Abstract Our paper experimentally tests whether ‘soft regulatory devices’ that rely on advisor identifiability and reciprocity are able to reduce misconduct by financial advisors. We also test whether advisor ability is negatively related to misconduct. We indeed find that low ability advisors are more likely to engage in misconduct in all treatments including the Baseline. Interestingly, a subset of advisors strategically exploits the identifiability mechanism in order to ‘game the system’, which undermines the effectiveness of the treatment. Furthermore, we find that reciprocity affects advisors heterogeneously, which causes a group of advisors to increase their misconduct and leads to a significant loss in efficiency in comparison to the Baseline treatment. Additionally, we discuss how low ability advisors have different incentives within each treatment compared to high ability advisors. Our analysis reveals substantial disparities in advisor misconduct across various contexts as interventions influence misconduct in a nuanced manner. Finally, we argue for the use of segmented policy approaches in order to respond to complex market environments involving heterogeneous advisors.