How The SNB Improved Payment System Stability With AI

A Stable Payment System for Switzerland’s Financial Sector

As the economic environment becomes increasingly complex, ensuring the stability of a country’s payment system is a critical aspect of maintaining financial integrity. This was the challenge faced by the Swiss National Bank (SNB) – to ensure the stability of the Swiss Interbank Clearing (SIC) payment system by detecting potential payment difficulties in key participants. The approach required a system that could effectively and efficiently analyze interbank transactions on a large scale.

Harnessing the Power of AI: The Solution

The answer to this challenge was realized in the form of a highly customized, AI-powered monitoring tool developed in a collaborative project with swissQuant. This innovative tool goes beyond simple data analysis – it offers the power of nowcasting. With nowcasting, the tool uses real-time data to make accurate predictions and assessments of the current financial landscape. Utilizing a five-step machine learning (ML) methodology, the tool could handle vast data sets, identify outliers, and assess the stability of network participants.

From Big Data to Actionable Insights: The Methodology

swissQuant’s sophisticated ML models transform large, unwieldy data sets into clear, actionable insights. Their innovative approach involved a layered ML model with a five-step process:

Robust Data Processing: Raw data was aggregated and adjusted for trends and seasons, resulting in a set of ‘stress event’ labels.

Outlier Identification: A heterogeneous outlier ensemble ML model was used to calculate individual outlier scores.

Predictive Performance: With the outlier score data, supervised ML models predicted the likelihood of bank stress events.

Evaluation: The models’ performance was assessed through the application of ROC curve metrics, qualitative analysis, and visualization techniques.

Insights to Actions: The final step involved transforming these insights into actionable strategies.

A Win for Data-Driven Decision Making

With the application of swissQuant’s methodology, the SNB achieved a high degree of accuracy in nowcasting stress events for banks within the SIC payment system. This breakthrough method was captured in an academic paper jointly published by swissQuant and the SNB, and the findings were presented at scientific conferences in Switzerland and beyond.

In Conclusion: Harnessing the Power of AI for Stability

The successful collaboration between the SNB and swissQuant is a testament to the transformative power of AI in the financial industry. It demonstrates the potential of machine learning to create robust, flexible tools that can analyze complex data sets and provide vital, timely insights. By doing so, these tools enhance the stability and efficiency of payment systems, ensuring the robustness of economies at both the national and global levels.