Comparative Analysis of Alternative Credit Scoring Models: Expanding Access and Identifying Limitations

Background: Access to credit remains a fundamental barrier to financial inclusion, with millions worldwide excluded from traditional banking systems due to inadequate credit histories. This study investigates whether alternative credit scoring models effectively expand credit access for underserved populations while examining the constraints that limit their equitable application across different markets. Methodology/Principal Findings: A comparative analysis was conducted of three leading fintech companies—Upstart, Tala, and Kaleidofin—using publicly available financial reports, impact assessments, and peer-reviewed literature. The analysis evaluated credit approval rates, borrower demographics, default rates, and customer outcomes. Results demonstrate that alternative credit scoring models significantly increase credit access. Upstart's AI-driven model approves 44.28% more borrowers than traditional models, with 28.8% of loans directed to lowto-moderate-income communities. Tala has served over 7.5 million borrowers globally, with 63% being first-time digital borrowers and 84% reporting improved quality of life. Kaleidofin serves 7.55 million customers (98% women), achieving 20-30% higher approval rates than traditional bureaus while maintaining portfolio-at-risk below 2%. Collectively, these models have reached approximately 18 million previously excluded borrowers. However, effectiveness is constrained by significant limitations: extensive data collection requirements raising privacy concerns, persistent algorithmic fairness challenges despite bias mitigation efforts, and regulatory barriers limiting international scalability. Conclusions/Significance: While alternative credit scoring models demonstrate substantial success in expanding financial inclusion, their transformative potential requires coordinated efforts to address underlying challenges.