The research and development (R&D) tax credit under Internal Revenue Code (IRC) § 41 requires taxpayers to establish both qualifying activities and eligible expenses. Phoenix Design Group, Inc. v. Commissioner, T.C. Memo. 2024-113, illustrates the evidentiary challenges faced by engineering, architectural, and design firms. Professional complexity does not itself establish qualified research. The decision applies the statutory four-part test, particularly technical uncertainty and a process of experimentation, to the evidence for three trial projects.

The R&D tax credit incentivizes qualifying domestic research by offsetting a portion of eligible costs. Technical job titles, project descriptions, and payroll totals do not alone establish entitlement. Phoenix Design Group (PDG) illustrates the need to connect specific work to the statutory requirements. It does not establish that engineering services are categorically ineligible or that courts have replaced a formerly permissive legal standard with a new documentation rule.

The Statutory Framework: The Four-Part Test as a Legal Threshold

Research must satisfy each part of the four-part test at the business-component level. If an entire component fails, qualifying subsets may still be considered under the shrinking-back rule. Separate expense requirements and statutory exclusions also apply. The following table describes the framework relevant to the historical years in Phoenix Design Group.

The Section 174 test applicable to the years at issue concerned research or experimental expenditures incurred to eliminate uncertainty about capability, method, or appropriate design. PDG illustrates that calculations using information already available do not necessarily establish investigative activity. The technological-information requirement concerns reliance on physical or biological sciences, engineering, or computer science. The research must be intended to improve a business component, and its experimental process must evaluate one or more alternatives. Neither industry-wide novelty nor a successful outcome is required. For tax years beginning after 2024, the statutory expense cross-reference in § 41(d)(1)(A) is to Section 174A; the historical Section 174 discussion should not be read as a complete statement of current deduction rules.

Case Analysis: Phoenix Design Group, Inc. v. Commissioner (T.C. Memo 2024-113)

Phoenix Design Group is a multidisciplinary engineering firm focused on mechanical, electrical, plumbing, and fire protection (MEPF) systems. Its credit study considered more than 200 potentially qualifying projects. The disputed credits arose from 2013–2016 activities and were used in the 2015–2019 tax years. The Tax Court issued its opinion on December 23, 2024, finding that qualified research had not been established for the three projects tried.

The trial projects were the Gerald Champion Military Psychiatric Unit, Baptist Memorial Health North Mississippi Oxford, and Vanderbilt University Engineering and Science Building. PDG argued that their design constraints and performance requirements involved qualified research. The parties agreed that the findings on those projects would not bind the remaining projects, although they would provide a framework for resolving the wider dispute.

Failure to Prove Objective Technical Uncertainty

A central weakness was PDG’s failure to establish qualifying technical uncertainty for the claimed activities. Uncertainty concerns whether the information available establishes capability, method, or appropriate design. The engineers’ professional knowledge and historical data were relevant to what was already known. The court distinguished investigation to resolve uncertainty from basic calculations using information already available.

Professional complexity is not a substitute for proof of technical uncertainty. The possibility that a design might change, that building details were not final, or that equipment had not yet been selected did not by itself establish uncertainty throughout the MEPF systems. PDG needed to explain the specific unresolved technical questions and the investigative work performed to address them.

The Dissection of the Six-Stage Design Process

PDG relied on a six-stage process: Basis of Design, Schematic Design, Design Development, Construction Documents, Bidding, and Construction Administration. A sequence of design and project-management stages does not automatically establish an evaluative process. The court examined the work actually performed, including discrepancies between general descriptions of those stages and the activity records.

The relevant distinction is whether activities evaluated alternatives to resolve a technical uncertainty. Iteration and engineering calculations can be part of qualified research, but their existence alone is insufficient. The evidence must explain what was uncertain, which alternatives were evaluated, and how the evaluation addressed that uncertainty. These stage descriptions therefore should not be treated as categorical exclusions for every engineering firm.

The “Substantially All” Requirement and the 80% Rule

The substantially-all requirement in IRC § 41(d)(1)(C) generally requires 80% or more of the research activities for a business component to constitute elements of a process of experimentation for a qualified purpose. Treasury Regulation § 1.41-4(a)(6) permits measurement by cost or another consistently applied reasonable basis. This is not an 80% test of all company operations or simply of a product’s new features.

Experimental activity ratio = measure of research activities constituting elements of a process of experimentation for a qualified purpose ÷ measure of total research activities for that business component ≥ 80%, using cost or another consistently applied reasonable basis.

In Little Sandy Coal Co. v. Commissioner, 62 F.4th 287 (7th Cir. 2023), the Seventh Circuit affirmed the denial of the credit while rejecting aspects of the Tax Court’s reasoning. Direct supervision, direct support, and pilot-model production activities are not categorically excluded from the experimentation analysis. Their inclusion depends on the work performed and its relationship to experimentation. The taxpayer nevertheless failed to provide a supported allocation of activities by vessel. The decision does not prescribe a universal ratio of engineers’ wages to all project expenses.

PDG’s records did not adequately connect the claimed activities to qualified research. The court considered shrinking back to narrower components and technical issues, but PDG did not substantiate a qualifying subset. Shrinking back concerns subsets of a business component, not merely a design phase or a selected employee’s work. Suitable component-level records can support that analysis even when the overall system fails.

Evidentiary Standards and the Burden of Substantiation

Contemporaneous technical and financial records can provide persuasive evidence of what occurred, but the law does not impose a blanket prohibition on interviews, later reconstructions, or oral testimony. Treasury Regulation § 1.41-4(d) requires records sufficiently detailed to substantiate eligibility and amount. The relevant question is whether the combined evidence reliably establishes qualified research and the claimed expenses.

The Role of Oral Testimony and the Cohan Rule

The Cohan estimation principle does not dispense with proof that qualified research occurred. Little Sandy Coal expressly recognized that, once qualification is established under § 41(d), eligible expenses under § 41(b) may be estimated where the evidence supports doing so. Arbitrary percentages cannot substitute for proving the substantially-all requirement. Failure at the whole-component level still requires consideration of any substantiated qualifying subset under the shrinking-back rule.

In PDG, inconsistencies between testimony and activity records weakened the proof. Demonstrating an eventual design solution does not establish the investigative or experimental steps used to reach it. Technical explanations should be reconciled with drawings, calculations, revisions, test records, and time or expense records where those materials exist.

Financial Consequences and Accuracy-Related Penalties

The accuracy-related penalties in Phoenix Design Group must be understood in light of the parties’ stipulation: the penalty outcome depended on whether the court found qualified research in any of the three trial projects. The opinion’s adverse findings therefore triggered the stipulated penalty result. This does not establish that every disallowed research credit automatically produces a 20% penalty under IRC § 6662.

The table shows the credit amounts used and the penalties identified for the disputed tax years. These are distinct from the years in which research generated the credits. Section 6662 generally measures the relevant penalty against an underpayment, not mechanically against every dollar of claimed credit. Section 6676 concerns certain excessive refund or credit claims, includes a reasonable-cause exception, and has coordination rules. Administrative review procedures do not make penalties mandatory whenever an R&D claim is disallowed.

The Impact of Funded Research and Contractual Rights

For client-service firms, the funded-research exclusion under IRC § 41(d)(4)(H) requires examination of payment risk and rights in the research results. Amounts payable regardless of research success may constitute funding; research for which the taxpayer retains no substantial rights is also excluded. The contract, incorporated documents, governing law, and actual arrangements matter. A fixed price, milestone schedule, or warranty alone does not resolve every case.

Milestone-Based Payments and Technical Risk

In Smith v. Commissioner, a December 18, 2024 order denied the IRS’s motion for partial summary judgment on funded research. Contractual obligations, payment conditions, and governing law required further consideration; the order did not establish that all milestone-paid architectural work qualifies. In System Technologies, Inc. v. Commissioner, an order dated January 3, 2025 likewise denied the IRS’s partial-summary-judgment motion. The court considered Indiana law and concluded that remedies for failed performance, including repayment, were not displaced by the contractual warranties. These procedural rulings addressed funding arguments rather than deciding all requirements for the credits.

Service providers should examine whether they bear the relevant financial risk and retain substantial rights in research results. Copyright ownership can be relevant, but it is not the sole way to retain substantial rights. Contract labels and payment milestones must be considered alongside enforceable obligations and the actual work; drafting alone cannot turn nonqualifying activities into qualified research.

Future Implications: Revisions to Form 6765 and Increased Reporting

The revised Form 6765 collects additional information about research credits. These filing requirements do not replace the statutory qualification test, and the form’s business-component selection rule is separate from the 80% experimentation requirement. The applicable IRS instructions determine what information a filer must provide.

Significant Changes to Form 6765

Under the December 2025 instructions, Section G is optional for tax years beginning before 2026 and generally required afterward, subject to exceptions. When applicable, it covers components representing at least 80% of QREs, capped at 50 components, in descending QRE order; remaining components are aggregated. Exceptions include specified qualified small businesses claiming the payroll-tax credit and eligible original-return filers with controlled-group QREs of $1. million or less and prior-three-year average gross receipts of $50 million or less. Amended research-credit claims have separate requirements.

Strategic Implications for Future R&D Tax Credit Applications

Phoenix Design Group illustrates why a research-credit study should connect technical evidence to the activities and expenses claimed. Interviews can help identify relevant work, but should be supported by available business records. Capturing technical issues, alternatives, evaluations, and costs while work proceeds is a useful practice rather than a newly imposed universal requirement to use a particular real-time tracking system.

Identification of Specific Technical Challenges

Identify specific uncertainty about capability, method, or appropriate design and explain the work undertaken to resolve it. The requirement is not that the problem be beyond all existing scientific knowledge. Routine code checks, aesthetic choices, and standard measurements do not independently establish qualified research; any claimed activity must be evaluated under the statutory requirements and exclusions. Explain the uncertainty and experimentation in the taxpayer’s own work rather than relying solely on similarities to other cases.

Activity-Level Mapping

Connect employee services to the relevant research activities and business components. Section 41(b)(2)(B) covers actual conduct, direct supervision, and direct support of qualified research. Direct supervision genera