Belgium Cracks Down on Social Security Fraud with Innovative Data Analytics Tool

A Pioneering Project in Belgium Uses Data Analytics to Combat Social Security Fraud

The National Social Security Office (NSSO) in collaboration with the Federal Public Service (FPS) Social Security, has implemented a groundbreaking project called MiningWatch. This data analytics tool uses predictive modeling to identify high-risk companies and detect social security fraud.

How MiningWatch Works

MiningWatch is designed to optimize social fraud detection while minimizing costs. The platform:

• Uses predictive risk modeling for fraud detection
• Integrates with existing data warehouse systems
• Provides a user-friendly interface for inspectors to target inspections based on predictive models
• Offers a real-time ranking system for companies by risk level (red, orange, green, and blue)

Key Benefits of MiningWatch

• Significantly increased detection rates, exceeding initial objectives
• Improved effectiveness of inspections through collaborative approach and knowledge sharing
• Enabled inspectors to prioritize efforts on high-risk cases

Cost-Effective Solution

The annual costs of the project amount to €200,000, including software licensing and service provider fees. A core team of 10 staff members is responsible for maintaining and improving the platform.

Evaluation Results

Evaluation results show that MiningWatch has exceeded its objectives, with detection rates increasing significantly since its introduction. In the construction sector alone:

• Average detection rates have risen from 16% to 45% among employer profiles identified as high-risk for temporary lay-off fraud

Lessons Learned

The project highlights the importance of:

• Not underestimating data complexity
• Factoring in sufficient time to optimize data quality
• Involving all users early in the development process
• Engaging users in validating predictive models

Transferability and Future Development

The level of transferability of MiningWatch depends on several factors, including the availability of administrative data, possibility of transferring confidential data, and collaboration between authorities. The success of this project serves as a model for other countries and sectors seeking to combat social security fraud using innovative data analytics tools.

Contact

Tom De Lust, Business Analyst, National Social Security Office Email: tom.delust@onssrszlss.fgov.be