Saudi banks that deploy AI-powered FinTech tools show stronger returns, higher valuations and greater stability, alongside improved ESG metrics; results hold across panel and dynamic models but stem from a small, single-country sample.

Main Finding

Banks’ internal adoption of AI-enabled FinTech-related digital tools (e.g., machine-learning credit assessment, robo-advisory, automated compliance) is positively and statistically significantly associated with better market performance (Tobin’s Q), higher accounting performance (ROA, ROE), improved financial stability (Z‑Score), and stronger sustainability outcomes (Bloomberg ESG Disclosure Score and LSEG ESG performance score) in the Saudi banking sector over 2015–2024. These relationships are robust across multiple estimation strategies that address potential simultaneity and dynamics.

Key Points

• Sample and period: Panel of 10 Saudi banks observed 2015–2024.
• AI-enabled tools studied: machine-learning credit scoring, robo-advisory systems, automated compliance and monitoring tools, and related FinTech digital solutions deployed internally by banks.
• Outcome variables:
  • Market performance: Tobin’s Q
  • Accounting performance: Return on Assets (ROA), Return on Equity (ROE)
  • Financial stability: Z‑Score
  • Sustainability: Bloomberg ESG Disclosure Score and LSEG ESG performance-oriented score
• Controls: bank size, age, leverage, liquidity, loan-to-deposit ratio, non-performing loans, net interest margin, market capitalization, board size.
• Estimation approaches:
  • Pooled OLS and Fixed Effects (FE)
  • Driscoll–Kraay standard errors to account for cross-sectional dependence and heteroskedasticity
  • Dynamic FE models including lagged dependent variables, lagged independent variables, and shock-interaction terms to mitigate simultaneity and capture adjustment dynamics
• Robustness: Positive associations persist across specifications and dynamic treatments, suggesting results are not driven by a single model choice.
• Interpretation: Adoption of AI-enabled FinTech tools is linked with efficiency gains, improved risk management and compliance, and enhanced disclosure/practices that cohere with better ESG outcomes.

Data & Methods

• Data: Bank-level panel (10 banks) covering 2015–2024; internal adoption of AI-enabled FinTech tools measured as banks’ internal deployment intensity (study-specific adoption metric).
• Dependent variables: Tobin’s Q, ROA, ROE, Z‑Score, Bloomberg ESG Disclosure Score, LSEG ESG performance score.
• Independent variable of interest: internal adoption of AI-enabled FinTech-related digital tools.
• Control variables: size, age, leverage, liquidity, loan-to-deposit ratio, non-performing loans, net interest margin, market capitalization, board size.
• Estimation strategy:
  • Baseline: Pooled OLS and bank fixed effects to account for time‑invariant heterogeneity.
  • Inference: Driscoll–Kraay SEs to address cross-sectional dependence and serial correlation.
  • Dynamics and endogeneity concerns: dynamic FE models including lagged dependent variables and lagged regressors, plus shock‑interaction terms (to reduce simultaneity bias and capture response to shocks).
  • Robustness checks: consistency of direction and significance across model families reported.

Implications for AI Economics

• Productivity and value creation: Evidence that internal AI-enabled FinTech adoption can raise firm-level market and accounting performance supports arguments that digital/AI investments generate measurable firm productivity gains in banking.
• Risk management and stability: Positive association with Z‑Score suggests AI tools can strengthen risk monitoring and resilience—relevant to macroprudential assessments of technology adoption as systemic risk mitigant or amplifier.
• ESG and sustainable finance: Correlated improvements in ESG disclosure and performance imply that digitalization and AI can complement sustainability objectives through better reporting, monitoring, and targeted product offerings.
• Policy and regulatory takeaways:
  • Encourage responsible AI adoption via regulatory sandboxes, guidance on model risk management, and standards for governance, explainability, and data quality.
  • Monitor systemic implications as adoption scales — regulators should track concentration, third‑party dependencies (cloud, vendors), and operational risk.
  • Support capability building (skills, data infrastructure) in banks to realize benefits while containing risks.
• Directions for further research:
  • Causal identification (natural experiments, instrumental variables) to more firmly establish causality and quantify effect sizes.
  • Larger and cross-country samples to test generalizability beyond Saudi Arabia.
  • Disaggregated analysis by AI application (credit scoring vs. compliance vs. advisory) to identify heterogeneous impacts.
  • Investigation of labor effects, vendor concentration, and potential distributional or competitive consequences of AI adoption in banking.