The Statutory Architecture of the Research and Development Tax Credit

Qualified research expenses, or QREs, include qualifying employee wages, supplies, certain computer-use costs, and eligible contract research payments, subject to statutory limitations. Qualification is assessed separately for each business component. Passing the research tests does not automatically make every associated expenditure creditable.

The Four-Part Test Framework

The following table summarizes the requirements. For taxable years beginning after December 31, 2024, Section 41(d)(1)(A) refers to domestic research or experimental expenditures under Section 174A. The earlier cases discussed below applied the Section 174 provisions governing their respective tax years.

The substantially-all threshold is at least 80%, measured on a cost or other consistently applied reasonable basis. The credit does not require an advance in industry-wide knowledge. Research exclusions, including routine quality control and certain adaptations of existing components, must nevertheless be considered.

The Wolf Precedent and the Burden of Substantiation

Wolf upheld a negligence addition under former Section 6653, rather than deciding a Section 6662 R&D credit penalty. Its discussion of reasonable care is relevant to general tax compliance, but it did not create a special research-credit documentation standard.

Negligence and the Duty of Care

Wolf considered whether the taxpayer acted prudently in investigating an investment and claiming associated tax benefits. It should not be characterized as a ruling against retrospective R&D interviews or wage allocations. For research credits, Treasury Regulation Section 1.41-4(d) directly requires records sufficiently usable and detailed to substantiate eligibility. Unsupported percentages are vulnerable; a reasoned reconstruction supported by reliable evidence is a different matter.

Implications of the Wolf Standard for Modern Claims

Section 6662 can impose a 20% accuracy-related penalty on an underpayment attributable to negligence or another specified ground. Disallowance of a research credit does not automatically establish a penalty. Reasonable cause and good faith under Section 6664(c), including the circumstances of any reliance on professional advice, require a separate analysis.

In Phoenix Design Group, T.C. Memo. 2024-113, the penalty outcome followed the parties’ stipulation linking penalties to the credit determination. It should not be described as an independent holding that missing contemporaneous time records necessarily establish negligence.

Investigative Activity and the Research Uncertainty Test

Under the framework applied in the cited cases, uncertainty exists when the information available to the taxpayer does not establish the capability or method for developing or improving a product, or its appropriate design. A claim must identify that uncertainty and the work undertaken to resolve it.

Objective vs. Subjective Uncertainty

Phoenix Design Group concerned mechanical, electrical, plumbing, and fire-protection engineering. The court rejected the claimed research for the trial projects because the evidence did not establish qualifying uncertainty and experimentation. The possibility of revising a design is insufficient by itself.

Complexity, demanding customer specifications, and technical difficulty do not automatically establish research uncertainty. Conversely, using established engineering principles does not itself disqualify research. The relevant question is whether the taxpayer systematically evaluated alternatives to resolve an uncertainty about capability, method, or design.

The Role of Investigative Activity

Project correspondence, test studies, calculations, design revisions, and meeting minutes can help establish what information was missing and how it was obtained. Their evidentiary value depends on their content and connection to the claimed activities. Prior experience with a similar system may narrow the uncertainty, but does not automatically eliminate uncertainty about a new component or its integration.

Dissecting the Process of Experimentation

A process of experimentation involves identifying uncertainty, identifying alternatives, and evaluating those alternatives through methods such as modeling, simulation, or systematic trial and error. Merely describing a design as innovative or listing tests performed does not demonstrate the required process.

The Scientific Method vs. Linear Design

In Siemer Milling Co. v. Commissioner, T.C. Memo. 2019-37, the taxpayer failed to establish the necessary process of experimentation for the projects at issue. The lesson is to explain how testing evaluates alternatives, rather than treating the existence of tests as conclusive.

A formal laboratory hypothesis-and-retesting sequence is not the only acceptable method. Nor is a project automatically disqualified because its management stages are sequential. The evidence must show what experimental work actually occurred. Routine design changes and compliance checks should be distinguished from systematic technical investigation; the adaptation exclusion also requires a fact-specific analysis.

Comparison of Engineering and Experimental Workflows

The examples below describe potentially qualifying experimental work and ordinarily nonqualifying routine work. Each activity remains subject to the complete statutory analysis; no table entry guarantees a credit.

The Substantially All Rule and the Fractional Test

The 80% requirement concerns activities, measured on an appropriate consistent basis. It is not a test of the percentage of a product that is new, and should not automatically be expressed as 80% of every construction hour.

The Little Sandy Coal Formula

Little Sandy Coal Co. v. Commissioner, 62 F.4th 287 (7th Cir. 2023), affirmed denial of credits for shipbuilding projects. The court rejected novelty and pilot-model status as substitutes for evidence of qualifying activities. A useful expression is: research activities constituting elements of experimentation divided by the relevant research activities for the business component, measured consistently, must reach at least 80%.

Inclusion of Supervision and Support

The Seventh Circuit rejected the Tax Court’s categorical exclusion of direct support and supervision from the numerator. Such activities can count when they constitute elements of the experimental process, with corresponding treatment in the denominator. The appellate court nevertheless affirmed the judgment because the taxpayer lacked a supported allocation demonstrating qualification. This was not a partial reversal.

Support and supervision are therefore neither automatically excluded nor automatically included. The taxpayer must establish their relationship to experimentation and a defensible allocation.

The Shrinking-Back Rule: A Strategy for Component-Level Qualification

If a business component fails the qualification tests, the shrinking-back rule permits examination of its most significant subset of elements, followed by progressively smaller subsets where appropriate. It does not excuse a failure to identify and substantiate qualifying research.

Mechanism of the Shrinking-Back Rule

The analysis continues until a subset satisfies the tests or the most basic element fails them. It applies to identifiable components or subcomponents, not simply to a selected group of employees or a favorable time percentage.

The Catalytic Products International dispute was Betz v. Commissioner, T.C. Memo. 2023-84. The court examined claimed research on air-pollution-control systems and considered particular subcomponents. The taxpayers failed to establish the claimed qualification and expenses; calling completed systems pilot models did not cure those evidentiary failures.

Application Challenges

These maritime examples are illustrative rather than findings about particular projects.

Preparing the component analysis before filing is prudent. However, shrinking back is not legally unavailable merely because it is first raised during an examination or litigation. Its usefulness depends on the supporting evidence and applicable procedural rules.

Contractual Risks and the Funded Research Exclusion

Section 41(d)(4)(H) excludes research to the extent funded by another person. Contract terms and the surrounding facts determine whether the researcher bears the relevant financial risk and retains substantial rights in the results.

Economic Risk and Milestone Payments

Payment genuinely contingent on successful research can support the researcher’s position, but milestone billing alone does not establish that contingency. Fixed-price labels and exposure to cost overruns do not resolve the issue. Payment obligations, acceptance conditions, termination provisions, and rights in results must be analyzed together. Phoenix Design Group should not be presented as establishing a general milestone-payment safe harbor.

Retention of Substantial Rights

The researcher generally must retain substantial rights to use the research results. Exclusive ownership is unnecessary, while increased experience alone is an incidental benefit. In Betz, contractual restrictions meant CPI lacked substantial rights on five projects. The rights analysis was an additional reason those projects could not support the claimed credit.

Modern Documentation Standards: FAA 20214101F and Subsequent Updates

The IRS’s 2021 Chief Counsel advice, commonly identified as FAA 20214101F, addressed information needed for research-credit refund claims. It is administrative legal advice, not a new statute or binding judicial precedent. The original five-item submission policy was subsequently modified.

Filing Requirements for R&D Refund Claims

Effective June 18, 2024, the IRS waived submission of the names of individuals performing each research activity and the information each individual sought to discover. Its published refund-claim guidance continues to require:

• Identification of the business components covered by the claim for the year.
• Identification of the research activities performed for each component.
• Total qualified wage, supply, and contract research expenses for the claim year, which may be provided using Form 6765.

The waived information may still be requested during examination. Filing requirements and the evidence needed to substantiate a credit are distinct. Taxpayers should also follow the applicable Form 6765 instructions; the 2021 memorandum should not be treated as an unchanged universal rule for all research-credit filings.

Comparison of Old vs. New Documentation Practices

The table retains the original comparison headings, but the second heading is historical shorthand: the entries distinguish current submission requirements from recommended substantiation practices. The memorandum did not universally mandate contemporaneous timesheets.

Industry Focus: Shipbuilding and Maritime Construction

Shipbuilding illustrates the challenge of separating development work from production. A first-of-its-kind vessel can include both qualifying research and substantial routine fabrication. The industry label and project value do not determine eligibility.

The Prototype vs. Production Dilemma

Pilot-model treatment under the research expenditure rules does not automatically establish the Section 41 experimentation test. Construction work should be evaluated according to its purpose and role in resolving uncertainty. Records should distinguish experimental support from production activities after the relevant design questions have been resolved.

Financial and Disclosure Impacts

A disputed credit can affect tax expense, interest, cash flow, and financial-statement disclosures. However, an individual company’s settlement does not establish why a claim failed or prove an industry-wide trend. The original discussion of Conrad Industries and Gulf Island Fabrication does not substantiate such a conclusion. Financial exposure should be assessed from the actual disputed positions and settlement terms.

Future Implications for R&D Tax Credit Applications

The practical lesson is to connect technical evidence, component-level qualification, expense allocations, and contractual rights. General negligence decisions should not replace the research-credit statute, regulations, and directly relevant case law.

Impact of TCJA Amortization and Section 174A

For tax years beginning in 2022 through 2024, the TCJA generally required capitalization and amortization over five years for domestic research expenditures and fifteen years for foreign research expenditures. Legislation enacted in July 2025 added Section 174A, generally permitting immediate deduction of domestic research or experimental expenditures for tax years beginning after December 31, 2024