*This is the twelfth post in a series on Michael Livermore and Richard Revesz’s new book, Reviving Rationality: Saving Cost-Benefit Analysis for the Sake of the Environment and Our Health. For other posts in the series, click here.

Reviving Rationality is an important read from two of the legal academy’s foremost experts on the administrative state. It provides a nice history of cost-benefit analysis, how it has evolved analytically, institutionally, and how it has mapped to our changing political landscape. It is this last aspect that represents the core motivation for the project.

For much of the last half-century, Livermore and Revesz (LR) persuasively contend, political forces channeled significant policy disagreements through cost-benefit analysis. Those on the right argued for more demanding forms of cost-benefit analysis; those on the left argued for looser cost-benefit analysis and less constrained agencies. But as executive orders from both Democratic and Republican presidents demonstrate, notwithstanding these disagreements a rough bipartisan consensus embraced reasoned decision-making and cost-benefit analysis, at least in some form and at least in many contexts.

President Trump upset this equilibrium in LR’s telling. At the start of his administration he had, as LR put it, a Hobson’s choice between hiring Republican technocrats and letting the establishment run the show or hiring loyal novices. Particularly early in his presidency, Trump adopted a mixed practice of appointing technocrats in some places and loyalists in others. But for the most part he appointed novices who, even if they could not yet be determined to be personally loyal to him, at least could be said to be only slightly encumbered by technocratic skills, ideals, or social networks.

The incompetence of these officials, when combined with intentional acts of sabotage and delegitimization, led to a “retreat from reason” in the administrative state under Trump. As LR show, he substantially weakened the guardrails that helped guide administrative decision-making and temper political influence. LR’s central interest, of course, is cost-benefit analysis, and there they chronicle the many ways in which the Trump administration mocked and undermined the analytical technique (for example, by declining to count benefits in regulatory analysis). But the guardrails fell apart in other places, too. All of this led, among other things, to an abysmal record for the administration before the courts, where at least for the short-term a commitment to reasoning remained largely intact.

To briefly lay my cards on the table, I am sympathetic to LR’s descriptive thesis. And also to their normative objective—to revive rationality in the administrative state as against the withering assault of recent years. As I contend elsewhere and in a forthcoming book, The Reasoning State, I see the distinctive margin of the administrative state precisely in its ability to credibly and publicly reason. Cost-benefit analysis should be seen as a form of structured reason-giving, and it represents an important ally in achieving public credibility.

So where I wish to offer some friendly (and also probably unfair) comments is in the third part of their book, in which LR offer recommendations for shoring up cost-benefit analysis and in the larger picture reviving rationality. I agree with many of their ideas about how cost-benefit analysis might be improved. To take one example, I very much think it a good idea to step back from an exclusive focus on efficiency—as with President Biden, distributional considerations should represent part of reasoned decision-making, despite the tricky analytical problems they implicate. Yet as LR acknowledge, resurrecting the guardrails, much less improving upon them, depends on the existence of a cooperative partner in the White House and possibly Congress. Maintaining those guardrails, I would add, would seem further to require the continuous existence of a cooperative partner in the White House. The outcome of the 2020 election—unknown to LR at the time of publication—delivered a cooperative partner in the White House for the moment. But what of the future?

The retreat from reason, after all, occurred in the context of decades of rough bipartisan consensus on the relevant guardrails. What is to stop Trump or a successor in spirit from flicking away resurrected guardrails? Indeed, if anything, he would be able to do so more easily in a new administration, with know-how in mind and more compliant courts on review. Recall also the Hobson’s choice that Trump faced early in his administration and Reagan’s maxim that “people are policy” (LR 36). Whatever analytical tools we decide upon, whatever norms we attempt to embed around decision-making processes, it is still and must be people who will be most relevant to what happens. Analytical steps only matter if people follow them; norms only matter if people value them. And not only did Trump infuse novice loyalists into the administration, he also undermined the decisional autonomy enjoyed by civil servants and exposed them to overt political influence.

A helpful step not only to reviving rationality, but importantly to maintaining it, therefore is to train attention squarely on personnel. This is a large area of inquiry, and I will only gesture to some topics of interest (indeed, as I suggest, it’s probably outside of the fair ambit of LR’s book). But as I see it, this means considering two basic questions. First, what type of person is appointed or hired? This means, among other things, pushing on the question of what qualifications Congress may impose on political appointments consistent with the Constitution; hemming the Vacancies Reform Act, perhaps along the lines of Representative Porter’s H.R. 2994; defending the Office of Personnel Management, which in many ways represents the beating heart of the civil service; and reviewing Executive Order 13,843, which potentially opens the door to politicized agency adjudications. A second question is, once appointed or hired, what incentives do we expose personnel to? This means, among other things, guarding against civil service protection end-runs, such as in Executive Order 13,957, by passing legislation perhaps along the lines of H.R. 302; and safeguarding whistleblowers and inspectors general. Many of these areas can be most effectively addressed through legislation. But even self-limiting agency regulations can be of use in creating space for decisional autonomy, as for example in the case of the special counsel regulations and elsewhere. Such steps promise agents of reason at least temporary shelter and space to regroup in future storms.

Jed Stiglitz is the Associate Dean for Faculty Research and a Professor of Law at Cornell Law School.