I want to follow up on Melanie’s post last week, about the OECD’s first-ever Foreign Bribery Report, and what its findings tell us about patterns and tendencies in firms’ illegal bribe activities in foreign countries. The Report is an important and informative document that presents, as its introduction says, “an analysis of all foreign bribery enforcement actions that have been completed since the entry into force of the” OECD Anti-Bribery Convention. There’s a lot in it, and I may do another blog post at some point on some other aspect of the report. But for now I wanted to focus on one thing about the report that jumped out at me: the way in which the report’s findings seem to be in some tension with my prior beliefs/stereotypes about the contexts in which foreign bribery is most frequent.

Let me start with my prior beliefs, which are not based on much firsthand information, but which I’ve absorbed from a lot of people who work in this area, and I think are fairly widely shared. These beliefs run as follows: Whatever the world was like a decade or two ago, these days most major multinational firms recognize the seriousness of anticorruption laws like the FCPA, and most such firms have fairly robust (though often imperfect) compliance programs. When such firms run afoul of the FCPA or similar laws–which they still do, probably far too often–it is less likely these to be the deliberate policy of senior management, and more likely to be low or mid-level employees “in the field,” under pressure to increase business in high-risk emerging markets. This doesn’t mean senior managers are blameless–they may have failed to set the right “tone from the top,” or failed to implement an adequate compliance program, or looked the other way. But at major multinationals, many (including me) were of the view that bribery is usually not the firm’s policy. By contrast, the thinking often goes, small and medium-sized enterprises (SMEs), expanding in to high-risk foreign markets for perhaps the first time, are much more likely to run afoul of the FCPA. They are less likely to be familiar with the statute, less likely to have sophisticated (and expensive) compliance programs in place, and less accustomed to managing the pressures of doing business in environments where corruption is prevalent.

The OECD Report strongly implies (but does not quite say) that this is (mostly) wrong. As the report states at the outset, “[c]orporate leadership [was] involved, or at least aware, of the practice of foreign bribery in most cases, rebutting perceptions of bribery as the act of rogue employees.” More specifically, in the 427 foreign bribery enforcement actions the OECD examined, in 12% the CEO was involved, and in another 41%, “management-level employees paid or authorized the bribe.” As for the firms involved, the OECD found that “[o]nly 4% of the sanctioned companies were … SMEs,” while in 60% of cases the company had over 250 employees, and in another 36% the company size could not be determined from the case records.

So, does this mean my prior beliefs were all wrong? Are the most likely foreign bribery culprits senior executives at large multinationals, rather than lower-level employees and SMEs? Maybe. But not necessarily. Whereas Melanie treated the Report as refuting the “rogue employee myth,” and spinning out the logical consequences of that refutation, I want to take a different tack, by raising a few questions about how we should interpret the report’s findings here for the types of foreign bribery problems that are most typical. Indeed, although the OECD Report’s findings are important and ought to provoke all of us to re-examine some of our assumptions, I want to suggest a few reasons to be cautious about not drawing overly broad and unwarranted inferences on these particular points.

• First, and most important by far, we need to bear in mind that the characteristics of the cases that result in enforcement actions (and that therefore end up in the OECD Report’s data set) are not necessarily representative of all the cases of foreign bribery that actually take place. There are some obvious forms of selection bias that are likely to be at work here (closely related to issues we’ve discussed on this blog previously). For instance, prosecutors are most likely to go after the cases with the most egregious violations of the statute, the biggest bribes, and the richest defendants. That means they may disproportionately (but perhaps appropriately) target larger corporations, and cases in which senior management was either directly involved or at least complicit. There may also be some selection bias in terms of which violations come to the attention of law enforcement. Larger firms may be subject to more scrutiny, or may be more likely to detect and self-report violations (precisely because they have better compliance systems and more effective internal auditing). Bribery schemes that involve senior management may be larger, and therefore not only viewed as more blameworthy, but perhaps also more likely to be discovered, because a larger number of people may be in a position to find evidence of wrongdoing. None of that is meant to imply that anything in the OECD Report is incorrect; the report is completely transparent about its data and methods. It does mean, though, that readers need to be careful in extrapolating from the Report’s data on cases that led to enforcement actions to all cases that could have led to an enforcement action.
• Second, the OECD Report’s definitions of some key terms may be capacious enough to reduce the tension between my prior beliefs and the Report’s apparent findings. For example, on the issue of who in the company is paying bribes, the OECD finds that the CEO was involved in only 12% of cases (still alarmingly high, but nonetheless a relatively modest number); to get to the conclusion that “corporate leadership is involved … in most cases,” the OECD Report must include the “41% of cases [where] management-level employees paid or authorized the bribe” (emphasis added). It’s not clear from the Report who counts as a “management-level employee.” If that category would include, for instance, a regional sales director stationed in a high-risk market, or an executive at a foreign subsidiary, then it would not be terribly inconsistent with the conventional view that most foreign bribery, at least in larger firms today, takes place because lower-level employees or agents evade the firm’s (perhaps inadequate) compliance system, rather than because the firm’s senior management has adopted bribery as part of its business plan.
• Third, though perhaps of lesser importance, it’s worth remembering that the OECD Report’s data spans a 15+ year period, from February 1999 to June 2014. My working hypothesis is that, as a result of the surge in enforcement (particularly FCPA enforcement) that started right around the beginning of that period, it is now the case that major multinational firms take foreign anti-bribery law seriously and have programs in place designed to prevent violations of those laws. But it took time for that to occur, so one would expect lots of senior executives at big firms directly involved in foreign bribery at the beginning of the period, and fewer at the end (correcting for other trends and factors). I’m not sure if there’s enough data to check this. It seems at least superficially plausible that taking this into account might alter the picture somewhat.

All that said, the OECD Report has caused me to re-examine some of my not-terribly-well-supported beliefs about patterns of foreign bribery. While I do think the above points are valid, I’m also a bit worried that I’m trying too hard to come up with reasons to maintain my prior beliefs in the face of contrary evidence–an all-too-common (and well-documented) human psychological failing. So for now I’ll just throw those ideas out there in the hopes that some better-informed people might be able to nail down more precisely what we do and don’t know about the empirics of foreign bribery. And again, kudos to the OECD for publishing this useful report, which contains much more useful information than this very narrowly-focused commentary addresses. Perhaps I’ll turn to some of those other findings and conclusions in a future post.