An important real-world conflict of interest experiment

In today’s NY Times, Michael Greenstone, an economics professor at MIT, writes about a study on auditor COIs that he –  together with Esther Duflo of M.I.T.;  and Rohini Pande  and Nicholas Ryan, both of Harvard – recently published.   The study was conducted in Gujarat, India, where industrial plants with high pollution risks are required  “to hire and pay auditors to check air and water pollution levels three times annually and then submit a yearly report to” a governmental body. In the study, for a randomly selected set of companies, but not for a control group, “auditors were paid a fixed fee from a central pool of money, a subset of the audits was chosen to have its findings re-examined, and auditors received payments for accurate reports, judged by comparisons with the re-examinations. The control group continued under the status quo system in which auditors were chosen and paid by the plants they were auditing.”

The results of this real-world experiment  powerfully demonstrate the impact on the ethicality of conduct that financial incentives can have – even on the judgment of individuals who, by virtue of their professional norms, are supposed to be resistant to COIs.  That is: “While many of the plants violated the pollution standards, few of the auditors in the control group reported these violations. In the case of particulate matter, an especially harmful air pollutant, auditors reported that only 7 percent of industrial plants violated the pollution standard. In reality, 59 percent of plants exceeded it.” However, “[t]he rules changes [in the experiment] caused the auditors to report more truthfully. In the restructured market, auditors were 80 percent less likely to falsely report a pollution reading as in compliance, and their reported pollution readings were 50 to 70 percent higher than when they were working in the status quo system. This difference was as large even when comparing reports of auditors working simultaneously under the two systems. Finally, and most important, the plants that were required to use the new auditing system significantly reduced their emissions of air and water pollution, relative to the plants operating in the status quo system. Presumably, this was because the plants’ operators understood that the regulators were receiving more accurate information and would follow up on it.”

Three comments on this important study.

First, while most directly relevant to auditors, these results can, I believe, be broadly applicable to COIs generally.  That is, if professionals who are trained to rise above COIs fare this poorly, one can only imagine the impact of COIs on the rest of us.

Second, the more important compliance and ethics program efforts become to society, the greater the need for not just C&E auditing but other forms of checking – such as monitoring, as was discussed in a piece in Corporate Compliance Insights.   But monitoring  (as a general matter) is even less independent than is auditing, so this recent study underscores  the considerable  challenges for making forms of checking beyond auditing effective.

Third, research to determine “what works”   is vitally important for the C&E field to mature and realize its full promise,  and real-world studies such as this one can be particularly valuable in that regard.  Interestingly, another article in today’s NY Times describes how in the UK there is now an government-run effort (headed by a “Behavioral Insights Team”) to use research to determine what works with respect to various public policies, including some compliance-related ones. I hope that the US and other countries will follow the UK’s lead here.

Finally, here is a prior post on auditor COIs.