Conflicts of Interest

INTRODUCTION 

Image: Giotto di Bondone, “Kiss of Judas,” via Wikimedia Commons

Bribery, kickbacks, and other clear-cut forms of corruption are serious problems for genuine bad apples; however, much of the problem with conflicts of interest is not intentional corruption but unintentional bias. Bias is widespread and is a problem even for well-meaning professionals (Cain & Detsky, 2008; Moore, Tanlu, & Bazerman, 2010). Human reasoning is easily pressed into the service of one’s own interests. For example, and as a general matter, whenever a person can reap rewards for recommending a particular course of action, he or she is more likely to recommend that action, even while honestly (but incorrectly) believing that he or she has acted objectively.

Disclosure is often proposed as a solution to conflicts of interest, but research finds that disclosure is often ineffective. Years of research on the “anchoring bias” (Tversky & Kahneman, 1974) suggest that once bad advice is let out of the bag, its impact on judgment is difficult to undo. Indeed, disclosure may even have perverse effects and can sometimes make matters worse. For example, disclosure can make advisors feel free to give worse (i.e., more biased) advice because advisees “have been warned” (Cain, Loewenstein, & Moore, 2011). Also, disclosure can pressure advisees into satisfying the advisors’ disclosed interests, because these interests are now common knowledge and are begging to be satisfied, just as a panhandler puts pressure on passersby to donate (Sah, Loewenstein, & Cain, 2013).

No one is arguing against transparency, but why does disclosure remain so popular as the primary remedy for conflicts of interest? One reason is that disclosure is cheap. It requires little substantive change. For those with financial conflicts of interest, disclosing that they are on the gravy train is preferable to getting off the train.

Often, the only reliable way to remove the pernicious effects of conflicts of interest is to remove the conflicts (although, in some instances, the costs of removing the COI may outweigh the pernicious effect of allowing it to continue). The New Yorker’s James Surowiecki (2005) put it this way, “Transparency is well and good, but accuracy and objectivity are even better. Wall Street doesn’t have to keep confessing its sins. It just has to stop committing them.”

CONTENTS 

Ideas to Apply

Areas of Research

Case Studies

Open Questions

To Learn More

IDEAS TO APPLY (Based on research covered below)

• Uncover conflicts of interest, then restructure the system to illuminate them. There are no cure-alls for conflicts of interest, and people often underestimate the severity of bias caused by such conflicts. Professionals often insist that they can properly navigate their conflicts, but objectivity is sometimes just not humanly possible. Therefore, ethical systems designers should be ruthless in identifying conflicts of interest and finding ways to create or restructure rules, procedures, other controls, and incentives to minimize them.
• Consider the opposite. While effective remedies for bias have been few and far between, one of the best decision strategies is called “consider the opposite” (Lord, Lepper & Preston, 1984; Milkman, Chugh, & Bazerman, 2009). It is similar to “playing the devil’s advocate” and goes like this: Ask yourself what you want to be true (i.e., what is in your personal interest) or what you are inclined to believe. Then consider several possible reasons to go against it. Do this early in your decision process, especially when the decision is important. It works better than you might think. “Can I disbelieve it?” seems like merely a logical flipside to “can I believe it?” but these questions turn out to be dramatically different to the brain; the former often leads to clearer thinking, while the latter is what we typically ask ourselves when “it” is something we want to be true.

AREAS OF RESEARCH

• To what extent do conflicts of interest influence professionals? Many doctors receive gifts and payments from pharmaceutical companies who want to influence their treatment choices. That money seems to be well spent. Exposure to a policy that restricts gifts (from representatives of pharmaceutical and device industries) during medical school is correlated with reduced prescribing of two out of three newly introduced psychotropic medications (King, Essick, Bearman, & Ross, 2013). Medical industry sponsorship negatively influences physicians’ perceptions of a drug trial’s methodological quality and reduces their willingness to believe in trial findings independent of the trial’s quality (Kesselheim et al., 2012). Another study showed that reminders of personal sacrifice can influence people’s willingness to accept gifts- another indicator of how context can affect decision making. (Sah & Lowenstein, 2010).
• How do issues of conflicts of interest vary by industry? COI issues can vary considerably by industry. For instance, here is a blog on conflicts of interest in the press. And here is a blog posting about a recent law review article comparing approaches to dealing with COIs in three different industries.
• How can we improve potential regulatory “fixes” (e.g., disclosure)? While disclosure has been shown to have its shortcomings, no one is arguing against transparency. So, how do we make it work? Disclosure may work better when advisees can punish biased advisors (Church & Kuang, 2009). And, it may work better for experienced advisees (Koch & Schmidt, 2010). Indeed, there have been many insights on improving disclosure’s efficacy, and this is where many fruits of future research lie (Loewenstein, Cain & Sah, 2011). Disclosure may also help in encouraging advisors to avoid conflicts, so that they have none to disclose (Sah & Loewenstein, 2014).
• Are academics immune to conflicts of interest? There is a small body of work on whether academics—particularly economists—are swayed in their scholarship when they receive money from industries they study. See this NYT article on the problem, which refers to this academic study (Carrick-Hagenbarth & Epstein, 2012). That study examined 19 leading economists who had taken clear positions on the topic of regulating financial firms. The correlational design was not able to prove that money biased their analyses, but it did demonstrate that 15 of the economists received money from financial firms, and few of them disclosed those financial relationships.

CASE STUDIES

• For vivid examples of how conflicts of interest at all levels—including government and the academy—contributed to the global financial crisis, see Charles Ferguson’s documentary Inside Job.

OPEN QUESTIONS

• Are there important variables such as the size of the conflict (large vs. small), or the domain (medical vs. financial, etc.)? Do some responses, e.g., disclosure, work better for some types of conflict?
• What role should rules, transparency, incentives, or nudges play in controlling conflicts of interest?

TO LEARN MORE

This page is overseen by Daylian Cain and Jeffrey Kaplan, although other contributors may have added content.

Miscellaneous Links & References

• See our blog post from 12/9/13 regarding an experiment in India showing how a systems approach to conflicts of interest can reduce corruption in the practice of auditing.
• A branch of economics called “agency theory” is about conflicts of interest between a principal and someone hired to act on the principal’s behalf, and how to design incentives to reconcile this conflict.