Background Saudi Arabia’s digital payment capability reached 85 percent of retail transactions in 2025. However, the firm-level development procedures remain unchanged. Purpose This study analyzes whether the relationship between digital finance and upgrading is influenced by accounting conditions at the firm level: digital finance induces traceability of transactions, while accounting formalization induces verification and information discipline that is conducive to traces translating to managerial signals. Method We build the Digital Finance Index (DFI) and Accounting Formalization Index (AFI) and test the complementarity prediction of the DFI × AFI in the spirit of Athey-Stern, which can account for 1,002 firms using the survey-weighted estimation alongside the Romano-Wolf family-wise correction across four pre-specified hypothesis families, Oster and Cinelli-Hazlett sensitivity bounds, and the wild-cluster bootstrap inference (B = 1,999) for the special features of the six sector clusters. Findings Supermodularity prediction was not supported. The DFI × AFI interaction on labor productivity is negative in the point estimate (α₃ = −0.038), but it loses massively under sector-level resampling (wild p = 0.517 versus cluster-robust p ~ 0.004). None of the AFI subchannels carried precise moderation. Outcome-Specific Patterns Two outcome-specific patterns were highly factual. Labor productivity and digital finance have a positive association (β = +0.075, wild p = 0.040, Romano-Wolf p = 0.046) and a negative association with product innovation (Romano-Wolf p = 0.046), which is robust to Oster but at the binding Cinelli-Hazlett benchmark. Conclusion The pattern is about specialization, one on the transactional versus the other on the experimental aspects, instead of complementarity to raise precision. Implications Findings are more of conditional associations rather than causal effects; bundled interventions, including onboarding and audit-readiness supports for digital finance, but add SME targeting capability vouchers warrant testing.
Ariyan et al. (Tue,) studied this question.