“As Necessary” Means Discretion: Upholding FRA’s Two-Person Default Crew Rule Under Deferential APA Review

Eleventh Circuit (Aug. 11, 2026) — consolidated petitions challenging FRA’s 2024 “Crew Size Rule”

I. Introduction

This consolidated proceeding arose from multiple petitions for review filed by railroads and industry trade associations (including Texas & Northern Railway Company, Union Pacific Railroad Co., BNSF Railway Company, and others) challenging the Federal Railroad Administration’s 2024 “Crew Size Rule,” published at 89 Fed. Reg. 25052 (Apr. 9, 2024) and codified principally at 49 C.F.R. § 218.123 et seq.

The Rule establishes a nationwide default of at least two crewmembers onboard for covered train operations, while permitting one-person crews if either (i) certain Class II/III operations qualify for notice-based pathways (including “legacy” operations), or (ii) railroads obtain FRA “special approval” by demonstrating that the proposed one-person operation is “as safe or safer” than two-person operations.

Petitioners’ central claims sounded in the Administrative Procedure Act (“APA”): that FRA (1) exceeded its statutory authority under the Federal Railroad Safety Act (“FRSA”), especially the phrase “as necessary” in 49 U.S.C. § 20103(a), and (2) acted arbitrarily and capriciously by relying on an allegedly insufficient record, failing to address costs, conflicting with another statute’s risk-reduction framework, and promulgating the Rule outside a purported 12-month deadline.

II. Summary of the Opinion

The Eleventh Circuit denied all petitions for review and upheld the Crew Size Rule. The court held:

• Statutory authority: FRA’s general power to issue safety regulations “as necessary” in 49 U.S.C. § 20103(a) is a broad discretion-conferring grant, not a strict “essential-to-safety” evidentiary bar.
• Reasoned decisionmaking: FRA reasonably explained why a two-person default with exemptions and pathways to one-person operations promotes safety, relying on research, operational trends, and extensive public comments; the APA does not require gold-standard quantitative proof before acting.
• Change-in-position doctrine: Petitioners could not trigger heightened scrutiny based on the 2019 Withdrawal Order because that order was vacated; in any event, FRA offered reasonable explanations for regulating now.
• Costs: Petitioners’ “foregone labor savings” theory rested on speculation that FRA would block one-person operations; FRA’s consideration of costs was adequate under the governing framework.
• No conflict with risk-reduction program: The Rail Safety Improvement Act’s risk-reduction programs coexist with, rather than displace, federal safety regulations.
• Timing: Missing a 12-month regulatory deadline does not invalidate the Rule absent a specified statutory consequence; the remedy for delay is generally an action to compel under 5 U.S.C. § 706(1).
• Short-line challenges: Several arguments were forfeited for failure to raise them during rulemaking; remaining objections to the hazardous-materials legacy framework and expanded alerter requirement failed on the merits.

A dissent argued the Rule was arbitrary and capricious because FRA relied on substantially the same evidence it had previously found insufficient in 2019 and because FRA’s cost analysis allegedly ignored labor costs.

III. Analysis

A. Precedents Cited (and How They Shaped the Decision)

B. Legal Reasoning

1) Statutory authority: “as necessary” is discretionary, not “essential”

The court’s statutory analysis turned on sentence-level context. Without “as necessary,” § 20103(a) would read as a blunt, arguably unrealistic mandate to regulate “for every area of railroad safety.” The parenthetical phrase “as necessary” was therefore read as a tempering modifier allocating discretion about when and on what to regulate—not as a “stringent evidentiary bar” requiring proof that a rule is “essential.”

Operationally, the court applied the familiar general-grant standard from Mourning v. Family Publ'ns Serv., Inc.: a regulation is valid if “reasonably related to the purposes of the enabling legislation.” Given FRSA’s strong safety purpose (49 U.S.C. § 20101) and Congress’s later instruction to prioritize safety (49 U.S.C. § 103(c)), the court treated crew-size regulation as comfortably within the statutory domain.

2) APA arbitrariness: reasonableness, not statistical certainty

The opinion emphasized that APA review is not a demand for statistical certainty or new agency studies. Relying on FCC v. Prometheus Radio Project and Citadel Sec. LLC v. SEC, the court held that FRA could act based on safety research, operational trends (including train length and human-factor accident rates), technological uncertainty (PTC outages and learning-curve concerns), and extensive public comments describing safety functions performed by a second crewmember.

The court’s reasoning is notably prophylactic: echoing Miami-Dade County v. EPA, Ethyl Corp. v. EPA, and Stilwell v. Off. of Thrift Supervision, it framed safety regulation as legitimately “precautionary in nature,” allowing agencies to prevent catastrophic outcomes without waiting for a statistically robust run of disasters.

3) Change-in-position: vacatur breaks the “existing policy” anchor (majority), but FRA also explained itself

Petitioners sought heightened scrutiny under FCC v. Fox Television Stations, Inc. and Encino Motorcars, LLC v. Navarro by arguing FRA contradicted its 2019 view. The majority rejected the premise that the 2019 Withdrawal Order constituted “existing policy” because it had been vacated in SMART-TD I, and, under vacatur principles (invoking United States v. Sigma Int'l, Inc.), a vacated action has no remaining legal force.

Even assuming a change, the court held FRA supplied “good reasons”: it reassessed safety research and considered more recent conditions (accident trends, longer trains, and PTC uncertainties), and adopted a calibrated approach (default two-person, but one-person allowed if “as safe or safer”).

4) Costs: rejecting “foregone labor savings” as a required quantified category

The majority treated petitioners’ “largest cost” theory as resting on two premises it found unproven: (i) that Michigan v. EPA compelled a particular cost analysis under FRSA’s “as necessary” text, and (ii) that FRA’s special-approval pathway was effectively illusory. The court rejected the first as a mismatch in statutory language and history (Congress removed “appropriate” from the FRSA rulemaking clause), and rejected the second as speculation insufficient to overcome the presumption of regularity.

5) Timing: no automatic invalidation for missing the 12-month target

On the “12-month” claim under 49 U.S.C. § 20103(b) and FRA’s implementing regulation, the court followed the reasoning of SMART-TD II and SMART-TD III and relied on Barnhart v. Peabody Coal Company: absent an express consequence, courts will not invent vacatur as a sanction for tardiness. The appropriate remedy would have been an action to compel unlawfully withheld or unreasonably delayed agency action under 5 U.S.C. § 706(1).

6) Short-line issues: forfeiture and reasonableness of hazardous-materials distinctions

The court enforced administrative issue preservation: arguments not raised during notice-and-comment were forfeited. On the merits, it upheld FRA’s differential treatment of hazardous materials as rational given their higher consequences and the reasoned choice to require either a multi-year operational record or an individualized risk assessment before allowing one-person hazmat operations.

7) Alerters: adequate explanation for expanded requirement

Short Line Petitioners argued FRA changed policy by requiring alerters even for trains operating under 25 mph (beyond the earlier 49 C.F.R. § 229.140(a) threshold). The court held FRA acknowledged the change and reasonably justified it: in a one-person operation, an incapacitated engineer can create an uncontrolled movement risk, and an alerter is a direct mitigation; FRA also addressed costs and extended compliance time for certain legacy operations.

C. Impact

• Textual-structural interpretation of “as necessary”: The opinion concretely frames “as necessary” in FRSA as a discretionary grant rather than an evidentiary gatekeeping requirement. Future challengers to FRA safety rules (and potentially other Title 49 regimes) will face an uphill battle arguing that “necessary” imposes a strict proof threshold.
• Reinforcement of precautionary safety regulation: By endorsing prophylactic rulemaking logic—especially where catastrophic harms are plausible but statistically rare—the decision strengthens agencies’ ability to regulate amid uncertainty without waiting for high-frequency, high-quality datasets.
• Administrative record strategy: The case underscores the evidentiary weight courts will allow to public comments and testimony in technical safety rulemakings, and the importance of preserving arguments in the administrative process (forfeiture rulings were outcome-relevant).
• Deadline challenges constrained: The court’s Barnhart-based analysis narrows “missed deadline” attacks on completed rules; litigants must generally pursue compel-delay remedies rather than vacatur after issuance.
• Practical regulatory consequences: FRA’s “two-person default, one-person by safety equivalence” architecture likely becomes the model for future FRA rules addressing operational shifts driven by automation and remote technologies.

IV. Complex Concepts Simplified

• “As necessary” (rulemaking clause): Here it means FRA may regulate when it reasonably judges regulation is warranted for safety. It does not mean FRA must prove a rule is the only way—or “essential”—to achieve safety.
• “Arbitrary and capricious” review: Courts do not decide whether a rule is the best policy. They ask whether the agency considered the important issues and gave a rational explanation grounded in the record.
• “Change-in-position doctrine”: When an agency changes policy, it must acknowledge the change and provide good reasons. Sometimes it must provide more detail—especially if it contradicts key factual premises it relied on before or disrupts reliance interests.
• “Vacatur”: When a court vacates an agency action, it is treated as having no legal effect. The majority used that to say the 2019 Withdrawal Order could not operate as an “existing policy” anchor for heightened change-of-position scrutiny.
• “Special approval” + “as safe or safer”: FRA does not ban one-person crews outright; it conditions them on an individualized showing of safety equivalence (or superiority) relative to the two-person baseline.
• “Forfeiture” in administrative law: If a party does not raise an issue during the rulemaking process (when the agency can respond), courts often will not consider it later in a petition for review.

V. Conclusion

The Eleventh Circuit’s decision cements a deferential framework for FRA railroad-safety rulemaking under FRSA: “as necessary” is a discretion-conferring phrase, not a demand for conclusive empirical proof. The court upheld FRA’s Crew Size Rule as a reasonable, precautionary safety regulation supported by research, operational trends, and public comment, while rejecting challenges based on alleged policy reversal, insufficient data, uncounted “foregone savings,” statutory conflict, and missed timing targets.

The dissent highlights the litigation frontier going forward—how much explanation and cost accounting should be required when an agency re-enters a disputed policy space—but the majority’s approach signals that, in railroad safety, courts will not require agencies to wait for catastrophe (or perfect data) before regulating.