publishes its report on big data and advanced analytics

The EBA has this month published a 'deep dive' review on the use of big data and advanced analytics ("BD&AA") in the banking sector. The aim of the EBA review is to share knowledge on the current use of BD&AA and to set out the key pillars and elements of trust that should accompany their use.

Observations on current use of BD&AA

Observing that:

BD&AA are driving fundamental change in institutions' business models and processes. Currently, BD&AA are part of most institutions' digital transformation programmes, along with the growing use of cloud services

The EBA nonetheless noted the following constraints on the use of BD&AA:

• Core banking data are currently the main flow feeding data analytics, rather than other data sources such as external social media data. This stems largely from institutions' concerns about the reliability and accuracy of external data.
• A further key constraint is the integration of BD&AA into existing business processes, and the need to develop relevant knowledge, skills and expertise. Firms are relying extensively on outsourced expertise.
• Institutions appear to be at an early stage of use of machine learning, with a focus on predictive analytics that rely on simple models. Although more complex models can bring better accuracy and performance explainability and interpretability issues are a concern for institutions.
• Institutions currently leverage BD&AA mainly for customer engagement and process optimisation purposes (including RegTech), with a growing interest in the area of risk management.

This report then identifies four key pillars for the development, implementation and adoption of BD&AA as follows:

• Data management arrangements providing for the control and security of data.
• Technological infrastructure in terms of data platforms and infrastructure that provide the necessary support to process and run BD&AA.
• Appropriate organisation and governance, and appropriate internal skills and knowledge, to support the responsible use of BD&AA across institutions and ensure robust oversight of their use.
• Analytics methodology to facilitate the development, implementation and adoption of advanced analytics solutions.

Trust Issues

The report notes that the roll-out of BD&AA specifically affects issues around trustworthiness which cut across the four key pillars covering questions of:

• ethics;
• explainability and interpretability;
• fairness and avoidance of bias or discrimination;
• traceability and auditability;
• data quality and protection and data security; and
• customer protection.

The EBA has noted that further work was needed on such issues to ensure responsible use of BD&AA, and its detailed comments on these topics draw heavily on the EU report on Trustworthy AI explained in our earlier briefings, (click here for links to these) and demonstrates a fairly consistent approach.

Next Steps

The EBA intends to continue monitoring developments. Where appropriate, it will perform additional work to enhance supervisory consistency and facilitate supervisory coordination across EU and international agencies.