The Docket · Brief · Working level
Siemer Milling v. Commissioner: the documentation cautionary tale
Siemer Milling, T.C. Memo 2019-37, denied research credits for a flour miller's product and process projects because the record showed no process of experimentation and no reliance on hard science — the standard citation for what happens when documentation is missing.
Siemer Milling Co. v. Commissioner, T.C. Memo 2019-37, is the research credit's documentation cautionary tale. An Illinois flour miller claimed credits for seven projects — new flour varieties, a heat-treatment process for wheat flour, milling process improvements — and the Tax Court denied every one, holding the taxpayer failed to prove either that the work was technological in nature or that it involved a process of experimentation under Section 41(d). Nothing in the opinion says milling cannot involve qualified research. Everything in it says this record did not show that it did.
The dispute
Siemer Milling claimed roughly $235,000 in research credits across two years for projects that sounded plausibly scientific: developing new flour formulations for specific baking properties, applying heat treatment to reduce microbial content, improving wheat-cleaning and milling throughput, and related efforts. The IRS disallowed the credits, contending the projects were routine product tweaks and quality-control work, and that the taxpayer could not carry its burden on the four-part test for any of them.
The holding
The Tax Court agreed with the government on every project. For most, the fatal prong was the process of experimentation: the record contained no evidence that Siemer Milling identified uncertainties, formulated hypotheses, designed tests, or evaluated alternatives in any systematic way. Employees testified in general terms that they tried adjustments and observed results, but the court found this indistinguishable from ordinary trial-and-error production management. Several projects independently failed the technological-in-nature prong because the taxpayer never showed the work fundamentally relied on principles of chemistry, biology, or engineering rather than accumulated operator know-how. With qualified research unproven, the court had no occasion to estimate expenses.
The reasoning that matters
The opinion is a tour of what Treas. Reg. §1.41-4(a)(5) means by a process of experimentation: identifying the uncertainty, identifying alternatives capable of eliminating it, and evaluating those alternatives through modeling, simulation, or systematic trial and error. The court stressed that "systematic" is the operative word — changing an oven temperature and seeing what happens is not a scientific method unless it sits inside a designed evaluative process. Equally important is what the court did not require: it did not demand lab notebooks or any particular record format, and it accepted that testimony could suffice in principle. The taxpayer's problem was that its testimony was conclusory. Witnesses asserted that experimentation happened; none could describe the hypotheses tested, the alternatives compared, or the criteria for success on any specific project.
What it means for claims today
Siemer Milling is the case examiners have in mind when an Information Document Request asks the taxpayer to "describe the alternatives evaluated and the method of evaluating them" for each business component — language that now shadows the Section G reporting on the redesigned Form 6765. Food processors, job shops, and other companies whose innovation lives in operator experience rather than engineering files are the most exposed. The practical prescriptions are direct. Capture the uncertainty and the candidate solutions at project start, even in a one-page memo. Record what was tested and against what criteria. Prepare witnesses to speak to specific projects in specific terms, because generalities lose. And treat Siemer Milling as the floor for a credit study: a study that merely lists projects and allocates costs, without evidence of the experimental process, documents a claim the Tax Court has already rejected once. The IRS research credit overview frames the same four-part test the court applied.
Related cases on the site
Siemer Milling is the failure-of-proof counterpart to Union Carbide and Suder, where documented experimental processes won; all three are situated in the research credit case law map. The prongs at issue are unpacked in the four-part test explained. Little Sandy Coal extends the evidentiary demand to quantifying the experimentation fraction. For building the record before exam, see research credit audit defense.
Frequently asked questions
- Why did Siemer Milling lose its research credit case?
- In Siemer Milling Co. v. Commissioner, T.C. Memo 2019-37, the Tax Court denied credits for all seven claimed projects because the taxpayer produced no evidence of a systematic process of experimentation — no hypotheses, no testing protocols, no evaluation of alternatives — and could not show the work fundamentally relied on principles of engineering or physical science.
- Does the research credit require written documentation?
- No statute or regulation requires contemporaneous written records, and testimony can carry a claim. But Siemer Milling shows the practical rule: without records or specific credible testimony describing hypotheses, alternatives, and evaluation methods, courts will find the process-of-experimentation and technological-in-nature prongs unproven and deny the credit entirely.