Some Doubts About the Duty of Reasoned Decision-Making

When federal courts review agency actions under the Administrative Procedure Act (APA), they must make three main findings. Factually, an agency action must be supported by substantial evidence. Legally, the action must be in accordance with the law. And as Richard J. Pierce, Jr., explains in his recent article, when an agency makes important policy choices, they must be the products of reasoned decision-making.

That last requirement is often called “hard look” review and associated with the U.S. Supreme Court’s 1983 decision in Motor Vehicles Manufacturers Association v. State Farm. Under State Farm, federal courts must ensure that agencies have articulated “satisfactory explanations” for their actions, including by providing a “rational connection between the facts found and the choices made.” State Farm also holds that when an agency changes a previously adopted policy, it is “obligated to supply a reasoned analysis for the change beyond that which may be required when an agency does not act in the first instance.” This last obligation is subject to the qualification, in the 2009 case FCC v. Fox Television Stations, that changed agency policies don’t get harder looks than new policies do.

In his article, Professor Pierce praises hard look review, and he does so worried that the doctrine is on shaky ground. Now that the Roberts Court has a working conservative majority, it may reconsider State Farm and hard look review at some point. For my part, I wonder whether State Farm should be overruled. James Burnham has recently argued for overruling State Farm as it applies to regulatory repeals, but I wonder whether the case should be overruled in its entirety and will explain my concerns here. Even if readers do not think these doubts are as serious as I do, I hope that by raising these doubts I help readers appreciate and evaluate Pierce’s argument.

Professor Pierce portrays the debate over State Farm as a “choice between a textual approach and a pragmatic approach” to hard look review. In other words, he assumes that hard look principles cannot be defended seriously on the ground that they follow from the text of the APA, but he finds those principles normatively valuable. I agree with Pierce on the text; we differ on the normative merits.

Consider first the text. Pierce finds “persuasive” doubts then-Judge Brett Kavanaugh raised about the textual foundations of the hard look doctrine in 2008 while sitting on the U.S. Court of Appeals for the D.C. Circuit. I find the same doubts persuasive. The hard look doctrine is supposed to carry into effect APA section 706(2)’s requirement that an agency action be declared unlawful and set aside if it is “arbitrary” or “capricious.” Pre-1946, before an agency action could be set aside as arbitrary, federal courts granted it a “presumption of the existence of” “any state of facts that would sustain it,” and the challenging party needed to “carry the burden of showing … that the action [was] arbitrary.” That passage, from the 1935 Supreme Court administrative law case Pacific States Box & Basket v. White, captures what “arbitrary” and “capricious” meant to competent administrative lawyers when the APA was enacted.

As Sidney Shapiro and Richard Murphy explain, circa 1946, the phrase “arbitrary and capricious” meant only “that no plausible, reasonable set of facts could be conceived of to support the rule.” And as Patrick Garry correctly concludes, State Farm “required courts to take a more scrutinizing look at” notice-and-comment rulemaking “than had been taken under earlier applications of the arbitrary and capricious test.”

I also agree with Pierce that State Farm is on shaky ground now given the make-up of the Roberts Court. Most of the Court’s “conservative” justices are conservative specifically in the sense that they think federal judges should have restrained roles. In federal administrative law, judicial-restraint advocates oppose federal courts having broad substantive discretion to second-guess decisions by federal executive-branch actors. Last year, in Seven County Infrastructure Coalition v. Eagle County, the Court repudiated four or five decades’ worth of lower-court precedents reviewing agency policy choices strictly under the National Environmental Policy Act (NEPA). As Justice Kavanaugh explained for the Court, a “course correction of sorts” was “appropriate to bring judicial review under NEPA back in line with the statutory text and common sense.” What Justice Kavanaugh said of NEPA could be said on a much broader scale about hard look review under the APA.

To be sure, judicial-restraint advocates will second-guess agency decisions when those decisions exceed the authority the agencies get from the U.S. Congress. In those situations, though, federal judges are not second-guessing agency policymaking. They are instead second-guessing agency statutory interpretation, which is, according to Loper Bright Enterprises v. Raimondo, “exclusively a judicial function.”

To make up for the textual deficiencies of State Farm and hard look review, Pierce portrays them as desirable responses to extreme political polarization. I agree that polarization is a problem. In my view, though, polarization is not ameliorated and is instead exacerbated by hard look review.

Regulation and administrative law are downstream of politics. Voters have strong opinions about many issues of regulatory policy. When Congress legislates and leaves gaps for Presidents and agencies to fill in the details through regulatory policy, voters’ preferences should be implemented by the Presidents they elect.

In his partial dissent in State Farm, then-Associate Justice William Rehnquist argued that a “change in administration brought about by the people casting their votes is a perfectly reasonable basis for an executive agency’s reappraisal of the costs and benefits of its programs.” That argument was convincing when made; it is even more convincing now in our more polarized political environment. New presidents, their political appointees, and the policy experts on their transition teams exercise far more control over what administrative agencies do than in 1983—let alone when the APA was passed. Today, I suspect that the error costs of State Farm hard look review are higher than those of pre-1946 arbitrariness review. State Farm review is designed to stop agencies from instituting policies that are arbitrary in the commonsensical meaning of that term. But State Farm creates another error cost, that federal courts will set aside policies that aren’t commonsensically arbitrary but seem so to judges given their political or ideological sensibilities. Even more so now than when Rehnquist wrote, the latter error costs are greater than the former ones.

To see why, read headlines and case dockets on any big topic in domestic regulation—say, energy policy. The federal government supports or frustrates different methods for energy production depending on which political party holds the presidency. In 2024, the U.S. Department of Energy, under the Biden Administration, stopped further exports of liquified natural gas to countries with which the United States does not have a free trade agreement. These policies stood contrary to those set by Donald J. Trump in his first term as President. And soon after Trump was sworn in for a second term, he issued an executive order halting new approvals of new wind-energy projects previously favored by the Biden Administration.

Those policy changes illustrate perfectly the stakes of State Farm and hard look review. In 2024, 14 states sued in federal court challenging Biden’s no-new-liquid-gas-exports policy, arguing that the pause violated the APA, and in 2025 18 states sued to have Trump’s wind-energy order declared unlawful and set aside. In both cases, the states relied on State Farm and its rendition of hard look review. In December 2025, a federal district court ruled on summary judgment that Trump’s wind-energy order was unlawful and set the order aside. Under pre-1946 doctrine, a court would have needed to reject an arbitrariness challenge to the order if any state of facts could be imagined in which it could have been upheld. In the wind-energy case, however, the Interior Department and other responsible departments had no justification for changing federal wind policy besides a directive from President Trump, and under State Farm that was not enough to uphold the change Trump wanted.

For Professor Pierce, in disputes like the ones cited here, hard look review is desirable because it prevents the “horrible effects on society” that will follow as the country “lurch[es] from far left to far right every four to eight years.” Change does not trouble me as much as it does Pierce.

To begin with, I doubt that the effects of change are as dire as Pierce assumes. I do not doubt that the affected energy producers, such as producers of natural gas and wind energy, suffer inconvenience and lose money when the federal government stops favoring the forms of energy they produce. But society will not suffer “horrible effects” because the federal government changes its mind about energy policy every four to eight years; there are too many sources of energy for specific government decisions to disrupt energy markets horrifically. Nor should anyone have sympathy for the energy companies. They are sophisticated actors. Their leaders know the political and regulatory risks they assume when they ask for favorable decisions from the federal government.

Two benefits would follow if federal courts reviewed agency policy choices more or less as they did around 1946. One consists of regulatory experimentation. The executive branch has the discretion at issue in the gas export litigation discussed above thanks to the discretion Congress gave in the Natural Gas Act of 1938, and the Outer Continental Shelf Lands Act and several other statutes create similar discretion about federal policies toward windmills. When Congress assigns executive actors’ discretion as broad as is given in those statutes, the discretion is a “tell.” Key members of Congress and regulated parties cannot work out among themselves what general standards seem best suited to permit new liquid gas exports or new wind energy projects. And if the key legislators and constituencies cannot work out such general rules, different executives should be allowed to experiment with different policy approaches.

The other benefit comes from political values often disregarded in administrative law and policy scholarship—civic participation and republican self-governance, values often associated with  Alexis De Tocqueville. The regulatory disputes I described above raise hard policy questions about the desirability of hydrocarbon-energy production and alternative-energy production. Those questions are important, they engage voters, and elections are supposed to settle them.

But State Farm and hard look review make the federal government less responsive to electoral politics. Voters may think that they conferred electoral mandates on Presidents when, in reality, State Farm and hard look review let federal courts stop presidential administrations from making good on those mandates. And that unresponsiveness can foster many undesirable attitudes in politics—despair, disengagement, and contempt.

State Farm and hard look review have never been consistent with the text of the APA. A change in judicial review of agency policymaking would give everyone more information about different approaches to important federal regulatory questions, and it would remove a check on voters’ voice in the federal government. More than 40 years after State Farm, I wonder whether we should welcome experimentation and political participation. And, whether we should bid a fond farewell to hard look review.

Eric R. Claeys is a professor of law at the George Mason University Antonin Scalia Law School.