#2: The Evolution of Payments Data Analytics and Alternative Data for Supervisory Purposes

Suptech Broadcast

29 Sept

3rd November 2022

With:

| Perttu Korhonen (Qatar Financial Centre Regulatory Authority)

| Clara Machado (Banco de la República)

In Session #2, Perttu Korhonen (QFCRA) and Clara Machado (Banco de la República) join FNA to explore the operationalization of payments and alternative data. As financial systems generate increasingly massive datasets, the challenge for supervisors has moved from "collecting information" to "extracting actionable signals."

The discussion highlights the transition from traditional, lagging econometric models to augmented decision-making tools. By utilizing network analysis to identify "super-spreaders" of liquidity and applying machine learning to detect anomalies in high-value payment systems, regulators can now conduct RTGS Stress Testing in real-time. Our guests provide practical case studies on how these advanced analytics allow supervisors to spend less time on manual data management and more time evaluating the systemic paths that could lead to financial instability.

Key Discussion Points:

The "Augmented" Supervisor: Why the future of oversight relies on a symbiotic relationship where AI provides the recommendations and humans provide the reasoning and judgment.
Network Science as a Risk Map: How metrics like hub and authority centrality are used to identify "too-connected-to-fail" institutions in the interbank market.
RTGS Stress Testing: Moving beyond static scenarios to simulate participant failures and liquidity shocks within the payment infrastructure to test system-wide resilience.
Alternative Data Integration: Exploring the use of granular, non-traditional datasets—such as property market indicators or micro-level loan data—to build more comprehensive risk models.
The Skills Evolution: The critical need for supervisors to move beyond Excel and develop proficiencies in Python, data science, and visual analytics to handle "Generation 4" Suptech.
Anomaly Detection Frameworks: Using both supervised and unsupervised learning to flag behavioral deviations in payment flows that could signal operational outages or financial stress.
Data Strategy: Best practices for building an analytical data model that ensures consistency, cohesiveness, and ease of exploration across different supervisory departments.