ABSTRACT
Using a unique dataset of FINRA-licenced individuals in Florida in 2015 that was enriched to include job classification information generally not contained in publicly-available regulatory data, a series of binary logistic regressions illustrate how unobserved differences among financial service professional roles may bias results in misconduct analyses. When using CFP® status as the sole predictor of misconduct among the full sample of licenced individuals, CFP® professionals are found to have 1.86 times higher odds of having engaged in culpable advisory-related misconduct compared to non-CFP® professionals. However, after controlling for other relevant factors and limiting the sample to only individuals identified as financial advisors, CFP® professionals are found to have 0.84 times lower odds of having engaged in culpable advisory-related misconduct. Because job classifications are generally not available in the standard SEC and FINRA datasets, these findings illustrate how the inability to control for unobserved differences in job roles may bias misconduct analyses.
KEYWORDS:
Disclosure statement
The authors wish to disclose potential conflicts of interest. Derek Tharp is a CFP® professional. CFP® professionals are subject to CFP Board oversight. Jeffrey Camarda sued the CFP Board and voluntarily relinquished his CFP® marks over a dispute regarding the integrity of the CFP Board's disciplinary process.
Notes
1 Although CFP® standards are higher than minimum industry requirements, Rhoades (Citation2020) notes that enforcement mechanisms are lacking.
2 Results were robust to Camarda’s (Citation2016) B-DIS and B-ADIS definitions of misconduct.
3 Comprehensive U4 data for Florida were obtained by purchase from the vendor Discovery Data which purports to collect comprehensive and accurate data on the financial services industry across the US. These aggregated raw data were sold by this vendor serving advisor firms’ recruiting and other needs. This vendor purchased the data from Florida, which obtained the data directly from FINRA to discharge Florida’s securities regulatory duties. Florida sold the data to vendors such as ours who enrich and resell it to industry firms for recruiting and other uses. The job description enrichment process for our 2015 data utilized a typical variety of unverified internet algorithmic techniques, such as keyword matching and job/income assumptions based on residential valuations. While there is no reason to believe these data are any less accurate than secondary data sources in general, they likely contain inevitable collection, modelling, and transcription errors. These unenriched data are generally consistent in content with those available from FINRA’s BrokerCheck and those referenced in other studies (e.g., see McCann, Qin, and Yan Citation2017). Given the inherent imprecision of such modelling techniques, we strongly support regulatory requirements to identify the client-facing status of securities registrants and other financial advisors.