Is misbehaving contagious?
While our research looked at financial misconduct, and the degree of its contagion, this has implications for any other setting, where people are grouped together after misconduct.
We measured the incidence of individual misconduct of financial advisors before and after merger of their firm’s branches, where in their new work groups, they encounter co-workers with a previous history of financial misconduct. The financial misconduct was based on customer-initiated complaints such as unsuitability, unauthorised trading, churning, and misrepresentation or omission.
We defined an advisor’s co-workers as the other advisors employed at the same branch of a firm at the same time. The important parameter in this is that given the shared office location and the small average branch size, it is reasonable to assume that nearly all group members will interact frequently.
Our results show that an advisor is 37% more likely to commit misconduct if he is merged into a new branch that includes co-workers with a history of misconduct, as compared to an advisor who does not work with new colleagues with a history of misconduct. This implies a social multiplier of 1.59, meaning that a case of misconduct by an advisor results in an additional 0.59 cases of misconduct by that advisor’s co-workers.
Co-workers influence misconduct in several ways. Firstly, through social learning, advisors could learn about the profitability of misconduct, techniques for committing misconduct, or how to reduce the risk of detection. Secondly, advisors could absorb the branch’s culture and its social norms, and conform their behaviour accordingly. Such social norms could be directly related to misconduct, but could also include indirectly related social norms, for example, a branch’s culture of intense competition or a lack of respect for clients. Advisors could also mimic their co-workers’ behaviour due to concerns about their relative position.
Establishing that there are peer contagion effects in financial misconduct has important policy implications. Clearly understanding how misconduct can “spread” is important, so firms can design effective internal controls. It also affects both the type and the severity of the penalties. If misconduct by one advisor creates a downstream consequence, then penalties for misconduct should be more severe than if there were no future consequences. The possibility of such “spread” should also affect the type of punishment, making suspensions or expulsions from the industry more effective than monetary penalties. Further, co-worker influence suggests that changes to the current disclosure regime could benefit investors. Currently, investors can only view information about an individual advisor. Requiring disclosure of information about an advisor’s co-workers could provide investors with additional useful information, while also creating a reputational penalty for the negative effects of co-worker influence.
There is a larger social application. If financial misconduct contagion arises between co-workers who spend a lot of time together, what about possible contagion between other individuals who are similarly put into the same groups, such as in schools, rehabilitation institutions and detention centres?
The specific results of our studies, in combination with the findings of other previous studies highlight some noteworthy issues.
Previous studies have shown that social influence is stronger between those with a similar demographic background or between individuals with the same ethnicity. In our studies too, an advisor is 26% more likely to commit misconduct if his new colleagues include a co-worker with a history of misconduct who also shares his ethnicity.
Other literature has shown evidence that peer effects are particularly important in the context of crime. In our research setting of financial misconduct committed by white-collar professionals, there are no formal training programmes to teach advisors how to commit misconduct; misconduct techniques and social norms can only be transmitted through informal channels such as social interactions. Similarly, for social crimes, because there are no formal schools that teach such skills, it is likely that social interaction is the most plausible means through which criminal behaviour proliferates. The peer-to-peer contagion mechanism is same: between individuals who spend a lot of time together.
In our studies, the average financial advisor is a mature 40 years. Would the social multiplier be significantly higher than 1.59 among groups of young offenders who are more easily influenced?
Our study also shows that contagion is asymmetric, that is it does not apply to good behaviour. For example, an advisor can learn about misconduct from meeting dishonest co-workers, but the effect is not reversed when he later meets honest co-workers.
Another question we flagged during our study was the extent that supervisors could influence misconduct, by setting the general ethical or through enforcement procedures. In a sub-sample study, where there was no change in supervisor, we found the same effect: advisors are more likely to commit misconduct if they encounter new co-workers with a history of misconduct. This result highlights the possible limited role of a supervisor-slash-mentor-slash-rehabilitating-officer, as the contagion mechanism resides more effectively across peers.
This evidence that the ethical conduct of many individuals is malleable and slanted towards bad behaviour—at a not-small figure of 37 per cent—is an alarming indicator of the potency of the social contagion mechanism for crime-oriented attitudes and behavioural habits. It brings into focus the usefulness, or dangers of settings—prisons, rehabilitation facilities, even school detention classes—where youths and offenders spend extended periods of time together, with ample time to share knowledge about illicit opportunities, or transfer criminal skills such as how to disable alarms or make crack.
Policy makers need to pay more heed to the mechanisms of social contagion. The proverbial bad apple can ruin a barrel.
About the author
Stephen Dimmock is Associate Professor of Banking & Finance at Nanyang Business School, NTU. The research mentioned in this article was co-authored with William Gerken, Assistant Professor of Finance & Quantitative Methods at the University of Kentucky and Nathaniel Graham, Professor of Finance and Decision Sciences at Trinity University, San Antonio.
This commentary was published in Harvard Business Review on 5 March 2018 and Entrepreneur Asia Pacific on 16 October 2018.