In July 2024, the United States moved to disallow Kyocera AVX Components Corporation’s entire Section 41 research credit claim. The company retained a third-party advisor to prepare a supporting study. It cooperated with discovery and had experienced counsel.
None of it was enough because the underlying records did not exist.
For CFOs and tax directors managing large credit claims, this dispute is not a legal curiosity. It aligns with increased scrutiny around documentation and methodology seen in recent enforcement and guidance. More importantly, it illustrates that this was not only a documentation issue. The claim methodology itself struggled to demonstrate eligibility.
Kyocera originally claimed a Section 41 credit of $398,985 for its tax year ending March 31, 2018. Two years later, it filed an amended return claiming $1,709,147 based on a study commissioned about sixteen months after the tax year closed.
The study relied on interviews with 36 subject matter experts who estimated research time for more than 1,200 employees across five business components: Tantalum Components, Ceramic Components, Advanced Specialty, Antenna, and Interconnect. Those estimates were then used to allocate wages, supply costs, and contract research expenses.
Kyocera did not maintain centralized time tracking for these projects. During discovery, it produced no time records, supply invoices, or contract documentation supporting the allocations. The advisor produced workpapers summarizing survey data, but the underlying survey responses and interview materials were not retained.
The government’s position was direct. Estimates are not records, and a study built primarily on estimates does not substantiate a Section 41 credit claim.
The motion focused on two core issues that relate to claim design rather than record volume.
Process of experimentation applied at the wrong level
Section 41 requires showing that substantially all research activities meet the process of experimentation test for each business component. Estimated time allocations without activity-level evidence have repeatedly been viewed as insufficient.
In this case:
Expense substantiation could not be demonstrated
Treasury regulations require records supporting both the amount and nature of claimed expenses. The claim included multiple roles without activity-level evidence showing participation in experimentation.
The record showed:
The company tracked employee time for certain government and defense projects but not for the Section 41 claim. This suggests the issue relates to claim discipline rather than tooling alone.
The taxpayer anticipated reliance on the Cohan doctrine, which allows courts to estimate amounts after eligibility is established. The sequencing matters.
Estimation is generally available only once qualification is demonstrated. Where the process of experimentation test cannot be shown, there may be no eligible amount to estimate. The dispute therefore centers on eligibility rather than precision.
Cohan does not function as a safety net for claims whose methodology has not established qualification.
The Kyocera dispute reflects themes seen across recent R&D credit cases. The reasoning aligns with issues courts have examined repeatedly, particularly around reconstruction, allocation logic, and component-level substantiation.
At a structural level, disputes like Kyocera reflect a broader industry tension rather than an isolated documentation lapse. Large R&D credit claims are often reconstructed because operational and accounting systems were not designed to capture Section 41 evidence as work occurs. Interview studies emerge as a practical response to scale, amended claim economics, and the difficulty of tracking experimentation activity across large organizations.
The result is a persistent tradeoff between claim size, feasibility, and evidentiary strength. That tension increasingly sits at the center of credit defensibility discussions.
Key boundaries highlighted by the case:
Cases like Kyocera often expose two distinct breakdowns: how the claim is designed and how supporting evidence is preserved.
National Tax Group’s role sits at the claim design level. This includes defining business components at the appropriate level, applying the substantially-all test consistently with case law, evaluating which roles can be supported as qualified services, and limiting reliance on estimation where activity evidence is unavailable.
TaxDrone.AI addresses the evidentiary layer courts frequently scrutinize. Rather than reconstructing allocations after the fact, the platform supports contemporaneous capture of activity tied to business components, preserves the inputs used to derive time percentages, and maintains links between employees, experimentation activities, and claimed costs.
Together, this approach is intended to reduce reconstruction risk, particularly where allocation logic exists but the underlying evidentiary trail does not.
Recent disputes indicate that Section 41 analysis continues to focus on whether experimentation can be demonstrated through evidence rather than reconstructed through memory. In Kyocera, the claim relied on interviews conducted well after the work concluded, covering employees beyond the interviewees’ direct knowledge and without preserved supporting records.
The lesson is methodological. Risk can arise at the filing stage when eligibility depends on reconstruction.
If a Section 41 claim relies heavily on retrospective allocation or limited activity evidence, it may be useful to evaluate how well the underlying documentation supports the eligibility narrative before filing.
