72 F.4th 1324
D.C. Cir.2023Background
- DOE issued a Final Rule (2020) tightening energy-efficiency standards for commercial packaged boilers under EPCA’s six‑year “lookback” authority and used a Monte Carlo life‑cycle‑cost model (10,000 trials) to compare Base Case vs New Standards Case.
- Key model inputs included random assignment of boiler efficiency levels to sampled buildings, fuel price projections from EIA, and an assumed building heat‑load of 30 Btu/h per heated square foot to derive burner operating hours.
- In APGA I (D.C. Cir. 2022) this Court remanded, directing DOE to respond to three main challenges: the use of random assignment (consumer choice modeling), whether fuel‑price data capture discounts paid by large purchasers, and anomalous burner operating‑hour estimates.
- On remand DOE published a Supplement relying on additional studies and datasets to defend random assignment, gave a fuller explanation of fuel‑price data, but did not explain the 30 Btu/h heat‑load assumption.
- Petitioners (APGA, AHRI, Spire) challenged the Supplement, arguing DOE violated the APA by relying on new evidence without notice‑and‑comment and again failed to justify critical assumptions. The D.C. Circuit granted the petitions and vacated the Final Rule and Supplement.
Issues
| Issue | Plaintiff's Argument | Defendant's Argument | Held |
|---|---|---|---|
| Whether DOE’s reliance on new studies/datasets in its remand Supplement required new notice‑and‑comment under the APA | DOE relied on new, critical empirical materials not previously in the record; notice‑and‑comment was required | DOE: Supplement merely bolstered an existing hypothesis; Building‑industry/Fabricare precedents allow additional support without new comment; good‑cause or tight remand deadline excused notice | The Court: DOE relied on entirely new information critical to its decision; APA notice‑and‑comment required and good cause not shown; failure prejudicial — vacatur warranted |
| Whether DOE gave a cogent, reasoned response to criticism of using random assignment for boiler purchases | Random assignment ignores rational purchaser behavior; DOE failed to show market failure specific to boilers | DOE: Supplement cited studies and datasets demonstrating market failures and corroborating data showing weak correlation of building characteristics with boiler efficiency | Court: Because DOE used new studies on remand without comment, the Court remanded on notice ground and did not fully resolve merits; DOE’s prior response was inadequate per APGA I |
| Whether DOE’s fuel‑price data overstated savings by failing to capture lower marginal prices for large industrial purchasers | Petitioners: EIA data exclude some manufacturing/industrial consumers and thus overstate energy savings from stricter standards | DOE: EIA data cover all utilities/customers in aggregate and include discounts; adjustments and marginal price factors address differences | Court: DOE’s supplemented explanation was cogent and adequate; petitioners’ new rebuttal argument was forfeited |
| Whether DOE adequately justified its 30 Btu/h heat‑load assumption used to derive burner operating hours | Petitioners: The 30 Btu/h assumption is unsupported and produced anomalous burner‑hour outputs; DOE failed to address these substantive comments | DOE: Burner hours are a derived quantity from building data; asserted minimal effect or relied on CBECs data explanations | Held: DOE failed to explain or justify the 30 Btu/h assumption on remand; omission is arbitrary and capricious and contributed to vacatur |
Key Cases Cited
- Am. Pub. Gas Ass'n v. DOE, 22 F.4th 1018 (D.C. Cir. 2022) (prior opinion remanding DOE to address three specific modeling and data concerns)
- Chamber of Com. of U.S. v. SEC, 443 F.3d 890 (D.C. Cir. 2006) (agency technical studies relied on must be disclosed for public evaluation)
- Bldg. Indus. Ass'n of Superior Cal. v. Norton, 247 F.3d 1241 (D.C. Cir. 2001) (when final rule is logical outgrowth, extra notice may not be required)
- Int'l Fabricare Inst. v. EPA, 972 F.2d 384 (D.C. Cir. 1992) (post‑proposal studies may be used without new comment where issue was noticed and commenters had fair opportunity)
- Methodist Hosp. of Sacramento v. Shalala, 38 F.3d 1225 (D.C. Cir. 1994) (good‑cause exception to notice‑and‑comment narrowly construed; tight deadlines alone insufficient)
- Mack Trucks, Inc. v. EPA, 682 F.3d 87 (D.C. Cir. 2012) (good‑cause exception to notice‑and‑comment to be narrowly applied)
- Sugar Cane Growers Co‑op. of Fla. v. Veneman, 289 F.3d 89 (D.C. Cir. 2002) (failure to comply with notice and comment is not harmless if any uncertainty exists)
- Daimler Trucks N. Am. LLC v. EPA, 737 F.3d 95 (D.C. Cir. 2013) (court typically vacates rules when agency entirely fails to provide notice and comment)
- Camp v. Pitts, 411 U.S. 138 (1973) (review focuses on administrative record before the agency)
- Allina Health Servs. v. Sebelius, 746 F.3d 1102 (D.C. Cir. 2014) (vacatur is the normal remedy for unlawful agency action)
