When Digital Access Becomes Investment Capability: Financial Literacy and Portfolio Diversification Across Countries (Research Square)

Summary

• The panel covers five Global Findex waves (2011, 2014, 2017, 2021 and 2024) and contains 291 country-year observations for 82 countries.p.10
• Economic value-enabling conditions positively predict digital investment adoption (β=1.854, SE=0.378, p<0.001).p.16
• Social value-enabling conditions are also positive (β=0.448, SE=0.098, p<0.001).p.16
• Adoption is positively associated with the portfolio-diversification environment (β=0.049, SE=0.018, t=2.66, p=0.008).p.17
• The adoption and financial-literacy interaction is positive (β=0.088, SE=0.031, t=2.81, p=0.006), supporting a conversion-capability reading.p.17
• In the cluster analysis the mature-market interaction is large but imprecise (β=0.563, p=0.207; N=139 country-years).p.20

Key insights

1. Functional and emotional proxies are negative in the within-country specification, so more enabling conditions do not always produce more adoption.p.16
2. The adoption model explains 56.0 per cent of within-country variation in the standardised adoption proxy (within R²=0.560).p.17
3. The diversification model has a modest within R² of 0.069, as expected for a market-level outcome shaped by many other factors.p.17
4. Pooled correlations are high, with ECON correlating 0.98 with EMO and 0.91 with FL, so coefficients cannot be read as fully independent effects.p.14
5. The selected two-cluster solution has an average silhouette value of 0.478.p.13
6. The low-access cluster has a smaller interaction coefficient (β=0.048) that is statistically precise (p<0.001; N=152).p.20
7. Banks should pair acquisition metrics with post-adoption indicators rather than treat app adoption as the final marketing KPI.p.26
8. Regulators should combine affordable access, financial capability and market integrity rather than measure success by account ownership alone.p.27