Consortium Data & Fraud Sharing¶
Definition¶
Consortium data sharing enables multiple financial institutions to share fraud signals and verified identity data — dramatically improving detection by revealing patterns invisible to any single institution.
What's Shared¶
| Data Type | Purpose | Privacy Approach |
|---|---|---|
| Known fraud indicators | Alert other institutions to confirmed fraud | Anonymized/hashed identifiers |
| Device fingerprints | Link fraud attempts across institutions | Device hash only |
| Face embeddings | Cross-institutional 1:N dedup | Template-protected embeddings |
| Velocity data | Detect rapid multi-institution applications | Aggregated counts |
| Confirmed mule accounts | Block known mule accounts | Account hashes |
Consortium Models¶
| Model | Example | How It Works |
|---|---|---|
| Industry utility | UK CIFAS, US Early Warning Services | Central database, all members contribute and query |
| Platform-mediated | Alloy, LexisNexis, Socure | Platform aggregates data from all clients |
| Bilateral | Bank-to-bank agreements | Direct data sharing between specific institutions |
| Blockchain-based | Emerging approaches | Decentralized sharing with privacy-preserving computation |
Key Takeaways¶
Summary
- Cross-institutional data sharing multiplies fraud detection power — catches fraud rings, mules, velocity attacks
- Privacy-preserving techniques (hashing, differential privacy) enable sharing without exposing raw PII
- CIFAS (UK) and Early Warning Services (US) are the largest operational consortiums
- Platform-mediated sharing (via eKYC vendors like Alloy, Socure) is growing fastest