Trinity Industries, Inc. v. United States, 757 F.3d 400

Background

Trinity Industries claimed R&D tax credits under I.R.C. § 41 on amended returns for taxable years ending March 1994 and March 1995 based on vessel development projects; IRS denied the claims and Trinity sued for refunds.
The research credit depends on claim-year QREs exceeding a base amount computed using a fixed-base percentage derived from base-period QREs divided by base-period gross receipts (base years 1985–1989).
The district court conducted a two-phase bench trial. Phase I found two of six claim-year vessels (Mark V and Dirty Oil Barge) qualified as QREs; four others failed the process-of-experimentation (80%) requirement when analyzed as whole projects because Trinity did not provide subcomponent-level cost records.
Phase II addressed two additional claim-year vessels (Queen of New Orleans — allowed; Penn Tugs — disallowed) and the proper computation of base-period QREs. Trinity’s expert (Bennett) provided a base-period QRE total of $49,483,136 (fixed-base ≈ 1.2847%).
Trinity argued two consistency-rule theories: (1) it had consistently used an all-or-nothing (whole-project) approach for claim and base periods (no shrinking-back), and (2) post-Phase I, four base-period vessels identified by former employees (Nuss, Charters) were as non-experimental as the claim-year vessels disallowed and thus should be removed from the base-period QREs (reducing base-period QREs to $26,706,987).
The district court adopted Bennett’s base-period figures, awarded Trinity $135,787.60 for 1994 and $0 for 1995, denied credit for the Penn Tugs, and rejected Trinity’s first consistency argument but did not resolve fully the second (Nuss) consistency argument; the Fifth Circuit affirmed in part, vacated and remanded in part.

Issues

Issue	Plaintiff's Argument	Defendant's Argument	Held
Whether the district court violated the consistency rule by applying shrinking-back in base years but not claim years	Trinity: If shrinking-back was applied to base years but not claim years, the comparison is inconsistent and unlawful	Gov: The record shows neither the returns nor Bennett’s report used shrinking-back; district court did not apply it	Court: No inconsistency — the district court (via Bennett’s figures) used an all-or-nothing approach for base years, so Trinity’s first argument fails
Whether four base-period vessels should be excluded under the consistency rule because they are as non-experimental as four disallowed claim-year vessels (Nuss/Charters testimony)	Trinity: District court’s Phase I standard renders those four base-period vessels non-QREs; they must be removed from base-period QREs	Gov: District court credited Bennett’s base-period calculation and did not adopt Trinity’s reclassification; burden on Trinity to prove inconsistency	Court: Remand required — district court must make factual findings whether to credit Nuss/Charters; if credited, remove those vessels and recalculate base-period QREs (to $26,706,987)
Whether Penn Tugs satisfied the §41 process-of-experimentation (80%) requirement	Trinity: Penn Tugs involved systematic experimentation (hull redesign, modeling, multiple revisions) sufficient to meet 80% threshold	Gov: Much of the work relied on third-party coupler; modifications were to existing hulls and limited pilot-house changes — not predominantly systematic experimentation	Court: Affirmed district court — Trinity failed to prove substantially all activities (80%) were a process of experimentation; Penn Tugs are not QREs
Whether IRS revenue agent report finding Penn Tugs met the experimentation test is binding or presumptively correct	Trinity: IRS report finding the Penn Tugs met the experimentation test is at least presumptively correct or binds the Government	Gov: The Commissioner’s ultimate assessment is reviewed de novo in refund suits; subsidiary agent reports are not binding or presumptively correct	Court: Rejected Trinity’s claim — agent report is admissible but not binding or entitled to a presumption of correctness in district court refund litigation

Key Cases Cited

Coe v. Chesapeake Exploration, L.L.C., 695 F.3d 311 (5th Cir.) (bench-trial fact/legal-review standards)
Petrohawk Props., L.P. v. Chesapeake La., L.P., 689 F.3d 380 (5th Cir.) (clear-error and mixed-question standards)
United States v. McFerrin, 570 F.3d 672 (5th Cir.) (process-of-experimentation regulatory interpretation)
Union Carbide Corp. v. Commissioner, 697 F.3d 104 (2d Cir.) (examples of systematic experimentation analysis under §41)
Ornelas v. United States, 517 U.S. 690 (U.S.) (mixed question of law and fact framework)
Dunn v. Commissioner, 301 F.3d 339 (5th Cir.) (standards for mixed questions of law and fact)