A great case for worst case analysis
Cross-posted at CPRBlog.
The National Environmental Policy Act (NEPA) is the nation’s look-before-you-leap environmental law, intended to make sure that we understand what environmental problems we might result before we act. To that end, federal agencies must prepare an environmental impact statement (EIS) before they take, authorize, or provide funding for actions that may have significant adverse environmental impacts. Useful as NEPA analysis is, the Deepwater Horizon disaster vividly illustrates the need to fix one of its shortcomings.
The White House’s Council on Environmental Quality (CEQ) oversees NEPA compliance. It has issued regulations prescribing how agencies should prepare EISs and what should be in those documents. The regulations are almost unchanged since they were originally issued during Jimmy Carter’s presidency, with one conspicuous exception. Where the impacts are uncertain or unknown, the regulations used to require that the EIS “include a worst case analysis and an indication of the probability or improbability of its occurrence.” Sensible as it sounds, that requirement was renounced in the Reagan administration in favor of a vague call to evaluate “reasonably foreseeable” environmental impacts, including low-probability but potentially catastrophic impacts, “provided that the analysis of the impacts is supported by credible scientific evidence, is not based on pure conjecture, and is within the rule of reason.”
The recent Gulf oil disaster offers a powerful argument for going back to the original requirement for worst-case analysis, which the current regulation allows agencies to avoid. According to the Washington Post,
BP’s own exploration plan, submitted to federal regulators in February 2009, minimized the danger of a spill. The company said “it is unlikely that an accidental oil spill release would occur from the proposed activities.” While it acknowledged that a spill could “cause impacts to wetlands” and to beaches, it added that “due to the distance to shore (48 miles) and the response capabilities that would be implemented, no significant adverse impacts are expected.” It said any effects on fish or shellfish would be “sub-lethal.”
BP was convinced that the blowout prevention systems would work. Therefore, it thought, any leaks would not be large, and could be contained long before they would reach shore. Interior’s Minerals Management Service bought that analysis.
Now that the worst case has come to pass, we can see the costs of not considering it earlier. NEPA itself does not require that environmental impacts be avoided or even minimized. Ventilating the worst case would not have precluded the oil lease or development. It might have shifted the political landscape a bit, perhaps making it more difficult for officials to approve this and other leases. But that’s not the most important function of worst-case analysis. What it surely would have done is make everyone prepare better for this kind of disaster. Faced with a worst-case scenario of an oil spill the size of the Exxon Valdez reaching the Gulf’s vibrant wetlands, surely BP and Interior would have done more both to ensure that the blowout prevention systems were reliable, and to prepare a rapid response to a catastrophic leak.
Coast Guard Commandant Thad Allen has continued to defend BP, saying in a Friday interview that “It’s hard to write a plan for a catastrophic event that has no precedent, which is what this was,” and that BP could not have been epxected to prepare a response for “what could never be in a plan, what you couldn’t anticipate.”
But of course both the company and US officials could have anticipated this spill. They didn’t because they weren’t forced to. It’s human nature to assume the best outcome and downplay the likelihood of catastrophic failure. NEPA can, and should, be used to counteract that tendency. CEQ was right in 1978. It should go back to its original worst-case analysis requirement. Without it, we now know, federal decisions will ignore some major and altogether foreseeable risks.