The Regulatory Foundation and the Four-Part Test

Section 41 provides a credit calculated by reference to qualifying research expenses, subject to statutory limitations and computational rules. Scientific or engineering work does not automatically qualify. The taxpayer must establish both qualifying activities and eligible expenses.

For the tax year at issue, the first part of the qualified research definition referred to expenditures eligible for treatment as expenses under Section 174. The research had to address uncertainty in the experimental or laboratory sense in connection with the taxpayer’s trade or business. This historical rule should not be confused with subsequent changes to research expenditure deductions: following the 2025 legislation, Section 41(d)(1)(A) refers to domestic research or experimental expenditures under Section 174A, generally for tax years beginning after December 31, 2024.

The technological requirement calls for reliance on physical or biological sciences, engineering, or computer science. The information sought must be intended to help develop or improve a business component, such as a product, process, computer software, technique, formula, or invention held for sale, lease, or license, or used in the taxpayer’s business.

The process of experimentation requirement involves identifying technical uncertainty, identifying alternatives, and evaluating those alternatives through methods such as modeling, simulation, or systematic trial and error. Its purpose must concern function, performance, reliability, or quality. Mere novelty, commercial success, or technical complexity is insufficient.

Under Treasury Regulation Section 1.41-4(a)(6), at least 80% of the research activities for each business component, measured by cost or another consistently applied reasonable basis, must constitute elements of a process of experimentation for a qualified purpose. The denominator is the relevant research activities, rather than automatically every activity associated with the entire product or project.

Substantially all fraction: Research activities constituting elements of a process of experimentation for a qualified purpose ÷ total relevant research activities for the business component ≥ 80%, using the same permissible measurement basis for both.

The remaining research activities must meet the applicable expenditure requirement and avoid statutory exclusions. This business-component test is distinct from the separate substantially-all rule for an employee’s qualified services. Neither should be replaced with a blanket allocation of departmental wages.

The Corporate Context: Kyocera AVX and Electronics Manufacturing

Kyocera AVX, formerly AVX Corporation, manufactures electronic components. Kyocera completed its acquisition of the remaining AVX shares in 2020, making AVX wholly owned; the Kyocera AVX name followed in 2021. Its business includes products such as capacitors, connectors, and sensors.

Electronic-component development can involve technical uncertainty, including material selection, manufacturing tolerances, and performance under environmental stress. However, general descriptions of the company’s products cannot establish which activities in a particular tax year satisfy Section 41.

The government’s July 2024 motion identifies the disputed year as the tax year ending March 31, 2018. It states that Kyocera claimed a $398,985 research credit on its original return and increased that amount to $1,709,147 in a July 2021 amended return following a PwC study. The incremental credit was $1,310,162. The available filing does not support describing this specific refund suit as an adjudicated aggregate claim for 2017 through 2020.

Comparative Financial Claims and the Litigation

The research credit claim, the Section 965-related issue, the government’s erroneous-refund recovery action, and the deficiency dispute represent different claims or procedural positions. They should not be added together as though they were independent amounts finally owed by the company. The table distinguishes the historical amounts and procedural descriptions; it does not present a verified final disposition of every proceeding.

An amended claim can lead to examination of the facts necessary to establish an overpayment. However, the coexistence of these disputes does not prove that the research credit amendment caused the erroneous-refund action, or that the company faced a threat to its financial stability.

Analysis of the PwC Study and the Government’s Documentation Arguments

The government argued that the PwC study depended on retrospective interviews and estimates that did not adequately establish activities or expense amounts. Accounts of the study describe interviews with 36 subject matter experts covering approximately 1,200 employees. The important evidentiary question is the foundation for each estimate, rather than the interview count alone.

According to the government’s motion, Kyocera lacked sufficient project-level evidence for the claimed activities, and the study did not adequately describe employees’ roles and activities by business component. The government also disputed whether interviewees had enough firsthand knowledge to estimate time for employees across different levels of the organization.

The motion alleged that supporting study materials were not retained or produced, including records generated by PwC. It also challenged substantiation of expense amounts. These are the government’s contentions, not findings adopted merely because a motion was filed.

The government opposed using the Cohan estimation doctrine to fill those gaps. Its argument distinguished proving entitlement to a credit from estimating an amount after entitlement has been established. Courts may estimate expenses in appropriate circumstances, but an estimate cannot simply substitute for evidence that the research qualifies.

The practical lesson is to preserve the underlying evidence and explain how the study reaches its allocations. Neither an accounting firm’s reputation nor a polished narrative independently proves eligibility. Retrospective interviews may contribute useful evidence when grounded in knowledgeable testimony and corroborating records; this dispute does not establish that all retrospective studies are invalid.

The Procedural Shift: District Court and Tax Court

Kyocera brought its refund action in the District of South Carolina in 2022. Contemporaneous accounts published April 2, 2025 describe Judge Timothy M. Cain’s dismissal of that action after the company petitioned the Tax Court concerning its 2018 tax liability. The event should not be described as a March 2025 merits ruling rejecting the research credit study.

Jurisdictional rules can prevent overlapping determinations of the same tax liability in different courts. A dismissal on jurisdictional grounds does not decide whether the taxpayer’s research satisfies Section 41 or whether the government’s factual criticisms are correct.

The government’s erroneous-refund recovery suit, filed separately in March 2025, must also be distinguished from the taxpayer’s refund action. The available materials do not establish that every proceeding was consolidated in Tax Court or that all amounts were finally resolved there.

Precedential Context: Little Sandy Coal and Park-Ohio

In Little Sandy Coal Co. v. Commissioner, decided March 7, 2023, the Seventh Circuit affirmed denial of a shipbuilder’s research credit. The taxpayer had not supplied a principled basis for determining the portion of activities that constituted experimentation. General claims about the vessels’ newness and unsupported time allocations were insufficient.

The appellate court disagreed with aspects of the Tax Court’s reasoning while affirming the result. Its decision does not establish a universal requirement to use automated timekeeping or one prescribed record format. The relevant lesson is to support the experimentation analysis with evidence permitting a reasonable, consistently applied calculation.

Park-Ohio Holdings Corp. v. United States, filed April 14, 2025 in the Northern District of Ohio, concerns a different issue: the procedural validity of research-credit refund requirements. Park-Ohio alleged that the IRS’s additional information requirements violated the Administrative Procedure Act. An allegation in that lawsuit is not a ruling invalidating the requirements.

The IRS’s published guidance also requires an important qualification. Effective June 18, 2024, it waived submission of individual researchers’ names and the information each individual sought to discover at the refund-filing stage. It continued to require identification of business components, research activities for each component, and aggregate qualifying wage, supply, and contract research expenses. The waived information may still be requested during examination.

Technology and Research Documentation

Project management, engineering, and timekeeping systems can help preserve evidence created during the work. Useful records may include dated test results, design revisions, technical discussions, work assignments, and links between activities and accounting records. These are practical aids, not a new legal mandate established by Kyocera.

A timestamp proves when a record was created; it does not automatically establish hours worked, qualified research, or the eligibility of a particular cost. Companies should explain what each data source measures and how allocations are derived.

AI tools may assist with organizing records, locating relevant materials, or preparing draft narratives. They cannot reliably determine legal eligibility or employee time merely from metadata without validation. Technical personnel and tax reviewers should check the underlying evidence, the allocation method, and the resulting study.

The Technical Challenges of Ceramic Capacitor Research

Capacitor development provides useful illustrations of potential technical uncertainty. The following examples are illustrative, rather than judicial findings about the projects at issue in Kyocera’s claim:

• Material reliability: Evaluating alternative ceramic or electrode materials to address uncertain electrical performance under temperature or humidity stress.
• Manufacturing scalability: Testing alternative process settings to resolve uncertainty about achieving required tolerances or performance at production scale.
• Signal integrity: Comparing component designs to address uncertain high-frequency electrical behavior.

To support a claim, the taxpayer should connect the actual uncertainty, alternatives, evaluation activities, and expenses for each relevant business component. Routine quality-control testing or ordinary production does not qualify merely because the product is sophisticated. A successful outcome is not required, but qualifying activities and costs must be substantiated.

Broader Litigation Context: Data Breaches and Environmental Matters

Separate litigation should not be treated as evidence that Kyocera’s research credit claim was deficient. In the employee data-breach action commonly captioned Dollar v. Kyocera AVX Components Corporation, the March 6, 2025 order dismissed the case without prejudice for lack of Article III standing. The court did not decide the merits of the underlying negligence allegations.

The United States also brought a CERCLA action concerning environmental response costs at the Olean Well Field Superfund Site. That matter concerns environmental obligations and provides no determination of Section 41 eligibility. Claims that these unrelated cases establish a tax-documentation standard or prove financial distress are unsupported.

Practical Recommendations for R&D Tax Credit Compliance

The following practices address the evidentiary weaknesses raised in the dispute. They are practical recommendations grounded in substantiation requirements, rather than new rules imposed by a Kyocera merits decision.

Maintaining Contemporaneous Activity Records

Capture reliable evidence of the work as it occurs where practicable. Time records can help, but technical and financial records should also identify activities and connect them to the claim. Separate experimentation from routine maintenance, production, and other excluded activities. Use a consistent, supportable method to allocate eligible expenses.

Using Subject Matter Experts Effectively

Select interviewees with firsthand knowledge, record the basis for their estimates, and corroborate their accounts. Preserve questionnaires, interview notes, assumptions, and supporting documents. Do not assume that a manager knows the activities of every employee in a broad organizational group. The reliability of testimony is a separate question from whether an employee’s services qualify as direct research, direct supervision, or direct support.

Implementing Record Retention Policies

Preserve technical evidence and the records used to prepare the study, including payroll data, supply invoices, contract documents, and allocation calculations. Retention should account for applicable limitations periods, credit carryforwards, amended claims, examinations, and litigation holds. Clarify which records the taxpayer and adviser will retain and how they will remain accessible.

Defining Business Components and Applying Shrink-Back

Identify the actual product, process, software, technique, formula, or invention being developed or improved. Apply the qualification tests separately as required. Broad departmental labels may obscure the relevant activities, but the law does not invariably require selecting the smallest possible component at the outset.

If the requirements are not met for the overall business component, Treasury Regulation Section 1.41-4(b)(2) provides for applying the tests to the most significant subset of elements and continuing to smaller subsets as necessary. This shrink-back analysis requires evidence about the subset and its activities; it does not automatically rescue unsupported costs.

Final Thoughts

Kyocera’s research credit dispute highlights the risks of unsupported allocations and incomplete study records. Its procedural history also demonstrates the need to distinguish allegations, motions, jurisdictional decisions, and merits judgments.

A defensible research credit study connects qualifying technical activities to eligible expenses and preserves evidence supporting that connection. Contemporaneous records and well-founded interviews can work together. The reviewed materials do not justify declaring retrospective studies obsolete, automated tracking mandatory, or the company’s credit finally disallowed on the merits.