Ever since the start of the Great Recession in 2007/08 there has been a debate about punishing those responsible for causing the disaster that ensued. Some people blamed greedy bankers, other lax regulators, while others placed a lot of the blame on borrowers. What is clear a decade-plus later is that most of the human agents responsible went unpunished, since regulators preferred to find institutional actors over individuals. I have always contended that this approach was incomplete, at best, and misguided, at worst, a thought that came back as I read a 2018 paper called “Punish One, Teach A Hundred: The Sobering Effect of Punishment on the Unpunished” by Francesco D’Acunto (Boston College), Michael Weber (Georgetown), Jin Xie (Chinese University of Hong Kong).
The authors’ paper looks at the effect that punishing CEOs of Chinese state-owned enterprises (“SOE”s) has on unpunished firms. In brief, the authors wanted to know if publicly punishing one CEO for a specific illegal financial maneuver (in this case, “tunneling” funds from public companies to cover loan losses in a related private entity) discouraged the CEOs of similar firms from committing the same crime.
As the authors note, the effect of public punishment was visible and measurable across non-punished firms:
When we implement our first design, we find that after Chinese regulators punish a listed firm for tunneling via inter-corporate loan guarantees, unpunished listed peer SOEs cut their loan guarantees. This effect is economically and statistically signicant. After a peer is punished, SOEs reduce the amount of loan guarantees over total assets by 2.4 percentage points – about 25% of a standard deviation of the scaled loan guarantees in the sample. Moreover, non-punished peer SOEs are more likely than non-punished peer non-SOEs to implement changes in their governance that favor minority shareholders. For instance, they are 43% more likely to move from CEO duality – the CEO chairs the board – to a more independent board structure. These results suggest that SOEs react strongly to the punishment of local peers by aligning their actions with the interests of minority shareholders. Instead, local non-SOEs barely react.
In other words, SOE CEOs are well aware that a peer has been caught and punished, and that punishment causes non-punished CEOs to alter their firm’s loan guarantee allocation to reduce the risk of a similar personal fate. Moving on, the authors note that, as China’s SOEs strictly enforce retirement at age 60, older CEOs are less likely to adjust behavior than younger CEOs, who obviously have longer potential CEO tenures ahead of them:
Consistent with this variation in the incentives to react based on career concerns, we find that older SOE CEOs react less than younger SOE CEOs to the punishment of the same peers in the same locations. Crucially, we do not find any systematic differences in the likelihood that firms headed by young or old CEOs are punished, either in the full sample or when we split firms between SOEs and non-SOEs.
The authors highlight another important factor that influences reaction, and it is whether CEOs are Chinese (and thus would have a hard time finding a CEO role outside of China) or whether they are foreign (and thus more likely to be able to find a new CEO role in another country). The results are those you would expect:
As a second test to exploit only variation in the incentives to react within SOEs, we consider SOE CEOs that have ever held managerial jobs overseas relative to other SOE CEOs. The rationale for this test is that SOE CEOs with foreign experience might have an easier access to the international CEO labor market through the connections they created while working abroad (Giannetti et al. (2015)). If they were punished in China, they could move abroad and proceed their managerial career there, and hence face weaker career concerns relative to SOE CEOs without overseas experience. Consistently, we find that SOE CEOs with overseas experience react less than other SOE CEOS to the punishment of the same peers in the same locations.
In sum, CEO’s who saw peers punished curtailed their most risky behaviour, and, crucially, their firms’ performance improved because of that alteration:
We find the cuts lead to signicant improvements in total factor productivity (TFP) in the medium run, which suggests SOEs were arguably not employing the firms’ resources effciently before the punishment events (Giannetti et al. (2017); Lagaras et al. (2017)). Event studies corroborate this interpretation because SOEs’ CARs are positive in the days around the punishment events and substantially higher than the (statistically insignicant) CARs of non-SOEs after peers’ punishments.
The authors are confident that public punishment of SOE CEOs resulted in a reduction in illegal financial dealings and a positive effect on the performance of non-punished firms. Public punishment has another positive effect, which is reducing the overall cost of policing SOE CEOs:
Overall, our evidence is consistent with a sobering effect of observing peers’ punishment on the behavior of Chinese listed SOEs. An interesting feature of peers’ punishment as a corporate governance mechanism is cost effectiveness. Under this mechanism, regulators would only need to monitor and punish a small set of listed firms to obtain broad compliance, which reduces dramatically the costs of monitoring listed firms on the part of regulators and activist shareholders. This form of governance could thus be especially viable in settings like China in which more traditional forms of governance are less effective than in other settings (e.g., Allen, Qian, and Qian (2005)). Punishing the wrongdoing of one firm reduces the scope for misbehavior of peer firms without any need to monitor or investigate peer firms directly. [Emphasis mine.]
Of course, the authors are aware that their paper raises many relevant questions:
Our results open a set of questions that beget further investigation. Is the sobering effect of peers’ punishment a permanent change in agents’ behavior, or does this effect revert over time? If the effect is permanent, to what extent could the salience-of-punishment mechanism – which is cost effective because it does not require universal monitoring or oversight on the part of the regulator – substitute more expensive mechanisms that aim to guarantee a level playing field in markets?
These questions and others are worthy of much more study and elucidation. Unfortunately, it seems that financial regulators are less interested in understanding the positive system effects of punishing individual bad actors in the corporate ecosystem than in collecting large, headline-producing, fines. From personal experience, I know that c-suite executives are keenly aware of how their peers are seen and treated by their boards, regulators and the public. They take signals from peers to set compensation demands, to create strategy, to focus M&A activity, etc. It stands to reason that they would also adjust risky, or even illegal behavior, in line with what peers are known, or suspected, to “have gotten away with.”
I hope that Professors D’Acunto, Weber and Jin continue their work in this line of research. I suspect that punishing one would indeed teach a hundred, and at much lower cost than expensive corporate prosecutions. After all, a corporation starts life as a piece of paper. It’s people who make it good, and it’s people who need to be accountable when things companies go bad.
See this post on LinkedIn.