With a new presidential administration comes a new wave of agency policy changes.1 Post-election agency reversals often attract legal challenges on a theory that the policy change was “arbitrary [or] capricious” under the Administrative Procedure Act2 (APA).3 Such “rationality review”4 litigation can put courts in a bind — charged with ensuring that agencies take a “hard look” at the facts and reach reasoned conclusions,5 but vulnerable to criticism for applying normally lenient rationality review too onerously in order to block an incoming administration’s agenda.6 In FCC v. Fox Television Stations, Inc.,7 the Supreme Court rejected the proposition that agency policy changes must clear a higher bar than agency actions on a blank slate. Rather, an agency need only provide “good reasons” for a new action, as long as the agency does not disregard prior factual findings on which its earlier action hinged.8 But whether an agency has disregarded such prior factual findings or has instead permissibly applied new normative judgments to the existing factual record is often a fraught question. On November 8, 2018, in Indigenous Environmental Network v. Department of State9 (IEN), a Montana District Court vacated the Trump State Department’s determination that construction of the Keystone XL pipeline was in the national interest — a reversal of the Obama State Department’s prior determination.10 The court’s holding that State’s new determination failed under Fox’s standard for rationality review of agency policy changes implicates the question of how — and when — judges should draw the line between agency disregard of prior factual findings and normative reevaluation. Under Fox, agencies reversing course must engage meaningfully with prior factual findings that the agency made as part of a technocratic inquiry, especially one delegated by Congress. But where, as in IEN, the prior agency findings at issue are judgment-based rather than technical and are made pursuant to core executive powers, or are highly specific to a point in time, courts should be more willing to accept conclusory agency treatment of prior findings. Requiring more in such cases risks ossifying precisely the sort of value-laden decisions that a new administration should be able to readily reevaluate under Fox.
On September 19, 2008, TransCanada Keystone Pipeline, L.P. applied for a presidential permit to construct the Keystone XL pipeline, which would carry Canadian crude oil from Alberta to Nebraska.11 So began a ten-year (and counting) permitting saga. Pursuant to Executive Order 13,337, the President has delegated authority to the State Department to determine whether certain cross-border projects are in the national interest — and if so, to issue a presidential permit.12 The “national interest” hurdle is a product of executive power,13 as opposed to, for example, a presidential finding required by statute.14
As State assessed the project,15 Keystone’s national and international profile grew. In the media and amongst environmental activists, Keystone became associated with increased production of particularly “dirty” (carbon-intensive) oil at the Canadian tar sands.16
On November 3, 2015, State issued a Record of Decision (ROD) determining that approval of Keystone was not in the national interest.17 State concluded that Keystone’s benefits did not “outweigh the fact that approval would undermine the United States’ successful foreign policy engagement in efforts to combat climate change.”18 The 2015 ROD made clear that State’s decision hinged on the widely held perception that Keystone would significantly increase greenhouse gas (GHG) emissions, rather than State’s own estimate of the project’s climate impact.19 The Paris climate summit — aimed at establishing a framework for reducing global GHG emissions — was scheduled to begin less than a month from the issuance of the 2015 ROD, making the permit denial particularly timely for diplomatic purposes.20
On January 24, 2017, President Trump formally invited TransCanada to reapply.21 On March 23, 2017, State issued a new ROD determining that approval served the national interest.22 On most issues — such as Keystone’s impacts on GHG emissions, environmental and cultural resources, and the economy — the 2017 ROD closely tracked the 2015 ROD and relied on the same factual record.23 But where the 2015 ROD devoted three pages to “Climate Change–Related Foreign Policy Considerations,”24 the 2017 ROD had a single paragraph acknowledging State’s 2015 determination that approval would “undercut” U.S. “credibility and influence” in climate diplomacy, and stating that “[s]ince then, there [had] been numerous developments related to global action to address climate change, including announcements by many countries of their plans to do so.”25 The paragraph continued: “In this changed global context, [approval] at this time would not undermine U.S. objectives in this area.”26 The 2017 ROD did not mention the Paris Climate Accord, from which the U.S. withdrew about two months later.27
IEN and other groups sued in the District of Montana, seeking to invalidate the permit under multiple legal theories.28 Plaintiffs argued the 2017 ROD was deficient under the National Environmental Policy Act29 (NEPA), the Endangered Species Act30 (ESA), and the APA.31
On November 8, 2018, Judge Morris enjoined construction of the pipeline and vacated State’s 2017 national interest determination.32 He rejected various plaintiff theories under NEPA and the ESA,33 but ruled for plaintiffs on six issues: under the APA regarding State’s reasoning for reversing course, on four issues under NEPA,34 and on one issue under the ESA.35 In earlier proceedings, the court had rejected the government’s argument that the permit decision was unreviewable as action pursuant to inherent presidential authority.36
In finding State’s explanation of its reversal defective, the court applied Fox’s framework for conducting rationality review of agency policy changes. At issue was whether State had provided “good reason” for adopting its new decision.37 Per Fox, when a “new policy rests upon factual findings that contradict those which underlay [the agency’s] prior policy . . . [a] reasoned explanation is needed for disregarding” those earlier findings.38 If a policy change is rooted in normative judgments, by comparison, appellate courts have read Fox to permit agency reversal “even on precisely the same record.”39 The challenge, post-Fox, is distinguishing those two scenarios.
Here, the court held that the 2017 ROD had impermissibly disregarded the 2015 ROD’s “prior factual findings related to climate change” — referring to the 2015 ROD’s discussion of climate diplomacy.40 While the court acknowledged that State could prioritize factors other than climate diplomacy, it concluded that the 2017 ROD unlawfully “ignore[d] the 2015 ROD’s conclusion that 2015 represented a critical time for action.”41 The court found insufficient the 2017 ROD’s paragraph addressing climate diplomacy, describing it as “conclusory.”42
The government, TransCanada, and multiple plaintiffs appealed.43 Then, on March 29, 2019, President Trump revoked the March 23, 2017 permit, and issued a new permit for Keystone without delegating a national interest determination to State.44 Further litigation seems likely.45
IEN’s application of Fox presents a puzzle. In Fox, the Supreme Court signaled that courts conducting rationality review should uphold agency policy changes grounded in the application of new normative judgments to existing facts,46 but must also ensure that agencies do not disregard prior factual findings underlying their earlier actions.47 Summarized, the policy change at issue in IEN was that President Trump — about to withdraw from the Paris Climate Accord — decided that risks to climate diplomacy did not justify blocking Keystone. The puzzle: How did the court in IEN, applying Fox, find itself preventing the sort of policy-driven reevaluation by a new administration that Fox facilitates? Under Fox, courts should reserve heightened scrutiny for agency reversals that implicate prior factual findings made pursuant to technocratic inquiries, especially ones delegated by Congress. In such cases, courts should require clarity from agencies about how their new policies square with those findings. But where prior findings — like State’s 2015 conclusions on Keystone’s consequences for climate diplomacy — amount to nontechnocratic judgment calls committed to the Executive, or are inherently context-specific to a point in time,48 courts can best implement Fox’s framework for rationality review of agency policy changes by crediting even conclusory agency statements that address such prior findings. Courts may thereby make the elusive line between normative reevaluation and disregard of factual findings by agencies easier to draw, and they will need to draw it less often.
In the Supreme Court’s deferential rationality review jurisprudence, Fox’s call for special scrutiny of an agency’s treatment of its prior factual findings is notable. In judicial review of agency action, an agency’s responsiveness to political control and its technocratic expertise usually both weigh in favor of deference.49 Rationality review is typically lenient; agencies generally win at the Supreme Court,50 and courts are especially deferential where agencies rely on technical expertise.51 Agency reversals represent a special case: they may pit democratic responsiveness against respect for agency expertise (and Congress’s decision to delegate a technocratic inquiry) — if the agency ignores prior technocratic findings to achieve a values-driven outcome.
Fox sought to balance that tension. On the one hand, the Court rejected the proposition that policy changes must be held to a tougher rationality standard than blank-slate agency action. Applying a higher bar where an agency implements a new administration’s normative priorities would undermine the administrative state’s responsiveness to supervision by the elected President.52 Fox consolidated an important shift. Courts conducting rationality review were historically suspicious of normative reasoning creeping into technocratic inquiries — and agencies accordingly pretended only technocratic analysis was at work.53 Fox invited agency frankness about normative reasoning.54 But still, to deliver the rule-of-law values of rationality review, judges must ensure that an agency has not, in the words of Justice Kennedy’s Fox concurrence, “simply disregard[ed] . . . factual determinations that [the agency] made in the past,” just as a court would not permit an agency to “ignore inconvenient facts when it writes on a blank slate.”55 Hence Fox’s call for close scrutiny of agency disregard of prior factual findings.
Deciding when to trigger that heightened scrutiny can be tricky: whether an agency has disregarded its prior factual findings is often in the eye of the beholder. Take Organized Village of Kake v. USDA,56 a post-Fox case in which the en banc Ninth Circuit split 6–5 in striking down a USDA policy reversal as arbitrary and capricious57 — and on which, along with Fox, IEN’s analysis of State’s policy change relied.58 In Kake, the USDA issued a 2003 ROD reversing its prior determination in a 2001 ROD, relying on “precisely the same record.”59 In deciding not to exempt certain forestland from a rule banning road construction in protected areas, the 2001 ROD determined that the environmental risks outweighed economic considerations; the 2003 ROD, in contrast, concluded that socioeconomic costs to lumber harvest–dependent communities outweighed the “minor” risks to conservation objectives.60
The fact pattern proved to be a Rorschach test. The majority saw disregard of prior factual findings; the 2003 ROD referred to “minor” conservation risks whereas the 2001 ROD had identified “high” risk.61 A dissent accused the majority of “latch[ing] onto” the word “minor,” when “it [was] clear” that the 2003 ROD permissibly reweighed conservation and socioeconomic interests on the existing record.62
State’s 2017 ROD presented just such a Rorschach test: Did its paragraph on climate diplomacy say that (i) Keystone in fact no longer posed a serious risk to U.S. diplomacy aimed at promoting aggressive action on climate change, or that (ii) State now believed the project’s benefits outweighed any such risk (or that such risk was irrelevant given new U.S. diplomatic priorities)?63 If the former, and State’s 2015 conclusions on climate diplomacy were “prior factual findings” requiring especially well-reasoned treatment under Fox, State’s paragraph would seem insufficient. The latter would be a permissible policy-based judgment, and clearly would provide the correct real-world account — given President Trump’s invitation to TransCanada to reapply and his subsequent withdrawal from the Paris Accord.64 State could have simplified the case by making its normative reevaluation clear65 or detailing new factual findings. But State declined to take the easy way out, instead penning its ambiguous paragraph.66 Though State’s ambiguity invited skepticism, there are good reasons the court in IEN should have accepted the 2017 ROD’s conclusory treatment of State’s 2015 climate diplomacy findings and avoided tough line-drawing.
The 2015 ROD’s climate diplomacy conclusions are not the sort of prior factual findings that should trigger close scrutiny under Fox — for two reasons. First, State made a judgment-based determination delegated by the President pursuant to his constitutional powers, rather than a technocratic inquiry delegated by Congress. As Justice Kennedy emphasized in Fox, agencies endanger the separation of powers when they wander outside the “specific and detailed” mandates assigned to them by Congress — aggrandizing the power of the executive or administrative state at Congress’s expense.67 Rationality review under the APA guards against that risk,68 which looms when an agency ignores the results of its delegated technocratic analysis. But where agencies act pursuant to inherent executive authority, such concerns do not exist.69
On this point, State Farm70 is a helpful foil to IEN. In State Farm, Congress had charged the National Highway Traffic Safety Administration (NHTSA) with a technocratic mandate — to regulate to “meet the need for motor vehicle safety” — and NHTSA had issued a rule mandating air bags or automatic seatbelts in cars.71 The NHTSA — under a new President — eliminated the rule with “[n]ot one sentence” addressing its extensive prior findings on the safety benefits of air bags.72 NHTSA’s disregard of its prior findings is exactly the sort of arbitrariness against which rationality review under Fox guards.73 The agency was empowered by Congress to govern in a particular sphere based on technical analysis; it made extensive findings, then ignored them.
Now contrast IEN. State acted pursuant to an executive order delegating authority to decide whether projects “serve the national interest.”74 State’s 2015 determination — and IEN’s holding under Fox — turned on State’s assessment of foreign perceptions of the pipeline and the diplomatic risks of approval. Whereas in State Farm, triggering heightened scrutiny under Fox would serve to ensure that NHTSA acted within its technocratic mandate from Congress, in IEN, heightened scrutiny ended up burdening the President’s inherent authority to modify (or delegate modification of) a diplomatic determination.
Second, State’s 2015 climate diplomacy findings should not have triggered heightened scrutiny under Fox because they were time-bound. Unlike findings about air bags’ safety benefits — presumably roughly as true three years later as when first made, absent significant shifts in technology and usage — diplomatic findings are inherently time- and context-specific. State’s decision to send a diplomatic message by rejecting a controversial project one month before the Paris talks hinged on a point-in-time assessment of the perceptions of multiple foreign actors. It seems odd to require agencies to consider in detail prior factual findings of a sort that inevitably and rapidly become stale.75
Had the IEN court determined that State’s 2015 climate diplomacy conclusions were not prior factual findings requiring rigorous treatment under Fox, the usual lenient form of rationality review would have applied. If the 2017 ROD — based largely on existing environmental and economic findings — satisfied that standard it should have been upheld. State should have been free in 2017 to treat the issue of climate diplomacy in a conclusory manner,76 as it did — and as it could have if perceptions of Keystone had been raised in a public comment.77
There is much to be said for implementing Fox’s lenient rationality review framework for post-election agency reversals.78 Fox allows for the readier injection of democratic responsiveness into agency policymaking, and allows courts to play a reduced role in policing agencies’ politically salient normative judgments — absent State Farm–style disregard for technocratic findings.79 Rationality review under the APA — aimed at preventing arbitrariness by technocratic agencies — should not ossify a prior decision to deny a pipeline permit for political reasons related to foreign perceptions of the pipeline. Courts can realize Fox’s promise by applying tough scrutiny only where the disregarded factual findings were made pursuant to a technocratic inquiry, especially one delegated by Congress. Where such findings are implicated, agencies making new value-based judgments on existing facts should make explicit that they are doing so — and courts may be well-advised to incentivize clarity by punishing agencies where they do not make their normative reasoning explicit,80 as did the court in Kake. But where an agency dismisses a prior finding inseparable from a specific point in time, or that was a judgment call implicating inherent executive authority, a conclusory statement showing awareness and rejection or de-prioritization of the prior finding should suffice, as long as the agency provides other good reasons for its new course of action.