Key Findings
Digital Footprint Matches Credit Bureau Scores
Simple digital footprint variables have similar predictive power (AUC 69.6%) compared to credit bureau scores (AUC 68.3%) for predicting defaults
Complementary Information Sources
Digital footprints complement rather than substitute credit bureau information, with combined AUC of 73.6%
Reduced Default Rates
Implementation of digital footprint screening led to 50% reduction in default rates while maintaining credit access
Predictive Power Comparison
- Digital footprint alone achieves slightly better predictive power than credit bureau scores
- Combining both methods provides significant improvement in predictive accuracy
- Based on analysis of 254,819 customers with credit bureau scores
Default Rates by Digital Profile
- Mobile users show highest default rates at 2.14%
- Desktop users demonstrate lowest risk with 0.74% default rate
- Android users (1.79%) show higher default rates than iOS users (1.07%)
Impact on Default Rates After Implementation
- Default rates dropped by over 50% after implementing digital footprint screening
- Improvement seen across both scoring methods
- Maintained similar levels of credit access while reducing defaults
Contribution and Implications
- Demonstrates digital footprints as powerful alternative credit assessment tools
- Potential to expand financial inclusion for 2 billion unbanked adults worldwide
- Provides lenders with easily accessible, cost-effective screening method
- Highlights importance of digital behavior in financial assessment
Data Sources
- AUC comparison chart based on Table 4 regression results
- Default rates by device type visualization derived from Table 2
- Implementation impact analysis based on Table 9 data
- Study analyzed 270,399 purchases from German E-commerce company between October 2015-December 2016
Infographics of Publications by Journal
