Institutional Legitimacy as a Heuristic for Following Digital Financial Advice: Evidence from a Preregistered Two-Module Experiment

We study how institutional identity cues, transparency disclaimers, and access rules to human support shape adherence to standardized digital financial advice. In a preregistered experiment with an interactive portfolio task (N = 403 in the pre-final analysis sample), we randomize advisor identity (bank vs independent) and disclaimer presence, and independently randomize a policy describing equal versus wealth-gated access to human support. Primary outcomes measure divergence from the recommendation (L1 and Euclidean distances) and exact adherence. We report confirmatory tests with Holm correction for the identity effect across divergence measures, and assess distributional robustness via hurdle and Tobit models. The access-policy manipulation strongly affects perceived fairness/legitimacy/inclusion (FALI), while identity effects on divergence are small and statistically indistinguishable from zero in the pre-final dataset.