Do Professional Norms in the Banking Industry Favor Risk-taking?

Excessive risk-taking in the financial industry is thought to be one of the key contributor to the recent financial crisis (e.g., Diamond and Rajan, 2009; Financial Crisis Inquiry Commission, 2011; Freixas and Dewatripont, 2012). Academics and policy makers have proposed several reasons for undesirably high levels of risk taking, including flawed compensation practices (Krugman 2009), poor corporate governance (Freixas and Dewatripont, 2012), and low capital requirements (Admati and Hellwig, 2013). One explanation that has received increased attention in recent years is the financial sector’s risk culture, i.e., the norms and informal rules of behavior that may favor inappropriately high levels of risk taking. Consequently, policy makers and regulators have called for a change in the risk culture (e.g., House of Commons Treasury Committee, 2008; Power, Ashby, and Palermo, 2013; International Monetary Fund, 2014). However, there is little or no empirical evidence showing that the unwritten rules of behavior in the financial industry encourage financial professionals to take greater financial risks.

We conducted experiments with financial professionals and professionals from other industries to test this hypothesis (Cohn, Fehr & Marechal 2017). Measuring the impact of business culture on behavior is notoriously difficult because professional groups with different cultures also vary along many other dimensions, some of which may not be observable. Our approach, which circumvents this issue, draws on identity theory (Akerlof and Kranton, 2000). The theory proposes that people have multiple identities based on, e.g., their gender, ethnicity or professional role. The idea is that each identity is tied to norms that prescribe appropriate behavior in a specific situation. Now, which identity guides behavior more strongly depends on the relative salience or accessibility of the identity and associated norms in a person’s mind. Consequently, if the professional norms in the financial industry encourage risk taking, we should observe that financial professionals become more willing to take risks when their professional identity is made more salient.

For our first experiment, we recruited 128 employees of a large, international bank and asked them to fill out a short online survey. After some initial filler questions, participants were randomly assigned to one of two conditions. In the professional identity condition, we asked them seven questions about their professional background and experience. The goal of this manipulation was to increase the saliency of participants’ professional identity (for which we have evidence that it worked). Those in the control condition answered questions unrelated to their professional role, such as what is their favorite leisure activity. All participants then received the opportunity to make an investment choice that had real monetary consequences. They were given US $200 of which they could invest any amount in a risky asset which (i) paid back 2.5 times the investment with 50% probability or (ii) nothing with 50% probability. Participants knew these probabilities and were allowed to keep all the money they did not invest. We use the dollar amount invested in the risky asset as a proxy for participants’ willingness to take financial risks. The key feature of the experiment is that it allows us to compare the behavior of professionals within the same profession and for the same risky asset. Thus, any difference in risk taking behavior between conditions can be unambiguously tied back to the manipulation, and not to differences unrelated to a profession’s culture.

We find that bank employees in the professional identity condition took significantly less risk. They invested about 20% less in the risky asset relative to the control group. Thus, the results do not confirm the widespread belief that the professional norms in the financial industry promote risk taking. To examine whether the effect is indeed unique to the financial industry, and not just due to participants thinking about their jobs, we ran the same experiment with 133 employees from other industries. However, the nonbanking employees did not respond like the bank employees. In fact, they invested even slightly more in the risky asset when reminded of their professional identity, though the effect is not significant. Since participants of the first experiment all came from the same financial institution, we wanted to test whether the effect replicates across a broader set of financial professionals. To this end, we also ran the experiment with 142 employees from several other banks. The results of the replication experiment are remarkably similar to the original experiment. We find that bank employees in the professional identity condition invested roughly 20 percent less in the risky asset compared to the control group. Thus, the results seem to apply to the broader financial industry, and not just a single bank.

Overall, we do not find evidence that the risk culture in the financial industry promotes risk taking. In fact, our results suggest the opposite, i.e., the cultural norms appear to advocate prudence rather than excessive risk taking. However, our results do not imply that the business culture in the financial industry is completely unrelated to the problem of excessive risk taking. In a companion paper, we find that bank employees who were reminded of their professional identity were more likely to cheat and break rules relative to the control group (Cohn, Fehr, and Marechal, 2014). In the past, there have been several cases of excessive risk taking for personal benefit (e.g., to achieve higher bonus payments) that occurred because traders disregarded the trading limits in order to make up for their losses which, however, resulted in even bigger losses for their institutions (e.g., the so-called “London Whale” at JPMorgan Chase). Thus, professional norms in the financial industry may contribute to excessive risk taking, but our results suggest that the channel may have been a lack of commitment to ethics standards rather than a problematic risk culture per se.

The complete paper is available for download here.