Introduction
As children, we tend to see ethics in simple, unequivocal terms. There are good people and bad people. Good people do good things and bad people do bad things. Children's books, movies and television usually reinforce this dichotomy of the world. As we grow older, we become more comfortable with the complexity of human nature, with its mix of positive and negative characteristics.
As a behavioural scientist, I research how and why good people – especially professional advisors – do bad things. Over and over again, I see good advisors striving to do their very best but nonetheless doing harm. The clear conclusion is that their ethicality is not a stable trait, but a dynamic, malleable variable. In addition to our moral identity (Aquino & Reed II, Reference Aquino and Reed II2002), the ethicality of our behaviour can be influenced by aspects of the situation (Ross & Nisbett, Reference Ross and Nisbett1991), our individual circadian rhythms and the time of day (Gunia et al., Reference Gunia, Barnes and Sah2014) and other seemingly irrelevant factors (Bazerman & Tenbrunsel, Reference Bazerman and Tenbrunsel2011), often in predictable, systematic ways.
Advisors, such as physicians, financial advisors, lawyers and accountants, often face a conflict of interest – that is, a clash between their professional and personal interests. Such conflicts can lead to biased and corrupt advice. For example, some of the accounting scandals in recent years are the result of accounting firms having the dual role of auditing clients' accounts while simultaneously collecting large consulting fees from those clients (Moore et al., Reference Moore2006). Indeed, the financial ties between the then-prestigious auditing firm Arthur Andersen and its client, Enron, contributed to one of the most notorious corporate ethics scandals in history and the downfall of both corporations. Even absent dual revenue streams, auditors charged with delivering fair, honest audits also have a motivation to please their clients thanks to their desire to be rehired for future auditing services. Similarly, credit rating agencies that evaluated mortgage-backed securities were hired and fired by the same firms whose bonds they were rating. Furthermore, concern has grown that health care professionals face comparable conflicts of interest because of industry payments to physicians and fee-for-service compensation arrangements (Sah & Fugh-Berman, Reference Sah and Fugh-Berman2013).
Although bad apples do exist and some people deliberately choose to act unethically, the bias created by conflicts of interest can influence anyone and often manifests in subtle ways, producing implicit and unintentional unethical behaviour. In this paper, I focus on how conflicts of interest can cause good people to unintentionally cross ethical boundaries and how we can avoid succumbing to this bias.
Imagine that each of us has a literal ‘moral compass’ that guides our ethical behaviour.Footnote 1 As with all working compasses, the needle always points north. If we go astray, it might be because we ignored our compass, or our compass malfunctioned. To illustrate the former case, Enron's former Chief Financial Officer, Andrew Fastow, asserted, “I lost my moral compass, and I've done terrible things that I very much regret” (Levine, Reference Levine2006). In the latter case, our compass may be broken without us even knowing it. Situational factors can act as a magnet that sways the compass needle away from north (Lessig, Reference Lessig2013; Moore & Gino, Reference Moore and Gino2013). For example, incentives that create conflicts of interest act as a magnet that deviates the needle, which could deter someone from his or her primary professional goal, such as delivering an impartial audit. Because we may not even be aware of this magnetic deviation, we may remain convinced that the needle on our moral compass faces north and that we are behaving ethicality. Due to the pervasive belief that succumbing to a conflict of interest is a matter of deliberate corruption, rather than unintentional bias, many professionals deny and remain unaware that they could be, and likely are, inappropriately influenced by conflicts of interest.
In this paper, I first discuss two mechanisms through which advisors may convince themselves to accept conflicts of interest while vehemently believing that they remain unbiased: (1) a sense of entitlement; and (2) a sense of invulnerability created by their belief in their own professionalism (Sah & Loewenstein, Reference Sah and Loewenstein2010; Sah & Larrick, Reference Sah and Larrick2017). I then discuss several proposed policy solutions for managing conflicts of interest, such as training, sanctions, second opinions and disclosure policies. Current policies dealing with conflicts of interest are largely based on inaccurate intuitions regarding the psychological processes that underlie them; consequently, these proposed solutions tend to fail or have unintended consequences. I then focus on solutions that are more likely to be successful, such as identifying and changing the professional norms that exert the most powerful influence on employee behaviour (Sah, Reference Sah2017). I typically draw on examples from the USA, although due to the focus on psychological behaviour, the nature of the evidence is transferable to other countries.
Rationalisations that increase the acceptance of conflicts of interest
People often view themselves as ethical, competent and moral (Aquino & Reed II, Reference Aquino and Reed II2002). This view often prevents them from recognising their own biases and conflicts of interests. Our strong desire to appear ethical to ourselves and to others may prevent us from engaging in behaviour that is obviously unethical, but it does not prevent us from rationalising behaviour that is more self-serving that we would care to admit. Two justifications – a sense of entitlement and a sense of invulnerability – manifest themselves in predictable, systematic ways and unconsciously liberate professional advisors to engage in, and accept, self-serving behaviour.
Sense of entitlement: because I'm worth it
Professionals sometimes feel entitled to accept gifts from third parties that could be construed as potential bribes. For example, office managers often accept samples and gifts from sales representatives who want to supply the company, and physicians may accept gifts from pharmaceutical companies or medical device manufacturers. Similarly, in the world of corporate hospitality, organisations provide potential clients with a steady stream of exclusive tickets to private boxes, suites or premium seats at top sporting events, concerts and other entertainment venues. Employees often view such gifts as a perk of their job. Indeed, interactions between physicians and industry are a cultural norm, which often starts early in physicians’ medical careers. One survey found that virtually all (94%) physicians have some type of relationship with industry (Campbell et al., Reference Campbell2007).
Although different rationalisations may psychologically enable advisors to accept conflicts of interest, a remarkable one is their sense of entitlement. For example, physicians use a variety of justifications to accept industry gifts that outsiders might view as barely disguised bribes (Chimonas et al., Reference Chimonas, Brennan and Rothman2007). Some physicians believe that they deserve these gifts because of the hard work and sacrifices they have made to become physicians. A study of third-year medical students found that such rationalising begins even before students become physicians: over 80% of the students believed that they were entitled to pharmaceutical industry gifts due to hardships of ‘considerable debt and minimal income’ (Sierles et al., Reference Sierles2005).
Sah and Loewenstein (Reference Sah and Loewenstein2010) conducted a randomised experiment to examine whether subtle reminders designed to make new physicians feel deprived could increase their willingness to accept industry gifts and other incentives. A control group of physicians answered a series of questions relating to conflicts of interest in order to gauge the acceptability of receiving gifts from industry. Just 22% of physicians in this group found industry gifts acceptable. Another group, prior to answering the conflict of interest questions, answered questions regarding the sacrifices they had made during their medical training (hours of sleep, debt incurred, working conditions, etc.). For this physician group, 48% found industry gifts acceptable. In a third condition, physicians, again prior to being asked the conflict of interest questions, answered the sacrifice questions, but then also responded to the following explicated rationalisation: “Some physicians believe that the stagnant salaries and rising debt levels prevalent in the medical profession justify accepting gifts and other forms of compensation and incentives from the pharmaceutical industry. To what extent do you agree or disagree that this is a good justification?” Interestingly, while a majority of these physicians disagreed with this rationalisation, when they answered the conflict of interest questions, 60% now said that industry gifts were acceptable. Thus, even rationalisations that are consciously rejected may still help justify behaviour that professionals would otherwise find unacceptable.
Sense of invulnerability: self-perceived professionalism
A sense of entitlement, although powerful, does not tell the full story of why professionals accept conflicts of interest. Another complementary and instrumental reason is that professionals, like most individuals, usually believe that they are immune to being influenced by conflicts of interest, gifts or other irrelevant factors. In other words, professionals often believe that such conflicts will not affect their ability to make impartial decisions. Indeed, professionals commonly cite their ‘professionalism’ as a primary argument against being affected by conflicts of interest. In 2000, CEOs and managing partners of large accounting firms, including Arthur Andersen and PricewaterhouseCoopers, while testifying before the Securities and Exchange Commission, argued that the professionalism and objectivity of professional auditors resolved the issue of auditor independence (US Securities and Exchange Commission, 2000). Similarly, physicians may concede that their colleagues could be susceptible to being influenced by industry, but believe that they themselves are personally invulnerable (McKinney et al., Reference McKinney1990). Although professionals may believe that small gifts will not impact their decision making, research on influence and reciprocity robustly demonstrates the contrary (Cialdini, Reference Cialdini2006; Sah & Fugh-Berman, Reference Sah and Fugh-Berman2013). People are unable to remain objective, even when motivated to be impartial (which suggests an unintentional element of bias), and people succumb to bias while concurrently denying that they have done so (Babcock & Loewenstein, Reference Babcock and Loewenstein1997).
In a series of experiments, Sah and Larrick (Reference Sah and Larrick2017) examined how beliefs about one's professionalism contribute to biased judgements in the presence of conflicts of interest. People, perhaps unsurprisingly, reported that others would be more susceptible to being influenced by accepting gifts than they would be. In fact, the managers who viewed themselves as more professional and less likely to be swayed by gifts from third parties reported that they were actually more likely to accept these gifts. Critically, these managers were also more likely to become biased by the conflict of interest, while simultaneously being less likely to register the bias in their judgements. In other words, the more ‘professional’ people feel they are (i.e. the greater they believe their ability to make objective decisions and self-regulate the influence of small gifts to be), the more likely they are to accept gifts and be influenced by those gifts. All the while, they remain unaware of the influence and their bias.
These results demonstrate that it is precisely those who believe that they are the least susceptible to outside influence (i.e. the most ‘professional’) who are in fact the most likely to accept conflicts of interest and demonstrate increased bias. Clearly, the problem of conflicts of interest is not always due to deliberate corruption, but often arises from unintentional bias. A belief in one's high level of professionalism does not protect one from unconscious bias and may actually make matters worse.
Popular policy responses
Education and training
In an attempt to ‘de-bias’ people, Babcock and Loewenstein (Reference Babcock and Loewenstein1997) educated study participants about self-serving biases. While participants' awareness of bias increased, they still believed, often more strongly, that others were susceptible to self-serving biases and that they themselves were not. This belief in our resistance to bias highlights not only that such biases operate unconsciously, but also that simply educating professionals about bias is unlikely to be successful.
Even if education or ethics training is incorporated at the start of employment, people often engage in a process called ‘ethical fading’ (Tenbrunsel & Messick, Reference Tenbrunsel and Messick2004). Ethical fading is a phenomenon in which difficult moral issues are pushed to the background or removed altogether from a decision-making process, while more straightforward, objective business issues come to the forefront. Such fading means that ethics training or codes of conduct cannot protect professionals from succumbing to bias from conflicts of interest. The fact that Enron had a 64-page ‘Code of Ethics' booklet serves as a powerful exemplar that culture has a stronger impact on moral decision making than codes.
Sanctions
Another popular response to conflicts of interest is to impose penalties on those who engage in corruption or fraud in order to discourage such behaviour in the future. For example, the US Government levies large fines on pharmaceutical companies that promote their products for uses not approved by the US Food and Drug Administration. Likewise, companies that pollute excessively often face fines or penalties. Unfortunately, introducing fines can worsen rather than improve employees' ethical behaviour by encouraging them to categorise a decision as a cost–benefit business decision, rather than as an ethical decision (Tenbrunsel & Messick, Reference Tenbrunsel and Messick1999, Reference Tenbrunsel and Messick2004). Indeed, some critics argue that the size of such penalties typically is not large enough to outweigh the benefits of self-interested behaviour, and that companies will simply view the fines as the ‘cost of doing business’ (Gagnon & Lexchin, Reference Gagnon and Lexchin2008; Feldman, Reference Feldman2016).
Another flaw in the reasoning behind imposing financial sanctions on undesirable behaviour is that it assumes people decide to succumb to bias after engaging in a conscious and careful cost–benefit analysis. Although this sometimes happens – Ford Motor Company infamously compared the costs of fixing defective Pintos to the benefits of avoiding lawsuits for deaths and injuries (Gioia, Reference Gioia1992) – more often people succumb to bias unconsciously, thus limiting the effect of penalties on behaviour.
Additionally, because it is difficult to verify the existence and effect of bias from conflicts of interest, it is unclear when sanctions are warranted. In medicine, for example, multiple investigation and treatment options sometimes exist, and it can be difficult to determine which advice is best and least biased (Sah, Reference Sah2015; Sah et al., Reference Sah2016). In addition, it is not always the case that biased behaviours have negative results, or that unbiased behaviours have positive results. To complicate matters further, often the negative effects of biased advice are not realised until years later, after the professionals who made harmful decisions have left their positions. For example, it was years after the tobacco industry paid physicians to promote the supposed health benefits of cigarettes that the causal link between smoking and lung cancer was accepted. In sum, given the difficulty of proving bias in a timely fashion and the limited impact of legal threats and penalties, it is not surprising that sanctions are rarely effective at mitigating unintended bias.
Second opinions
Second opinions are often proposed as an antidote to potential bias caused by a primary advisor's conflict of interest. For example, financial journalist Ross Sorkin described the most talked-about idea at an annual corporate attorney conference: “If a company's advisor has a conflict, the best practice for a corporate board is to hire a second advisor to help cleanse the first advisor” (Ross Sorkin, Reference Ross Sorkin2012). But Sah and Loewenstein (Reference Sah and Loewenstein2015) found that second opinions may not be the panacea they appear to be. In this research, while clients who received a high-quality second opinion made better decisions than clients who did not, a curious thing happened when primary advisors were aware that their clients might seek a second opinion. These primary advisors adopted a profit-maximising frame and gave even more biased advice. Further, the more costly it was for clients to seek a second opinion, the less likely they were to do so, and the more biased their primary advisors became.
These findings suggest that the percentage of clients who opt to pursue second opinions is critical to determining whether they benefit from doing so. Second opinions are rarely mandatory, and clients do not tend to seek them frequently for a number of reasons, including cost, effort, time and in order to avoid the possibility of offending their primary advisors. The unexpected impact of second opinions on primary advisors' behaviour could therefore have substantial negative consequences for their clients.
Conflict of interest disclosures
Disclosure is one of the most commonly proposed and implemented solutions for dealing with conflicts of interest across a range of industries and professions. For example, the American Medical Association states in its code of ethics that physicians should disclose to their patients if they receive referral fees for enrolling their patients in clinical trials, and the US Securities and Exchange Commission requires registered investment advisors to disclose any commissions if they refer clients to particular solicitors, brokers or other agents.
Disclosure is popular for a variety of reasons. At least in theory, it provides useful information to the advisee and involves minimal disruption of the status quo. However, a number of unintended consequences of conflict of interest disclosures have been documented, including advisors giving even more biased advice with disclosure than without (Cain et al., Reference Cain, Loewenstein and Moore2011), and advisees reporting increased trust in advisors who disclose bias or conflicts (Sah et al., Reference Sah, Fagerlin and Ubel2016). In addition, advisees who are informed of their advisors' conflict of interest sometimes feel increased pressure to comply with their advisors' recommendations, despite trusting the advice less than advisees who are not informed of their advisors' conflicts (Sah et al., Reference Sah, Loewenstein and Cain2013, Reference Sah, Loewenstein and Cain2015). These social pressures (desires to avoid appearing unhelpful or untrusting) to comply with advice accompanied by a conflict of interest disclosure are reduced when interactions between advisors and advisees are not face to face (i.e. in private) (Sah et al., Reference Sah and Fugh-Berman2013). Even when social pressure is low, however, advisees may not react optimally to disclosures. Advisees could ignore disclosures if they do not know what to do with the information or believe the information is not important to their decision making (Hampson et al., Reference Hampson2006), or advisees could discount too much and discard valuable advice that is accompanied by a conflict of interest disclosure (Sah & Feiler, Reference Sah and Feiler2017). Although transparency is desirable in many situations, disclosure is not a panacea to managing conflicts of interest.
Eliminate the conflict of interest
Perhaps the best way to mitigate a conflict of interest is to eliminate it, either by extracting oneself from the situation or realigning incentives to remove the conflict of interest. Prior research has shown that advisors who are able to reject or accept a conflict of interest are more likely to reject it if they are aware that they will have to disclose the conflict to their advisees (Sah & Loewenstein, Reference Sah and Loewenstein2014). In many instances, however, rejecting the conflict of interest is not possible. Physicians, for example, cannot change the fee-for-service model, which is the predominant form of physician reimbursement in the USA, and financial advisors have limited input in determining their incentive compensation structure. Striving for independence and eliminating conflicts of interest whenever possible are desired goals, but innate barriers ultimately exist that make abolishing all conflicts of interest impossible.
Professional norms and the appropriateness of bias
The policy solutions presented above paint a depressing picture: overall, there are few clearly effective and inexpensive solutions available to reduce conflicts of interest in our society. However, reason for optimism emerges when we consider that people do not always act in ways that are consistent with their own self-interest. In fact, people often comply with ethical principles even when doing so incurs real costs or sacrifices opportunities for gain. This compliance regularly occurs if professional norms dictate it.
Norms are understood rules for accepted and expected behaviour. Injunctive norms, or prescriptive norms – those that are socially sanctioned and commonly approved of – reliably predict behaviour (Cialdini et al., Reference Cialdini, Bator and Guadagno1999) and are closely related to the ‘logic of appropriateness' perspective, which March (Reference March1994) proposed to describe the role of context in organisational decision making. The logic of appropriateness “sees human action as driven by rules of appropriate or exemplary behaviour, organized into institutions. Rules are followed because they are seen as natural, rightful, expected and legitimate” (March & Olsen, Reference March and Olsen2004, p. 2).
These moral norms do not depend upon cost–benefit analyses or anticipation of future consequences. Instead, people take reasoned action by answering (implicitly or explicitly) three questions: (1) recognition – ‘what kind of situation is this?’; (2) identity – ‘what kind of person am I?’; and (3) rules: ‘what does a person such as myself do in a situation like this?’ (March, Reference March1994; March & Olsen, Reference March and Olsen2004). Internalised prescriptions, or ‘scripts', of what one ought to do determine the appropriate behaviour.
Professional advisors, across a wide variety of settings (medicine, auditing, finance, law, etc.), learn norms that define their professional roles. We believe that medical doctors care for their patients and ‘do no harm’, and we hope that registered investment advisors uphold their fiduciary duty to place their clients' interests above their own. However, professional norms are often learned through socialisation practices within the institution in which one works. In some institutions, advisors' primary roles may be to make money for their institutions, rather than to serve their clients (Friedman, Reference Friedman2007). Indeed, bankers or other sales people may be socialised into a culture in which they can only succeed by offloading stocks or goods, regardless of their worth or benefit to the client (Cohan, Reference Cohan2016).
Occupations are characterised by commonly perceived standards of ethical behaviour. One study demonstrated that when bankers' professional identities were activated – by reminding them that they were indeed bankers – they were more likely to behave unethically than when their professional identities were not activated (Cohn et al., Reference Cohn, Fehr and Maréchal2014). In a series of studies, Sah (Reference Sah2017) demonstrated that advisors who were required to disclose a conflict of interest reacted differently depending on the perceived professional norms of the context. When advisors, subject to real monetary conflicts of interest, were informed that they were giving financial advice, the advisors who had to disclose their conflicts of interest gave more biased advice, and reported that doing so was more appropriate, than those who did not have to disclose their conflicts. However, when advisors, subject to the same financial conflicts of interest and monetary payoffs, were informed that they were giving medical advice, the advisors who had to disclose now gave significantly less biased advice, and reported it was significantly less appropriate to give biased advice, than those advisors who were not required to disclose. The conflict of interest disclosure reminded advisors of their professional ethical dilemma and served as a signal that triggered prescribed professional norms for advisors – whether to give self-interested advice or put the advisee first. Disclosure prompted advisors in the financial context to prioritise their own self-interest, whereas advisors in a medical context prioritised the patient.
These findings demonstrate the powerful effect of activating perceived professional norms and expectations on behaviour. This is an important concept as it suggests that advisors' perceptions of institutional norms encourage them to either indulge in self-interest and bias or prioritise their advisees. Perceptions of an ethical organisational culture can therefore serve as a powerful defence against conflicts of interest. Even when legal regulations are weak or unenforceable, or awareness of bias is low, if professionals are reminded of relevant ethical norms, they may comply with standards simply because putting the advisee first is the right thing to do.
Thus, one way to improve professional advisors' ethical behaviour is to activate feelings of obligation and responsibility towards advisees. In contexts with strong ethical norms, conflict of interest disclosure may do so inherently (Sah, Reference Sah2017). In other contexts, changing professional norms and the ethical climate from the top down will be necessary (Mayer et al., Reference Mayer2009). Implementing situational cues to behave ethically, such as with honour codes or religious reminders (Mazar et al., Reference Mazar, Amir and Ariely2008), may increase the ethicality of decisions. In addition, perspective taking or imagining the client is a relative you know well and love, as well as giving advice on a one-to-one basis, tend to increase empathy towards advisees, helping to reduce bias in advice (Sah & Loewenstein, Reference Sah and Loewenstein2012). However, these changes must be aligned with other policies that encourage ethical behaviour. For example, fiercely demanding sales goals may pressure employees to behave unethicality, even when ethical codes or reminders are in place (Schweitzer et al., Reference Schweitzer, Ordonez and Douma2004; Cohan, Reference Cohan2016).
Once ethical motivations are activated, cost–benefit analyses and calculations of self-interest are of little importance in ethical decisions (Tyler, Reference Tyler, Moore, Cain, Loewenstein and Bazerman2005). If an action is viewed as appropriate, right, moral and just, professionals will simply comply. Due to the billions of dollars lost due to poor retirement advice in the last decade, in 2016, President Obama signed an Executive Order mandating that retirement investment advisors put their ‘clients first’ above profit (Arnold & Geewax, Reference Arnold and Geewax2016). We have yet to see if this move will have an impact, but this type of focus on ethical values and professional norms should create a threshold assessment of what is right, rather than what is most profitable for the institution.
The professional norms and appropriateness of bias perspective suggest that when conflicts of interest cannot be eliminated, the amount of bias caused by such conflicts can be dramatically reduced by activating the advisor's ethical values. Once activated, advisors will frame the decision of how to behave as an ethical one, and self-interested concerns will diminish in importance. Conflict of interest disclosure can activate ethical values, but only when strong ethical norms are in place and professionals have concern for their clients. In other words, policies, such as disclosure, must consider the culture and professional norms of each institution. Transparency alone is not enough; professional norms must also put clients first.
