A famous designer. A nine-step production process. An expert witness who admitted, under oath, that he never verified the facts in his own report.
That’s how the credibility of an R&D tax credit claim begins to unravel before the court even reaches the technical questions.
Leon Max built one of the more recognizable names in American fashion, producing up to a million garments a month across multiple brands sold at Macy’s, Dillard’s, and Belk. In 2011 and 2012, his company claimed $748,955 in research credits for its design process. The Tax Court disallowed the entire claim. The 2021 decision is one of the clearest warnings on record about what happens when an R&D tax credit claim is built on assumptions instead of evidence.
The Business Behind the Claim
Leon Max, Inc. ran a genuinely structured, nine-step process to turn a sketch into a finished garment:
- Concept planning and mood boards
- Design sketching
- First pattern and pattern cutting
- First sample and fitting
- Sale sample
- Production prototype
- Marking and grading
Along the way, the company tested fabric shrinkage, colorfastness, seam strength, and pilling, and worked through real trial and error on things like pleat placement and twisted shoulder straps.
In 2013, the company engaged a tax consulting firm to evaluate whether its design process qualified for the R&D tax credit. The firm sampled 35 garments and concluded 32 of them included qualifying activity. Mr. Max’s amended returns matched those findings almost exactly.
The Expert Witness Problem
Mr. Max hired a fabric science professor to back up the claim. On the stand, he admitted something that sank the whole case:
- He relied heavily on the consulting firm’s description of the company’s production process
- He never verified it against what actually happened at the company
- He had access to depositions from LMI’s own employees and didn’t use them
- He based his conclusions largely on one clothing line, despite the claim covering the company’s entire portfolio
The Commissioner’s expert, by contrast, had 35 years in the fashion industry and testified that everything in the study was standard, ordinary practice across the apparel trade, not evidence of scientific research.
In plain terms, the court concluded that the claim relied too heavily on an industry-wide study instead of company-specific evidence demonstrating what LMI actually did
Why the Court Said No
Section 41 requires research to clear four separate tests. The court concluded that Leon Max, Inc. failed to satisfy several of them.
- No technical uncertainty: The court found that the company’s designers and patternmakers generally relied on existing knowledge, experience, and routine methods rather than attempting to resolve technical uncertainty within the meaning of Section 41.
- Not investigative in nature: The court leaned on the dictionary definition of “investigate,” close examination and systematic inquiry, and found that cutting fabric a certain way to see how it works, or trying a different thread size until one fits, doesn’t clear that bar. As the opinion put it, these were common solutions to common problems.
- Not fundamentally based on the physical or biological sciences, engineering, or computer science: The court rejected the idea that fit testing was engineering, draping was material science, or shrink testing was chemistry. The judge compared it to a baseball player catching a fly ball: physics is happening in the background, but that doesn’t make the outfielder a physicist.
- Primarily driven by style and taste: Much of the design work existed to make clothes look good and match seasonal trends, which the statute rules out on purpose.
What the Company Could Have Done Differently
- Commission a company-specific credit study rather than relying on an industry-wide approach: A claim built on interviews with your actual designers, patternmakers, and engineers, tied to specific technical problems they faced, holds up very differently than a sampled study applied broadly across a portfolio.
- Separate the genuinely technical work from the aesthetic work: LMI’s pintuck template, built with perforations and fluorescent powder guides, arguably involved more systematic problem-solving than the broader claim suggested. Isolating activities like this, with real cost tracking, is far stronger than bundling style-driven design in with functional testing.
- Vet your expert before they take the stand An expert who admits he relied entirely on someone else’s report, without independently checking it, hands the other side its strongest argument for free.
Where TaxDrone.AI and NTG Actually Fix This
Leon Max, Inc. wasn’t a small operation cutting corners. It ran a sophisticated, well-documented process. That’s exactly why this case matters. The claim didn’t fail because the company lacked process. It failed because the credit study was generic, industry-wide, and never independently checked against what LMI’s own people actually did.
Here’s how TaxDrone.AI and NTG are built to avoid that specific failure:
- Projects are documented at the business component level, helping organize activities around the Section 41 requirements instead of relying on industry-wide assumptions.
- Supporting documentation is tied back to your own projects, employees, and records, helping demonstrate what your business actually did rather than what companies in your industry generally do.
- As IRS documentation expectations continue to expand, including more detailed project level reporting, contemporaneous records have become increasingly important for supporting R&D tax credit claims.
Your R&D Tax Credit Should Tell Your Story
