think the worst!

I talked about the Deepwater Horizon oil spill in the Gulf of Mexico with Maurizio Catino, a colleague from the University of Milano—Bicocca, whose work on organizational disasters and organizational myopia offers interesting insights on the topic. You should consider Maurizio the real blogger today, since all the ideas in this post come from him.

The conversation started out from a quite casual question I had, namely whether BP is really the best actor to handle the crisis. Most people, and the US President among them, have taken for granted that BP should be in charge and lead the oil spill “response” efforts. Indeed, BP is the most qualified actor available – the company has a better technical understanding and expertise than any public agency – and has all the interest in cleaning up its mess (or, more realistically, to contain the oil spill). Although I buy these arguments, I wonder whether some of the organizational problems that lead to the spill, are also in place in the planning and execution of the efforts to stop it. Maurizio’s answer raised interesting considerations concerning BP’s risk management philosophy, and the causes of its sloppy management.

First of all, it is relevant here to consider the distinction between a probability-based and a possibility-based approach to risk management. In the first case, companies base their risk-management strategies on the probability that an adverse event would occur. In the latter case, companies consider all the possible adverse events – i.e., all the events that have some chance to occur – and devise strategies for their prevention as well as crisis management plans. From a normative point of view, if we want to protect society and the environment from negative externalities due to organizational failures, we need to make sure that high-risk organizations move away from a probabilistic risk management philosophy, and develop a “think the worst” attitude in the consideration of high-impact events. The problem with the probabilistic approach is that it leads to underestimate the level of perceived risk of disastrous events in case they have a low probability of occurrence. In other words, to prevent future disasters, high-risk companies should give less attention to “what is the likelihood of an adverse event to happen?”, and focus on “what should be done in case an adverse event happens?” This is an important consideration especially in the light of the fact that oil reserves are shrinking and more hazardous drilling sites will be explored in the near future.

This leads to a second set of considerations concerning governmental regulation and supervision. President Obama compared the oil spill to the 9/11 terroristic attack, while some of his detractors are trying to cast it as his Katrina. Of course, Obama has no interest in being associated with G.W. Bush’s mismanagement of the Katrina disaster, and would rather prefer to appear as the commander in chief of a “mission accomplished”. Nonetheless, this might not be the best way to frame the events. Let us consider their nature. Both 9/11 and Katrina are extremely rare events, but 9/11 was also unexpected, and unpredictable (a black swan), while Katrina was a natural disaster of exceptional magnitude that has been underestimated by the government. Relying on the classical distinction between risk and radical uncertainty (which goes back to Keynes), while Katrina was a predictable event, the 9/11 attack was not (although security agencies had valuable information, they lacked the interpretative framework in which civilians planes could be used as weapons).

Since the BP oil-spill was a predictable event, we should conclude that it bears closer resemblance to Katrina than 9/11.  And, we argue, it is in the government interest to frame it in these terms. A company cannot be blamed for not preventing an unpredictable event. In contrast, in order to promote “possibility-based” risk management strategies, governments should require that high-risk companies provide ex ante a detailed plan for the prevention and crisis management of adverse predictable events. For instance, governmental authorizations of offshore drilling should be contingent upon the presentation of a safety and recovery plan in case of an oil spill. In sum, while it might be politically wise (in the short-term) for Obama to cast the Gulf of Mexico oil spill as a 9/11 type of disaster, it might be counterproductive in the long term, because it gives BP an alibi. In contrast, what the government should do is to promote regulation that would force companies to think about solutions to potential disasters and devote resources to this goal.

Getting companies to internalize negative externalities is not an easy task, and now it is even harder due to the diffusion of legal responsibility (i.e., see a list of “guilty parties” involved in the BP operation), and the physical separation between the management and the place in which the activity takes place. A successful example of risk management is Du Pont’s solution to the problem of securing the production of explosive material. The company forced its managers to live with their families inside the factories. In striding contrast, Tony Hayward, the chief executive of BP went sailing this week-end, but he was far away from the Gulf of Mexico. To some extent, management’s physical distance from the disaster translates into cognitive distance. The sloppy management and organizational failure of BP is partly due to the fact that their headquarters did not supervise the network of suppliers in a reliable way. Moreover, the organization was blind and deaf to signals coming from latent factors and near miss (see BP worker Tyrone Benton’s interview) and its response efforts were, at times, hilarious…

In sum, what I got out of this conversation is that, yes, BP should be in charge of cleaning up its mess, but the government should make sure that in the future high-risk organizations give up some of their optimism, and start thinking the worst. Only in this way, the worst can be “contained”.