Little Sandy Coal v. Commissioner: the substantially-all ratio has teeth

Little Sandy Coal Co. v. Commissioner, 62 F.4th 287 (7th Cir. 2023), aff'g T.C. Memo 2021-15, is the decision that gave the "substantially all" requirement of Section 41(d)(1)(C) real bite. A shipbuilder claimed credits for eleven first-of-a-kind vessels — a novel tanker design, a dry dock — and lost entirely, not because the vessels were unoriginal, but because the taxpayer never showed that substantially all of its development activities constituted elements of a process of experimentation.

The dispute

Corn Island Shipyard (whose results flowed to Little Sandy Coal) designed and built vessels it had never built before, claiming that because each vessel was new and its development uncertain, the entire cost of designing and constructing it was qualified research. The IRS conceded some genuine uncertainty existed but argued the claim swept in vast amounts of routine fabrication — welding, assembly, production labor — that was not experimentation, and that the taxpayer had made no attempt to show the 80 percent threshold of Treas. Reg. §1.41-4(a)(6) was met.

The holding

The Tax Court denied the credits and the Seventh Circuit affirmed. The court of appeals held that "substantially all" modifies activities: at least 80 percent of the research activities for the claimed business component, measured on a cost or comparable basis, must constitute elements of a process of experimentation. Activities that merely support or result from experimentation — building the vessel a design experiment specified — do not automatically count. Because the shipbuilder presented no evidence allowing the fraction to be estimated for any vessel, and did not invoke the shrink-back rule to claim discrete subsystems, the entire claim failed.

The reasoning that matters

The Seventh Circuit rejected the taxpayer's central syllogism — the vessel was novel, novelty implies uncertainty, therefore everything was research. Uncertainty may have pervaded the design, but Section 41(d)(1)(C) asks how much of the activity was experimentation. A production worker welding plates to an already-resolved specification is not experimenting, even on a first-of-a-kind hull. The court acknowledged that some construction labor can be part of experimentation — building a prototype to test it, as in Union Carbide's plant trials — but the taxpayer must connect the labor to the evaluative process and quantify it. The opinion also underscored the escape valve the taxpayer ignored: the shrink-back rule of Treas. Reg. §1.41-4(b)(2), which applies the test to the most significant subset of elements that satisfies it. A claim for the novel bow-loading system might have survived; the claim for the whole vessel could not.

What it means for claims today

Little Sandy Coal is now the government's lead citation against whole-project claims in shipbuilding, construction, aerospace, custom machinery — any industry where large fabrication costs ride alongside genuine engineering uncertainty. Its practical demands are concrete. Define business components deliberately, at a level where the 80 percent test is plausible. Document which activities were experimentation — alternatives considered, tests run, redesigns triggered — and estimate the experimentation fraction with a defensible method. Where the whole product fails, shrink back and claim the subsystems that pass. The Section G business-component detail on the redesigned Form 6765 effectively forces this analysis at filing, before an examiner asks.

Related cases on the site

The substantially-all doctrine sits within the framework covered in the four-part test explained; the full litigation landscape is in the research credit case law map. Union Carbide shows prototype and trial construction qualifying when tied to experiments; Siemer Milling is the parallel failure of proof on the experimentation prong itself. For responding to a substantially-all challenge on exam, see research credit audit defense.