31 F.4th 1203
9th Cir.2022Background
- NRDC petitioned EPA in 2009 to cancel registrations for tetrachlorvinphos (TCVP) in pet collars, citing neurodevelopmental risks to children from transfer of TCVP dust from treated pets.
- EPA took years to act; after delays and a voluntary remand, EPA issued a 2016 Risk Assessment identifying potential risks, then delayed further, prompting a second mandamus. The Ninth Circuit ordered EPA to issue a final response within 90 days in April 2020.
- EPA issued a 2020 Risk Assessment and denied NRDC’s petition, relying on Hartz’s post-market changes (one collar voluntarily canceled; six modified) and two Hartz studies: a Torsion Study and a Normal Wear Study.
- Key agency assumptions included: (1) only 0.38% of collar mass is released as dust, and (2) dust contains 14.6% TCVP (equal to collar-average TCVP), and (3) owners trim 20% of collars when fitting.
- NRDC challenged the denial, arguing EPA selectively used and misinterpreted Hartz’s studies, adopted unjustified assumptions, and relied on post-hoc rationalizations; the Ninth Circuit vacated EPA’s denial and remanded for a revised response within 120 days, ordering unsealing of the Torsion and Normal Wear Studies.
Issues
| Issue | Plaintiff's Argument | Defendant's Argument | Held |
|---|---|---|---|
| Whether EPA’s denial was supported by substantial evidence given selective use of the Torsion Study | EPA: denial unsupported because EPA ignored the Torsion Study’s finding that ~97.2% of released solids were TCVP and selectively adopted only some results | EPA: relied on Torsion Study for total dust released but rejected its composition result as exaggerated; pointed to Normal Wear Study as more representative | Vacated: EPA’s selective use of the Torsion Study was unexplained; agency failed to articulate rationale in the administrative record so denial lacks substantial evidence |
| Whether EPA reasonably assumed dust released from collars contained only 14.6% TCVP | EPA: measured evidence shows higher TCVP proportion in released dust; thus EPA’s 14.6% assumption understates exposure | EPA: post-hoc contends Normal Wear Study supports lower TCVP-in-dust fraction; Torsion Study exaggerates dust composition due to testing conditions | Vacated: Court rejected EPA’s 14.6% assumption as arbitrary and inadequately justified in the record; post-hoc rationalizations not considered |
| Whether EPA reasonably assumed pet owners trim collars by 20% when fitting | EPA: trimming reduces exposure; assumption drawn from a Hartz Efficacy Study showing 20–43% trimming and provides conservative estimate | EPA: defended using 20% as typical trimming | Vacated: Assumption departs from EPA’s prior practice and SOP without explanation, is unsupported by applicable data (Efficacy Study used different collars and sizes), and is arbitrary |
| Whether the agency may rely on explanations raised only in litigation briefing | NRDC: agency must provide the rationale in its decision; courts cannot accept post-hoc rationalizations | EPA: offered justifications in briefing relying on Normal Wear Study and alleged exaggeration in Torsion Study | Held: Court will not credit post-hoc rationalizations; decision must rest on reasons articulated in the administrative record |
Key Cases Cited
- Motor Vehicle Mfrs. Ass’n of the U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29 (agency must articulate rational connection between facts and decision)
- SEC v. Chenery Corp., 332 U.S. 194 (courts review agency action on the grounds the agency invoked)
- Fed. Power Comm’n v. Texaco Inc., 417 U.S. 380 (courts cannot accept post-hoc rationalizations)
- In re Nat. Res. Def. Council, Inc., 956 F.3d 1134 (9th Cir.) (prior NRDC mandamus and risk-assessment background)
- Nat. Res. Def. Council v. U.S. EPA, 735 F.3d 873 (agency action reviewed for substantial evidence)
- Nat. Res. Def. Council v. U.S. EPA, 857 F.3d 1030 (agency cannot rely on unsubstantiated assumptions or post-hoc reasoning)
- Ctr. for Biological Diversity v. Haaland, 998 F.3d 1061 (agency must provide reasoned explanation in record)
- Trs. of Cal. State Univ. v. Riley, 74 F.3d 960 (courts need not defer to arbitrary or highly inaccurate calculations)