Holly Doremus

The outgoing Bush administration leaves in its wake a systemic problem at key federal environmental agencies. Career scientists are thoroughly demoralized. The morale problem will be helped by case-by-case reversal of some of the most extreme Bush-era anti-environmental actions, but morale will not be fully restored until institutional systems are in place to make better use of scientific evidence and scientific personnel. That sounds easy, but it will take more than simply declaring a commitment to scientific integrity. The Obama Administration should draft an executive order on scientific integrity that highlights the roles of science and politics in policy decisions and provides mechanisms for separating the two.

A confluence of circumstances produces the science problem. First, environmental policy decisions typically have both scientific and non-scientific elements. They require both the best possible understanding of the relevant facts about the natural world—the realm of science—and judgments about goals, priorities and distribution of costs—the realm of politics. Second, those elements are difficult to separate. Data are often remarkably thin, which means that scientific conclusions may be influenced by the scientists’ underlying political views. Third, a slew of actors inside and outside government agencies have incentives to blur the distinction between science and politics. Sometimes the people who should be making political decisions—the politically-appointed leadership of a regulatory agency—want to fob those choices off on agency scientists or scientific advisors. Sometimes political actors want scientific cover for their political choices. Scientists may play along with this “science charade,” assuming that doing so will increase their influence. Finally, the scientific and political judgments inherent in environmental policy decisions ought to be made by different persons and overseen in different ways.

Judgments about the facts of the natural world are best made by people with expertise in the relevant science. Scientific training and reputational interests constrain those judgments to some extent, but oversight may be needed to limit the role of scientists’ political preferences and other biases. Political judgments are legitimately made by the legislature, the president, and the political appointees who run agencies on the president’s behalf. But agency decisions are legitimate only if they both stay within existing statutory boundaries and are publicly acknowledged so that Congress, the president, and the voters can react.

The Obama transition team is aware of the science problem, and has floated the idea of an early executive order on scientific integrity. That’s a good idea, but it’s not likely to be enough. The executive order proposed by transition officials calls for a statement that “all federal policy and information provided to the public by the federal government will be based on the best scientific evidence.” That’s already the law. Many environmental laws explicitly invoke the best science standard. For science-intensive decisions, it’s likely implicit in the general Administrative Procedure Act requirement that agency action not be arbitrary or capricious. Furthermore, a Clinton-era Executive Order which remains in effect already requires that agencies base their decisions “on the best reasonably obtainable scientific, technical, economic, and other information.” And the Data Quality Act already calls for government-wide guidelines for “ensuring and maximizing the quality, objectivity, utility, and integrity” of federally-disseminated information.

Of course, if the Obama Administration is more committed than the last one to scientific integrity, it could accomplish more with the same tools. But the executive order approach will inevitably come up short. It will not be possible for even the most committed administration to directly oversee every decision by every agency. Systematic institutional changes will be needed to ensure broad-spectrum improvements. Beyond calling for use of the best science, the next administration needs to articulate what that means and adjust agency structures and procedures to encourage it. It should tackle the science problem on three key fronts.

First, agencies should be directed to clarify the scientific and political elements of their environmental policy decisions, and to more clearly define role boundaries for agency personnel. It will not always be possible to cleanly separate science from politics, but acknowledging that the two are intertwined will help raise awareness of the role distinction. Providing training for agency actors, both career scientists and political appointees, should help to cabin the actions of each group within their appropriate role. An executive order should mandate these steps, and should require that agencies report periodically on implementation.

Second, improved institutional separation is needed between political appointees and career scientists. The biggest scientific scandal of the Bush administration was the way that Julie MacDonald, Deputy Assistant Secretary of Interior for Fish, Wildlife and Parks, repeatedly injected herself into the scientific elements of endangered species policy decisions. An investigation by the Interior Department’s Inspector General found that MacDonald routinely called agency scientists directly and tried to bully them into reaching her favored conclusions. Even more troubling, the investigation concluded that MacDonald’s boorish behavior did not violate any laws or department regulations. Again, better training could help. So could key personnel decisions. The president and agency heads could make sure that political appointees understand the need for role separation and are committed to it. They could put high-level career scientists with the stature to resist improper political interference in key science advisory positions, or create a scientific ombudsperson position to provide an avenue for agency scientists to complain about political interference. An executive order could require that agencies issue procedural regulations detailing the role of political and career staff in key decisions.

Third, increased transparency is needed to allow effective oversight of role separation. The MacDonald Report gave the Bush administration a high-profile black eye, and even encouraged a federal court to set aside a decision tainted by political interference. But MacDonald’s meddling, although widely known within the Interior Department, only came to public light through the Inspector General’s decision to investigate. Even the courts have been reluctant to require that the details of agency scientific recommendations be revealed in litigation, with some ruling that the deliberative process privilege shields those details. New mechanisms are needed to reveal when political appointees overstep their bounds. An executive order could require that agencies disclose the unvarnished recommendations of agency scientists that feed into policy decisions, and could explicitly and generally waive the deliberative process privilege for such recommendations.

To solve the science problem, the incoming administration will have to change agency cultures. That is not an easy task and cannot be accomplished by the stroke of a pen, although the right pen strokes in an Executive Order would provide a very good start at restoring and recognizing the proper roles of scientific and political judgment. That order should be followed with the institution of ongoing personnel training, oversight of boundary recognition, and appointments that reinforce effective role-separation.

Holly Doremus is Professor of Law at UC Berkeley and UC Davis.