Relation between financial advisory designations and FINRA misconduct FSR201714 Jeffrey M. Camarda, PhD, CFA EAa,* a4371 US 17, Fleming Island, FL 32003, USA Abstract Registered representatives have no general fiduciary duty. CFP, ChFC, and CFA designees have higher ethical duties and education requirements. Registrants’ criminal, regulatory, complaint, and other misconduct history is public. This study examines misconduct disclosures of undesignated versus designated Florida securities salespeople, and finds adverse disclosure materially decreases for designees; it incidentally finds misconduct increases with maleness, dual investment advisor/regis- tered representative status, and life insurance sales licensure. This appears to be the first such study of adverse disclosure association with financial designations, adding to the emerging misconduct and advisors’ ethics literature. These findings offer important policy and consumer choice insight. © 2017 Academy of Financial Services. All rights reserved. JEL classification: G020; G180; G220; G280; G290 Keywords: Advisor; Fiduciary; Designation; Best interest; Misconduct 1. Introduction Financial Industry Regulatory Authority (FINRA) registered representatives (RRs) have come to be known as financial “advisors.” RRs, as agents of Broker Dealer (BDs) organi- zations, are typically commission salespeople generally not required to put customers’ interests ahead of their own compensation, in contrast to Investment Advisor Representatives (IARs), who are agents of fiduciary Registered Investment Advisor organizations (RIAs). Many RRs are also IARs, with conflicting rules and duties, often to the same clients, * Corresponding author. Tel.: �1-904-278-1177; fax: �1-278-1070. E-mail address: j@camarda.com Financial Services Review 26 (2017) 271–290 1057-0810/17/$ – see front matter © 2017 Academy of Financial Services. All rights reserved. depending on the hat worn. The respective codes for Certified Financial Planner (CFP), Chartered Financial Consultant (ChFC), and Chartered Financial Analyst (CFA) designees require greater financial expertise and ethical duties to clients than do RR rules. Many RRs are designees. RR criminal, regulatory, complaint, and other misconduct history is public record. This study examines the comparative misconduct of undesignated versus designated RRs in Florida. It appears to be the first such study of adverse disclosure association with financial designations, so adding to the emerging FINRA misconduct literature (Egan, Matvos, & Seru, 2016). As the financial advisory industry evolves and regulators and other stakeholders seek to enhance practice quality and duties to investors, these findings may offer important contributions to policy, and help consumers to make better advisor choices. 2. Background More than 650,000 financial advisors help manage over $30 trillion of investable assets in the United States (Egan, Matvos, & Seru, 2016), with over one-half of all households and nearly 90% of consumers with investable assets over $100,000 seeking such help (Smith, Vibhakar, & Terry, 2008). Advisory wrongdoing has not undergone rigorous investigation and analysis (Zingales, 2015). RR sales licensing is relatively simple, and even exam cheaters may become licensed (FINRA, 2016). Even sophisticated consumers can easily confuse RR sales reps for fiduciary advisors (Government Accountability Office [GAO], 2011). RR misconduct is widespread, and tends to concentrate at firms that may enable it (Egan et al., 2016) or even depend on it as a profit center (Woolley, 2016). Consumers seem generally ignorant of these factors (McCann, Qin, & Yan, 2016), and social cost is not yet known (Ritholtz, 2016). Ethically challenged advisor behavior is estimated to waste some $17 billion per year of consumer wealth (The White House, 2016). While RR misconduct data are public, limited accessibility may damage consumers (PIABA Foundation, 2015). Designations sponsors’ (CFA Institute, 2016, CFP Board, 2016, The American College, 2016), send consumer quality signals (Terry & Vibhakar, 2011) implying lower misconduct for designees. Findings of valid signals could have important consumer welfare and policy implications, provide justification to adopt broad fiduciary standards across the multiple advisory channels, and nurture standardized professionalism. Unlike established professions such as law, medicine, and accounting, advisors have wide variations in ethical duties and required education. Regulatory gaps are significant and persistent (GAO, 2011). RIA companies and their IAR agents are required to protect clients’ best interests by being held to a fiduciary standard, but RRs— even when dully licensed as IARs—are not. RRs may appear to consumers as fiduciaries, but are generally held only to the lower suitability standard, allowing them to put their own interests first. The data here precede implementation of the emerging DOL rule that would apply a limited fiduciary RR duty to retirement accounts only. Very many study RRs have the CFP, ChFC, and CFA designations (the “study designa- tions”). These require meaningful study, examination, and allegiance to ethical codes with much stronger duties than RR rules. Designated and undesignated RRs are hired to sell commission products and operate under the same regulations; neither are legally generally 272 J.M. Camarda / Financial Services Review 26 (2017) 271–290 required to act as fiduciaries. Many RRs offer advice under marketing flags that inaccurately imply fiduciary duty (Raymond James Mission Statement, 2016), flown by some BDs with the very highest misconduct rates (Egan, Matvos, & Seru, 2016). Such misleading claims seem common. (Vystar Ad, 2016, p. B-8). Hauptman and Roper (2017) find such claims endemic to the BD industry. FINRA advisors exhibit a high level of misconduct, and an anomalous concentration of repeat offenders, compared with physicians (Egan et al., 2016). 3. The question of financial advisor professionalism There are no uniform professional standards to which those calling themselves financial advisors must adhere, and on which the public can rely. The occupation, as a whole, does not rise to established standards of professionalism. While some, such as study designees, do or purport to, it is important to note that these represent an occupational subset, and that the public may not be aware of the difference. This study distills six theoretical attributes of a profession from the literature (Dean, 1997; Flexner, 1976; Flexner & Metzger, 1976; Khurana, Nohria, & Penrice, 2005, Khurana & Nohria, 2008; Ragatz & Duska, 2010; Schaefer, 1984) and tests advisors against them: 1. Specialized, arcane, deep, socially useful expert knowledge requiring hard and constant study, that is codified, evolves and is perpetuated: While a rigorous body of academic knowledge is developing (Kitces, 2014b), there is no requirement that a practitioner obtain it (Kitces, 2015b; Moisand, 2008). 2. Rigorous testing, vetting, and certification of expert knowledge and ability: No prac- titioners’ requirement (GAO, 2011; Kitces, 2015b). 3. The placing of explored client interests before the professional’s: There is no general requirement of fiduciary duty (Cummings & Finke, 2010; GAO, 2011). 4. Formal system for professional standards promulgation and oversight: While a number of detailed financial designation ethical codes have emerged as cataloged by Ragatz and Duska (2010), these do not apply to all financial advisory practitioners. (GAO, 2011; Ragatz & Duska, 2010). 5. Monopoly power derived from social contract: This aspect is completely nonexistent in the current environment (Cummings & Finke, 2010; GAO, 2011). 6. Commitment to excellence, service, collegiality, and dignified conduct: For the reasons discussed under 4, above, this test is not met. Moisand (2012) analyzes financial planning and also concludes it is not a profession. Others (Financial Planning Coalition, 2014; Frumento & Korenman, 2013) concur and suggest the lack of real governmental regulation as a profession promotes unethical and fraudulent activity. While a profession is clearly emerging, a still-nascent commitment to formalized academic development is critical to its success (Warschauer, 2002). Undefined terms like financial advisor are freely used without regulation (GAO, 2011). Consumers may not understand the difference between fiduciary IARs and RRs who are free to enrich themselves at clients’ expense (Cummings & Finke, 2010; Finke & Langdon, 2012; GAO, 2011) but market and sell as if they put clients first (Hauptman & Roper, 2017). 273J.M. Camarda / Financial Services Review 26 (2017) 271–290 Established monopolistic professions like law and medicine use strong signals to establish professionalism. In nonmonopolistic markets, signal theory purports that agents signal skills with educational credentials; acquiring a credential is essentially a reputation-for-quality purchase (Spence, 1973; Spence, 2002). In cases like financial advisory where uniform professional standards are lacking, practitioners can distinguish themselves from the merely licensed by associating with certifying bodies (Mauldin, Wilder, & Stocks, 2000). To consumers, signals like MD or CPA offer clear indications of baseline expertise, and study designations act similarly to these with the key difference that they are not required by regulation. Mauldin, Wilder, and Stocks (2000) and Brockman and Brooks (1998) find designations signal objectivity, expertise, and ethics to consumers. Smith, Vibhakar, and Terry (2008) find advisors obtain designations to establish professional expertise and cred- ibility. Designations serve as an “umbrella brand for the . . . CFP or CFA . . . (that’s) worked successfully for CPAs for decades (Smith et al., 2008, p. 308). The effect of the study designations on RR misconduct is this study’s primary research question. The CFP mark has become the preeminent financial planning designation (Kitces, 2015), largely because of expensive and effective CFP Board marketing efforts (Kitces, 2015c). It is now most demanded by consumers, and more recognized by consumers by a factor of over two compared to CFA and over four compared to ChFC (CFP Board, 2015). It is perceived to be more appropriate for financial planning than the more specialized investments-expert CFA (Terry & Vibhakar, 2011). Non-CFA CFPs’ confidence in their investment skills may be overstated as compared to those who have acquired both and perhaps better recognize the limits of their knowledge (Cordell, Smith, & Terry, 2011). It is worth noting that this study deals with investment sales agents, and that the CFA investment knowledge set is much greater than those for CFP and ChFC. The CFA designation requires three successive eight hour exams on investments man- agement. The six-hour CFP exam tests nine subject areas, including investments, insurance and risk management, income taxes, estate planning, retirement planning, employee benefits, professional conduct, financial planning principles, and financial plan development. The ChFC designation requires the completion of nine courses in financial planning, including income taxation, insurance, retirement planning, investments, estate planning, and case studies (The American College Chartered Financial Consultant, 2016b). The CFP Board Standards of Professional Conduct requires certificates to “at all times place the interest of the client ahead of his or her own,” (CFP Standards of Professional Conduct, 2008/2014), but it is worth noting that CFP Board’s public position has been that a fiduciary duty is only owed for financial planning engagements as opposed to product sales. The CFA Code of Ethics requires charter holders to place the interests of clients above their own, to act with integrity, competence, and respect, and to maintain and develop professional expertise (CFA Institute, 2015). “ChFC advisors are required to do the same for clients that they would do for themselves in similar circumstances, the standard of ethical behavior most beneficial for their clients” (The American College The Highest Standard, 2016a, p. 1). From a fiduciary perspective, the CFA, ChFC, and CFP ethical requirements are functionally equivalent, and well beyond the RR regulatory standard. The designation codes do not carry the force of law, and compliance—and ethics code understandings—may vary considerably among designees. 274 J.M. Camarda / Financial Services Review 26 (2017) 271–290 Conferring institutions’ claims of designations’ signal validity are compelling and uncompro- mising. ChFCs are said to have “the most extensive educational program . . . strict ethical standards, and . . . serve you with the highest level of professionalism” (The American College the Highest Standard, 2016, p. 1; The American College Code of Ethics, 2016c). CFP Board notes that “. . . CFPs . . . have . . . extensive training and experience . . . and . . . held to rigorous ethical standards . . . will make recommendations in your best interest” (CFP Board, 2016, p. 1). CFAs are promoted as setting “. . . the global standard for . . . integrity, dedication, and advanced skills . . . no credential is as widely respected . . . (to) serve the best interests of investors and society” (CFA Institute Value of the CFA Charter, 2016, p. 1). FINRA misconduct studies are nascent. Barry and Eaglesham (2014) find that data restrictions confound study, and that industry-controlled FINRA does this intentionally. McCann, Qin, and Yan (2016) agree, finding this done to promote the illusion of transpar- ency, and conceal that misconduct is recurring and predictable. Eagan, Matvos, and Seru (2016) build a regressable database by accessing individual RRs’ BrokerCheck records, one at a time, to facilitate statistical analysis, finding that 12% of RRs have misconduct disclosures and 7% have been disciplined for misconduct or fraud, and that those with misconduct are five times more likely to repeat it. They find such misconduct elevates in firms targeting customers in areas with concentrations of people who are elderly, and/or have high incomes. Of note to this study, Florida has a high concentration of such populations. 4. Research design and methodology This study uses secondary FINRA data. RRs must keep current form U4, which requires answers to 57 misconduct questions, involving criminal, regulatory, civil, complaint, and other adverse items. FINRA maintains records of disclosure events that may be indicative of unethical practice, which this study refers to as disclosure items, or misconduct. Misconduct measures comprise the dependent variables using a unique model developed for this study that compiles adverse yes disclosure answers from RRs’ U4s. This study’s independent variable (IV) is the presence of at least one of the study designations. Control IVs include age and gender, years of RR registration, employee verses independent contractor status, the presence or absence of a life insurance sales license and the presence of an active fiduciary investment advisor’s IAR license in addition to the RR investments products sales license. Given their additional education and ethical requirements, designees should be more effec- tive in subordinating their interests and dispensing better advice than undesignated reps. This might reasonably be expected to be associated with lower disclosure. Because this study restricts analysis to RR sales agents, it controls for the difference between suitability sales agents and legally dedicated fiduciary non-RR IARs, because these latter do not also sell securities products under suitability. It should be emphasized that some study RRs wear both hats, enhancing conflicts for these. Including only those licensed as RRs is intended to highlight any pure designation effect differences associated with those bound by designation fiduciary ethical codes who operate in a suitability standard product sales environment. There is potential endogeneity, but it is unclear what, if any, bias this has on results. Would-be designees with misconduct may be prevented, via prescreening, suspensions, or revocations, 275J.M. Camarda / Financial Services Review 26 (2017) 271–290 from using a designation, resulting in a mechanical suppression of designee misconduct per se. This could induce a survivorship bias and a cross section study flaw where misconduct is found to decline for designees. It is possible than any misconduct or revocation effect may be somewhat self-correcting because designees disclosing misconduct may find their desig- nations revoked after a lag, who would subsequently be measured as undesignated RRs, and that these fluctuations would damp out in large samples over time. If endogeneity is a significant factor—because designees are closely scrutinized for disclosure events, and because observations of disclosed misconduct are correlated with revocations—one would expect lower scores for surviving designees, which should affirm the desired signaling mechanism. It is also quite possible that a number of errant designees slip through the designators’ enforcement nets. It should be noted that professional bodies such as the CFP Board and the CFA Institute may only on third party complaints and the self reporting of ethical violations for this information; perhaps routine screening via the BrokerCheck algorithm would be prudent. It should be also be noted that while the RIA/IAR disclosure form ADV discloses designation suspensions or revocations, the RR form U4 does not. There are no reasonably clear endogeneity effects affecting this study’s results. The regression models equate various measures of misconduct against the study and control independent variables available in the data. The primary dependent variable (DV) is a construct referred to as Disclosure Incidence Score (DIS); all other DVs are derived from DIS. The DIS model assigns a point value subscore to each U4 yes answer depending on nature and severity. DIS results shows a censored range from 0 to 56. Most registrants’ misconduct scores are at or near zero, producing a very non-normal distribution. The target population are those licensed as retail RRs in Florida. Those modeled as nonretail RRs, such as analysts, are removed. From 35,361 RRs in the raw data, 26,667 are modeled as retail. Of these, 116 (0.4%) have the CFA, 2,534 (9.50%) the CFP, and 970 (3.64%) the ChFC. Florida is one of the three highest misconduct states, has one of the highest concentrations of RRs (Egan et al., 2016), and is one of only 12 states imposing a limited fiduciary duty on RRs (Finke & Langdon, 2012). Given this, the greater Florida misconduct incidence noted by Egan et al., (2016) may underscore any designation effect found here. U4 data are obtained from a vendor serving industry recruiting interests, and are generally consistent with those on BrokerCheck and referenced in other studies (Barry & Eaglesham, 2014; Egan et al., 2016; McCann et al., 2016). Given this study’s data comprise virtually the entire population, traditional reliability concerns are not an issue. Data are such that misconduct scores are cumulative and not transitory. Note they allow for only one response to each question, regardless of the number of incidents to which it might apply. For example, a RR whose license has been revoked six times would show the same score for the applicable question as one revoked only once; this will likely produce under-measurement distortions. 5. Model specification, dependent and independent variables DIS is comprised of the sum of the subscores for all U4 disclosure questions, advisor- related or not, and whether indicative of mere allegation or clear wrongdoing. DIS is thus the broadest misconduct measure in the study. DIS scoring methodology is described below. 276 J.M. Camarda / Financial Services Review 26 (2017) 271–290 ADIS (Advisor DIS), is a DIS subset restricted to the sum of all DIS subscores that relate only to advisor functions; for instance, financial fraud or regulatory suspension would be included, but domestic abuse or personal bankruptcy would not. ADIS includes both allegations and indications of culpability. CAD (Culpable Advisor Disclosure) sums only the subscores that are both advisor-related and indicate findings of wrongdoing or other clear indications of culpability. As such it is this study’s most specific and important misconduct measure. The B-DVs are binary transformations of the corresponding DVs. For each, if the correspondent is greater than zero, the B- value would be one; if the correspondent is zero, the B-value is zero. Note scalar information is lost; DIS � 56 or DIS � 1 both become B-DIS � 1. The binary transformations are needed for Logit and facilitate informative descriptive statistics. B-DIS – the binary transformation of DIS. B-ADIS – the binary transformation of ADIS. B-CAD – the binary transformation of CAD. AnyDes is an indicator IV for RRs holding at least any one of the three study designations. This is the study IV. This study also uses control IVs of age, gender, years FINRA registered, life insurance sales licensure, independent contractor versus employee status, and dual RR/IAR registration status. Employment status seeks to control for possible higher employer supervision effects verses perhaps reduced compliance oversight for nonemployee contrac- tors. Note dual registration means the subjects can do business as both RRs and IARs, a conflicted situation given respective suitability versus fiduciary duties. RRs with insurance licenses have the ability to also sell fixed and variable life insurance products such as annuities, often marketed as investments, which have a robust history of complaints. While the control variables are not hypothesized in this study as explanatory, some of their results seem quite interesting and worthy of further study. 6. Construction of dependent variables There are 57 individual disclosure categories queried on the form U4. The excerpt in Fig. 1 offers an example of these questions. Some of these—such as “have you ever been charged with any felony?”—disclose allegation, but not culpability, of a non-advisor–specific disclosure, while others, such as “have you ever been convicted of . . . a misdemeanor involving investments . . . or any fraud . . .” connote a finding of clear advisor-related wrongdoing. To assess the potential impact on advisory quality, the study quantifies each yes disclosure based on the applicable two of four possible conditions: is the disclosure of mere allegation, or does it reasonably indicate culpability? Does the disclosure specifically relate to theft or consumer investment harm, or instead relate to nonadvisor-specific unethical behavior? Each condition is assigned a DIS factor value of 1–4, and the product of the factors corresponding to a specific combination of conditions determines the misconduct subscore for a particular disclosure question. For each registrant, each question’s subscores are summed to determine a regis- 277J.M. Camarda / Financial Services Review 26 (2017) 271–290 trant’s total disclosure score, which is used as the basis for the dependent variables in this study. To reiterate, this study uses four conditions and four DIS factors to determine the DIS and other study DVs: 1. Does the item generally relate to non-investments–specific amoral or unethical behav- ior? DIS factor � 1 2. Does the item directly relate to registrant’s actions with respect to investments or theft? DIS factor � 2 3. Does the item relate to allegations against registrant without other implication of culpability? DIS factor � 1 4. Does the item relate to allegations against registrant with reasonable implication of culpability? DIS factor � 2 Both a sum and a product version of the DIS calculation were considered for this study. The product version is used as it offers a more appropriate severity score gradient. For instance, as seen in Table 1, conditions one and three—mere allegation of a non-advisory item—only produces a subscore of 1 using the product method, one-fourth as much as a finding of an advisory item like financial fraud that would yields a subscore of 4. The sum method would produce a subscore of 2 for a disclosure of a mere allegation of non-advisory Fig. 1. Typical FINRA form U4 questions. 278 J.M. Camarda / Financial Services Review 26 (2017) 271–290 issue, twice the value produced by the product form. This seems intuitively disproportionate. These subscores are ascribed to specific U4 questions as illustrated in Fig. 2. 7. Descriptive statistics As seen in Table 2, only about 12% of the Florida RR population holds any study designation at all. Of those that do, CFPs are by far the most numerous, with less than half as many ChFCs, and with the CFA barely represented. The study population is overwhelmingly male, mostly employee, with over half holding insurance licenses and IAR registrations. Table 1 DIS subscores as function of condition combinations and product form DIS subscore Condition combination Condition 1 or 2 Condition 3 or 4 DIS question subscore as product of factors 1,3 Not investment/theft specific No clear indication culpability 1 1,4 Not investment/theft specific Clear indication culpability 2 2,3 Investment/theft specific No clear indication culpability 2 2,4 Investment/theft specific Clear indication culpability 4 DIS � Disclosure Incidence Score. Fig. 2. DIS subscoring of U4 questions. 279J.M. Camarda / Financial Services Review 26 (2017) 271–290 Of note in Table 3, non-CFA designees tend to be older and longer-registered than other non-designees. This is also true for insurance licensees and somewhat true for independent contractors. Study designees are predominantly male as seen in Table 4, with designation holders over 80% male. Given the sample gender distribution from Table 2, it is not surprising that all IV categories are mostly populated by males. We also see in Table 4 that most CFPs and CFAs are employees, but that most ChFCs are independent contractors. It is perhaps not surprising that most CFPs and ChFCs, and hence most designees, are insurance licensed given these financial planning designations include strong insurance components. It was unexpected that over half of CFAs are licensed to sell life insurance, but this does underscore the study model’s utility in identifying CFAs who play retail advisory roles, and who are perhaps required to be licensed to sell these products. Table 4 also reveals that most independent Table 2 Independent variables counts and sample weights Independent variable Count in 26,667 sample Percent of sample Holds at least one of any designation 3,197 12.00% Holds CFP 2,534 9.50% Holds ChFC 970 3.60% Holds CFA 116 0.40% Male 19,815 74.30% Female 6,852 25.70% Employee 16,871 63.30% Independent contractor 9,796 36.70% Dually licensed as RR and IAR 14,964 56.10% Holds life insurances/annuities license 15,324 57.50% CFP � Certified Financial Planner; ChFC � Chartered Financial Consultant; CFA � Chartered Financial Analyst; RR � registered representatives; IAR � Investment Advisor Representatives. Table 3 Mean ages and years registered by independent variable and individual designations Independent variable � Age � Years registered Any designation 51 23 No designation 42 10 CFP 53 22 Non-CFP 47 15 ChFC 57 26 Non-ChFC 47 15 CFA 48 18 Non-CFA 48 15 Male 48 16 Female 47 13 Independent contractor 52 17 Employee 45 14 IAR 48 17 Non-IAR 48 14 Insurance license 52 19 No insurance license 42 10 CFP � Certified Financial Planner; ChFC � Chartered Financial Consultant; CFA � Chartered Financial Analyst; IAR � Investment Advisor Representatives. 280 J.M. Camarda / Financial Services Review 26 (2017) 271–290 contractors are insurance licensed, and that a slight majority of all RRs are also registered as fiduciary IARs, with this tendency substantially increased for designees, particularly CFPs and CFAs. RR/IARs tend to be employees. Finally, Table 4 also shows there’s a tendency for RRs to also be licensed as fiduciary investment advisors and to sell commissionable life insurance as well as commissionable securities as RRs. Average misconduct scores for the sample and by independent variable are reviewed in Table 5. It should be noted that the heavy skewness of the discrete DVs toward the zero bound of the distributions limits the interpretative value of these means. An alternative perspective is offered by the means of the binary DVs that report the percentage of subsample disclosing misconduct by DV, without magnitude. It seems noteworthy that scores and percentage misconduct uniformly rise for the CFP and ChFC designations, but that the opposite is observed for CFAs. In other words, CFPs and ChFCs, respectively show higher misconduct than non-CFPs and non-ChFCs, and the percentages of sample showing any misconduct at all for each measure is higher for these designees than for those without them. While these observations are mitigated somewhat by regression against control IVs, it is interesting that the three designations share similar age and gender profiles, and that gender is clearly the single most powerful confounding factor appearing in these descriptive statistics, with males showing consistently sharply higher misconduct. Besides maleness, IAR and insurance licensee status also show strong associations with higher misconduct. 8. Test hypotheses The null hypothesis is that the study misconduct measures are not affected by having any one of the study designations, so that that there is no disclosure difference associated with Table 4 Gender, employment, insurance license, and RR/IAR status distribution by independent variable Independent variable Male Female IAR Not IAR Independent contractor Employee Insurance license No insurance Any designation 82% 18% 70% 30% 49% 51% 89% 11% No designation 73% 27% 54% 46% 35% 65% 53% 47% CFP 81% 19% 74% 26% 45% 55% 91% 9% Non-CFP 74% 26% 54% 46% 36% 64% 54% 46% ChFC 86% 14% 59% 41% 65% 35% 92% 8% Non-ChFC 74% 26% 56% 44% 36% 64% 56% 44% CFA 84% 16% 76% 24% 25% 75% 55% 45% Non-CFA 74% 26% 56% 44% 37% 63% 57% 43% Male 100% 0% 59% 41% 39% 61% 60% 40% Female 0% 100% 47% 53% 30% 70% 50% 50% Independent contractor 79% 21% 44% 56% 100% 0% 74% 26% Employee 72% 37% 63% 37% 0% 100% 48% 52% IAR 79% 21% 100% 0% 29% 71% 64% 36% Non-IAR 69% 31% 0% 100% 47% 53% 49% 51% Insurance license 78% 22% 62% 38% 47% 53% 100% 0% No insurance license 70% 30% 48% 52% 23% 77% 0% 100% CFP � Certified Financial Planner; ChFC � Chartered Financial Consultant; CFA � Chartered Financial Analyst; RR � registered representatives; IAR � Investment Advisor Representatives. 281J.M. Camarda / Financial Services Review 26 (2017) 271–290 having any designation compared with having none. The alternate is that scores will be different for holders of at least one of these designations. It is expected that designated RRs should show lower misconduct. H0: �AnyDes � 0 H1: �AnyDes � 0 The primary regression method is Tobit; data are censored, with results heavily skewed, with most at or close to zero. Logit is the primary, and OLS a secondary, robustness check. The regressions explore if having any of the study designations is associated with lower misconduct after controlling for other factors available in the dataset, including those identified in the descriptive statistics to be associated with higher levels of misconduct, such as insurance licensure, IAR registration, and maleness. As noted, higher misconduct scores are generally associated with designations, the exception being CFAs, which are quite scarce in the sample. The regressions seek to unbundle the effects. The Tobit and OLS regression models specifications are: DIS � �1 � �Yr � �A � �G � �IC � �RR&IAR � �Ins � �AnyDes � � (1) ADIS � �1 � �Yr � �A � �G � �IC � �RR&IAR � �Ins � �AnyDes � � (2) CAD � �1 � �Yr � �A � �G � �IC � �RR&IAR � �Ins � �AnyDes � � (3) Table 5 DIS, ADIS, CAD, and binary analog misconduct by designation, gender, employment status, and insurance and IAR licensure Independent variable � B-CAD � B-ADIS � B-DIS � CAD � ADIS � DIS Sample mean 0.12 0.15 0.24 0.75 0.86 1.10 Any designation 0.18 0.22 0.28 1.10 1.23 1.41 No designation 0.11 0.13 0.23 0.71 0.81 1.06 CFP 0.19 0.22 0.28 1.11 1.25 1.42 Non-CFP 0.11 0.14 0.23 0.72 0.82 1.07 ChFC 0.18 0.21 0.27 1.10 1.23 1.43 Non-ChFC 0.12 0.14 0.24 0.74 0.85 1.09 CFA 0.09 0.13 0.17 0.52 0.62 0.70 Non-CFA 0.12 0.15 0.24 0.75 0.86 1.11 Male 0.14 0.17 0.26 0.91 1.04 1.28 Female 0.05 0.07 0.17 0.29 0.34 0.57 Independent contractor 0.13 0.15 0.27 0.82 0.93 1.24 Employee 0.11 0.14 0.22 0.71 0.82 1.02 IAR 0.14 0.18 0.26 0.90 1.04 1.27 Non-IAR 0.08 0.10 0.20 0.56 0.64 0.90 Insurance license 0.16 0.20 0.29 1.02 1.17 1.41 No insurance license 0.06 0.07 0.17 0.39 0.44 0.68 CFP � Certified Financial Planner; ChFC � Chartered Financial Consultant; CFA � Chartered Financial Analyst; IAR � Investment Advisor Representatives. CAD is for Culpable Advisory Disclosure and measures regulatory or judicial findings or other strong indications of clear advisory-related misconduct. ADIS is for Advisory Disclosure Incidence Score and measures allegations and findings of advisory-related misconduct. DIS is for Disclosure Incidence Score and measures allegations and findings of advisory- and non-advisory–related misconduct. B-versions of DIS, ADIS, and CAD are binary transformations of the corresponding continuous variables. 282 J.M. Camarda / Financial Services Review 26 (2017) 271–290 The logit regression models specifications are: B-DIS � �1 � �Yr � �A � �G � �IC � �RR&IAR � �Ins � �AnyDes � � (4) B-ADIS � �1 � �Yr � �A � �G � �IC � �RR&IAR � �Ins � �AnyDes � � (5) B-CAD � �1 � �Yr � �A � �G � �IC � �RR&IAR � �Ins � �AnyDes � � (6) Where YR � Years registered A � Age G � Gender IC � Independent contractor RR&IAR � dual registration as RR and IAR Ins � RR is life insurance sales licensed Anydes � RR holds any of CFA, CFP, or ChFC designations 9. Regressions results As seen in Table 6, having at least one of the study designations is associated with lower misconduct by all three measures with high significance in the Tobit. Having an insurance Table 6 Tobit DIS, ADIS, and CAD variable coefficients, significance, and (standard errors) by years registered, age, gender, employment status, IAR and insurance licensure status, and presence of at least one study designation Independent variable DIS ADIS CAD Years registered 0.213 *** 0.449 *** 0.522 *** (0.009) (0.014) (0.017) Age 0.008 — �0.039 *** �0.038 *** (0.007) (0.012) (0.014) Gender 1.818 *** 3.542 *** 3.893 *** (0.155) (0.263) (0.317) Independent contractor 0.460 *** �0.343 — 0.030 — (0.137) (0.210) (0.246) IAR 0.841 *** 1.940 *** 1.921 *** (0.13) (0.21) (0.25) Insurance license 0.937 *** 2.805 *** 2.840 *** (0.146) (0.235) (0.279) Any designation �1.155 *** �0.984 *** �1.020 *** (0.191) (0.268) (0.311) IAR � Investment Advisor Representatives. *, **, *** left-adjacent coefficient significant at the less-than 10%, 5%, or 1% level, respectively. Associated standard error appears beneath the corresponding coefficient in parentheses. CAD is for Culpable Advisory Disclosure and measures regulatory or judicial findings or other strong indications of clear advisory-related misconduct. ADIS is for Advisory Disclosure Incidence Score and measures allegations and findings of advisory-related misconduct. DIS is for Disclosure Incidence Score and measures allegations and findings of advisory- and non-advisory–related misconduct. B-versions of DIS, ADIS, and CAD are binary transformations of the corresponding continuous variables. 283J.M. Camarda / Financial Services Review 26 (2017) 271–290 license, being male, and being dually registered as an IAR is associated with higher misconduct, also all with high significance. Please see the results section for more detailed discussion. The OLS results are presented in Table 7. Having a designation is associated with lower misconduct by all three measures with high significance in the OLS. Please see the results section for detailed discussion. Readers are reminded that because the data are bounded at zero, there is a non-normal error distribution, rendering OLS inappropriate as a primary regression technique because of possible Gauss-Markov violations. OLS is still a reasonable estimator for these data, and presented as a supplemental robustness check on the Tobit results. Logit regression results are presented in Table 8. Logit interpretations are different from typical regressions such as OLS, and generally focus on a measure called the odds ratio instead of regression coefficients. Essentially, the odds ratio gives the probability that a specific study IV is associated with a yes or no value for the misconduct DV. If an IV odds ratio is 0.25 for a misconduct measure of 1 or yes, then that misconduct is not likely to be present, and this should be interpreted as a good or 75% chance that a particular subject does not have this misconduct disclosed compared with a subject that lacks the categorical IV. Odds ratios lower than one are proportionately associated with less likelihood of the condition being present, and those higher than one with proportionately greater likelihoods; those close to one are analogous to zero coefficients in typical regressions. The odds ratios results here show a reduced probability of all three measures of misconduct associated with having at least one study designation, all with high significance. Probabilities of higher misconduct for all measures are associated with maleness, dual IAR registration, and Table 7 OLS DIS, ADIS, and CAD variable coefficients, significance, and (standard errors) by years registered, age, gender, employment status, IAR and insurance licensure status, and presence of at least one study designation Independent variable DIS ADIS CAD Years registered 0.077 *** 0.080 *** 0.076 *** (0.002) (0.002) (0.002) Age �0.003 * �0.005 *** �0.005 *** (0.002) (0.002) (0.002) Gender 0.456 *** 0.424 *** 0.372 *** (0.040) (0.037) (0.035) Independent contractor 0.046 — �0.077 ** �0.053 — (0.038) (0.036) (0.034) IAR 0.141 *** 0.134 *** 0.102 *** (0.036) (0.034) (0.032) Insurance license 0.050 — 0.053 — �0.001 — (0.040) (0.037) (0.035) Any designation �0.375 *** �0.292 *** �0.261 *** (0.055) (0.052) (0.049) IAR � Investment Advisor Representatives. *, **, *** left-adjacent coefficient significant at the less-than 10%, 5%, or 1% level, respectively. Associated standard error appears beneath the corresponding coefficient in parentheses. CAD is for Culpable Advisory Disclosure and measures regulatory or judicial findings or other strong indications of clear advisory-related misconduct. ADIS is for Advisory Disclosure Incidence Score and measures allegations and findings of advisory-related misconduct. DIS is for Disclosure Incidence Score and measures allegations and findings of advisory-and non-advisory–related misconduct. B-versions of DIS, ADIS, and CAD are binary transformations of the corresponding continuous variables. 284 J.M. Camarda / Financial Services Review 26 (2017) 271–290 insurance licensure. Please see the results section for detailed discussion. Logit is the primary robustness check as unlike for OLS, the data can be made appropriate to the Logit technique via binary transformations. Logit odds ratios (�) as the probability of misconduct are discussed, and � plots are provided in Figs. 3 (B-DIS), 4 (B-ADIS) and 5 (B-CAD). As seen in Fig. 3 for B-DIS, the odds of misconduct go down with a study designation, but meaningfully up for dual RR/IARs, insurance licensure, and for maleness, with the latter quite pronounced. Fig. 4, the odds ratios for B-ADIS, shows the probability of misconduct goes down with a study designation, but substantially up for dual RR/IARs, insurance licensure, and for maleness, with the latter two quite pronounced. In Fig. 5, for B-CAD, the odds of misconduct go down with a study designation, but substantially up for dual RR/IARs, insurance licensure, and for maleness, with the latter two quite pronounced. As seen in these three figures, Logit results are very consistent for each misconduct measure; please see results section for detailed discussion. 10. Results discussion, summary, and conclusions DIS is this study’s broadest measure of misconduct disclosure, including all items both advisory and non-advisory–related and including allegations as well as findings of miscon- duct. Having any study designation is associated with lower DIS in all of the controlled Table 8 Logit B-DIS, B-ADIS, and B-CAD variable coefficients, significance, odds ratios (�), and (standard errors) by years registered, age, gender, employment status, IAR and insurance licensure status, and presence of at least one study designation Independent variable B-DIS Psi B-ADIS Psi B-CAD Psi Years registered �0.047 *** 0.954 �0.083 *** 0.920 �0.090 *** 0.914 (0.002) (0.003) (0.003) Age �0.001 — 0.999 0.008 *** 1.008 0.007 *** 1.007 (0.002) (0.002) (0.003) Gender 0.213 *** 1.532 0.351 *** 2.018 0.363 *** 2.065 (0.019) (0.027) (0.030) Independent contractor 0.041 ** 1.084 �0.067 *** 0.874 �0.033 — 0.935 (0.016) (0.020) (0.022) IAR 0.121 *** 1.274 0.217 *** 1.544 0.203 *** 1.502 (0.016) (0.021) (0.023) Insurance license 0.137 *** 1.315 0.314 *** 1.875 0.293 *** 1.796 (0.017) (0.023) (0.026) Any designation �0.129 *** 0.773 �0.084 *** 0.846 �0.079 *** 0.854 (0.022) (0.025) (0.027) IAR � Investment Advisor Representatives. *, **, *** left-adjacent coefficient significant at the less-than 10%, 5%, or 1% level, respectively. Associated standard error appears beneath the corresponding coefficient in parentheses. CAD is for Culpable Advisory Disclosure and measures regulatory or judicial findings or other strong indications of clear advisory-related misconduct. ADIS is for Advisory Disclosure Incidence Score and measures allegations and findings of advisory-related misconduct. DIS is for Disclosure Incidence Score and measures allegations and findings of advisory-and non-advisory–related misconduct. B-versions of DIS, ADIS, and CAD are binary transformations of the corresponding continuous variables. 285J.M. Camarda / Financial Services Review 26 (2017) 271–290 regressions with high significance. It should be noted, however, that RRs who are male, have an insurance license, or are also registered as IARs are associated with higher DIS scores. The negative influence of the non-designation factors appears to exert a strong influence on misconduct scores, even for those with designations. It seems noteworthy that being male or Fig. 3. B-DIS and AnyDes logit odds ratio graph. Fig. 4. B-ADIS and AnyDes logit odds ratio graph. 286 J.M. Camarda / Financial Services Review 26 (2017) 271–290 also an IAR is associated with higher DIS misconduct to p � 0.0001 in all three tests, and that having an insurance license also is with the same p value in both Tobit and Logit. Exploration of non-designation effects is left for subsequent research. For DIS and B-DIS, the findings suggest that the null hypothesis H0: �AnyDes � 0 can be rejected with a high degree of confidence, with associated p values less than 0.0001 in all three regression tests. Consequently, lower DIS misconduct seems associated with having at any least one of the study designations. The AnyDes ADIS and B-ADIS findings are similar: the null that associated misconduct does not change when any study designation is present can be rejected with high confidence, in this case with p � 0.001 for Tobit and Logit, and OLS p � 0.0001. The AnyDes CAD and B-CAD findings are generally similar to those for DIS and ADIS. Recall that CAD is the study’s most serious misconduct measure, limited to advisory items only where misconduct is not merely alleged but nearly certain. For the CAD DVs, the reduction for AnyDes is quite robust, with p � 0.001 for Tobit, �0.004 for Logit, and �0.0001 for OLS. Once again, the null that misconduct does not change when any study designation is present can be rejected with high confidence. As for DIS and ADIS, similar strongly higher misconduct associations are seen for the CADs for gender (all ps � 0.0001), IAR (highest p is from OLS at �0.002) and with insurance licensure from Tobit and Logit (both ps � 0.0001; but note that OLS finds no significance for insurance). This study’s results support a strong and robust finding that having at least any one of the study designations is associated with lower misconduct. This is found with high statistical significance using the three different regression techniques. These findings are uniform across the severity spectrum from DIS to CAD, when controlling factors are considered. Both allegations and findings for advisory and non-advisory misconduct matters fall Fig. 5. B-CAD and AnyDes logit odds ratio graph. 287J.M. Camarda / Financial Services Review 26 (2017) 271–290 uniformly for those in the AnyDes category. For Florida RRs at least, holding at least one of the study designations seems to be a strong and reliable quality signal, consistent with the marketing claims of the various conferring institutions, in which we can have confidence. It should be noted that this finding is not consistent with those from the descriptive statistics section, where higher misconduct for CFPs and ChFCs was found. This may be because of the confounding influence of misconduct-increasing factors such as insurance licensure, IAR registration, and maleness. The fact that all three of these characteristics are associated with higher misconduct affirms the descriptive statistics findings, and they all appear more influential than the designation effect. It is possible that these compensation conflicts somewhat outweigh the good that designations may confer. The temptation of high and poorly disclosed insurance or securities commissions, perhaps misleadingly obtained under a false IAR or designation fiduciary flag where consumers are allowed to believe their interests are put first, may influence or even enable the designee to act less ethically than implied by the signal, inducing complaints and other misconduct markers. This may be one explanation for the descriptive statistics finding that misconduct rises for ChFCs and CFPs over non-designees, but the question is left for further research, including a study of misconduct by specific designation now underway by this author. The discovered gender effect has now been confirmed by subsequent research (Egan, Matvos, & Seru, 2017), but an exploration of its likely complex causes is deferred to later study. References Barry, R., & Eaglesham, J. (2014). Regulatory Breakdowns in Oversight of US Stockbrokers. Wall Street Journal [Whitepaper]. Brockman, C., & Brooks, R. (1998). The CFA charter: Adding value to the market. Financial Analysts Journal, 54, 81–85. CFA Institute. (2015). Code of Ethics and Standards of Professional Conduct. (available at http://www. cfainstitute.org/ethics/codes/ethics/Pages/index.aspx) CFA Institute. (2016). Value of the CFA Charter. (available at https://www.cfainstitute.org/programs/cfaprogram/ charter/value/pages/index.aspx) CFP Board. (2015). 2015 Annual Brand Tracking Report: Executive Summary. (available at http://www. letsmakeaplan.org/docs/default-source/default-document-library/2015-annual-brand-tracking-report-executive- summary.pdf) CFP Board. (2016). CFP Certification: The Standard of Excellence. (available at http://www.cfp.net/about-cfp- board/cfp-certification-the-standard-of-excellence) CFP Board Standards of Professional Conduct. (2008/2014). (available at cfp.net/docs/ethics-enforcement/ standards-of-professional-conduct-final.pdf?sfvrsn�2) Cordell, D., Smith, R., & Terry, A. (2011). Overconfidence in financial planners. Financial Services Review, 20, 253–263. Cummings, B., & Finke, M. (2010). The Economics of Fiduciary Investment Advice (SSRN Working Paper 1701181). (available at ssrn.com/abstract�1701181) Dean, P. (1997). Examining the profession and practice of business ethics. Journal of Business Ethics, 16, 1637–1649. Egan, M., Matvos, G., & Seru, A. (2016). The Market for Financial Adviser Misconduct. (available at http:// www.nber.org/papers/w22050) FINRA. (2016). Registered Representatives Brochure. (available at http://www.finra.org/sites/default/files/ registered-representatives-brochure.pdf) 288 J.M. Camarda / Financial Services Review 26 (2017) 271–290 Financial Planning Coalition. (2014). Consumers are Confused and Harmed: The Case for Regulation of Financial Planners. (available at http://financialplanningcoalition.com/wp-content/uploads/2014/06/Financial- Planning-Coalition-Regulatory-Standards-White-Paper-Final.pdf) Finke, M., & Langdon, T. (2012). The impact of the broker-dealer fiduciary standard on financial advice. Journal of Financial Planning, 25, 28–37. Flexner, A. (1976). Is social work a profession? Proceedings of the National Conference of Charities and Corrections. (available at http://scholar.google.com/scholar?hl�en&q�Is�Social�Work�a�Profession� flexner�and�metzger�college�and�university�1976&btnG�&as_sdt�1%2C10&as_sdtp�) Flexner, A., & Metzger, W. (1976). Is social work a profession? What is a profession? College and University, 52, 42–55. Frumento, A., & Korenman, S. (2013). Professionalism and investment advisors. Journal of Investment Com- pliance, 14, 32–41. Government Accountability Office. (2011). Consumer Finance: Regulatory Coverage Generally Exists for Financial Planners, but Consumer Protection Issues Remain (GAO-11–235). Washington, DC: Government Printing Office. Hauptman, M., & Roper, B. (2017). Financial Advisor or Investment Salesperson? Washington, DC: Consumer Federation of America. Khurana, R., & Nohria, N. (2008). It’s time to make management a true profession. Harvard Business Review, 86, 70–77. Khurana, R., Nohria, N., & Penrice, D. (2005). Is Business Management a Profession? Harvard Business School website. (available at hbswk.hbs.udu/item/4650.html) Kitces, M. (2014b). The Rise of the Financial Planning Academic. (available at kitces.com/blog/the-rise-of-the- financial-planning-academic/) Kitces, M. (2015). CFP Board’s Reduced Experience Requirement Quietly Takes Effect With FPA And NAPFA Silence As Tacit Support? (available at https://www.kitces.com/blog/cfp-boards-reduced-experience- requirement-quietly-takes-effect-with-fpa-and-napfa-silence-as-tacit-support/) Kitces, M. (2015b). Are the Licensing and Other Requirements to Become a Financial Advisor Too Easy? (available at https://www.kitces.com/blog/are-the-licensing-and-other-requirements-to-become-a-financial- advisor-too-easy/) Mauldin, D., Wilder, W., & Stocks, M. (2000). Does AICPA accreditation of nonaudit services add value? The case of personal financial planning. Accounting Horizons, 14, 49–67. McCann, C., Qin, C., & Yan, M. (2016). How Widespread and Predictable is Stock Broker Misconduct? (available at Securities Litigation and Consulting Group: http://papers.ssrn.com/sol3/papers.cfm?abstract_id�2768942) Moisand, D. (2008). The Financial Planning Act of 2008. Journal of Financial Planning, 21, 78–90. Moisand, D. (2012). So close and yet so far. Journal of Financial Planning, 25, 26–27. PIABA Foundation. (2015). PIABA BrokerCheck Letter to SEC re File 265–28. (available at https://www. sec.gov/comments/265-28/26528-98.pdf) Ragatz, J., & Duska, R. (2010). Financial codes of ethics. In J. Boatright (Ed.), Finance Ethics: Critical Issues in Theory and Practice. New York, NY: J. Wiley and Sons. Raymond James Mission Statement. (2016). Mission Statement. (available at http://www.raymondjames.com/ about/mission.htm) Ritholtz, B. (2016). Brokers Behaving Badly. (available at https://www.bloomberg.com/view/articles/2016-03- 02/broker-misconduct-is-worse-than-we-thought) Schaefer, T. (1984). Professionalism: Foundation for business ethics. Journal of Business Ethics, 3, 269. Smith, R., Vibhakar, A., & Terry, A. (2008). Demarcating designations: Chartered Financial Analyst and Certified Financial Planner. Journal of Financial Services Marketing, 12, 299–310. Spence, M. (1973). Job market signaling. The Quarterly Journal of Economics, 87, 355–374. Spence, M. (2002). Signaling in retrospect and the informational structure of markets. The American Economic Review, 92, 434–459. Terry, A., & Vibhakar, A. (2011). A comparative analysis of the CFA and CFP designations. Financial Services Review, Fall, 66–81. 289J.M. Camarda / Financial Services Review 26 (2017) 271–290 The American College. (2016b). ChFC: Chartered Financial Consultant. (available at https://www.theameri cancollege.edu/designations-degrees/ChFC) The American College. (2016a). ChFC: The Highest Standard. (available at http://www.chfchigheststandard. com/) The American College. (2016c). Code of Ethics. (available at https://www.theamericancollege.edu/about-the- college/community-responsibility/code-of-ethics) The White House. (2016). The Effects of Conflicted Investment Advice on Retirement Savings. Washington, DC: Government Printing Office. VyStar Ad. (2016). Florida Times-Union, B-8. (http://jacksonville.com/news/metro/2017-08-09/morris-publishing- sells-florida-times-union-gatehouse-media) Warschauer, T. (2002). The role of universities in the development of the personal financial planning profession. Financial Services Review, 11, 201–216. Woolley, S. (2016). It Just Got Harder to Trust Financial Advisors. (available at https://www.bloomberg.com/ news/articles/2016-03-01/it-just-got-even-harder-to-trust-financial-advisers) Zingales, L. (2015). Does finance benefit society? The Journal of Finance, 70, 1327–1363. 290 J.M. Camarda / Financial Services Review 26 (2017) 271–290