The federal research and development (R&D) tax credit under Internal Revenue Code Section 41 rewards qualifying research expenditures, but technical complexity alone does not establish eligibility. Phoenix Design Group, Inc. v. Commissioner, T.C. Memo. 2024-113, illustrates the importance of proving technical uncertainty and a qualifying process of experimentation for the business components claimed.

This study examines the engineering activities, sampling disputes, documentation issues, shrinking-back analysis, and funded-research questions raised by the source material. It also distinguishes the unrelated Balint decision and IRS theft-loss advice from research-credit authorities. These authorities do not establish a universal prohibition on retrospective evidence or a new requirement to use a particular project-management system.

Deconstructing Phoenix Design Group, Inc. v. Commissioner

Phoenix Design Group (PDG) designs mechanical, electrical, plumbing, and fire protection (MEPF) systems. Its disputed credits arose from activities during 2013 through 2016. The credit study considered 409 projects, reduced the potentially qualifying population to 238, and examined a sample of 24. The consultant treated 20 sampled projects as qualifying and extrapolated the results. The dispute concerned whether the underlying activities met Section 41, rather than whether engineering work was professionally sophisticated.

Procedural History and the Scope of Discovery

PDG sought to restrict discovery and trial to three to six example projects. In its August 29, 2023 order, the Tax Court declined that request. The taxpayer had not demonstrated that those examples represented the larger population or would produce statistically valid results. The IRS could seek relevant information beyond the taxpayer’s proposed examples.

A sampling method used to prepare a credit study does not itself bind the IRS or the court. An agreement should specify project selection, the issues being tried, and the effect of sample findings on other projects. Sampling may remain useful, but its statistical validity and its procedural effect are separate questions.

The Sampled Projects and the Threshold of Qualification

The December 2024 opinion examined three projects: Gerald Champion Military Psychiatric Unit, Baptist Memorial Health North Mississippi Oxford, and Vanderbilt University Engineering and Science Building. The parties’ arrangement used the findings as a framework for other projects; it did not make the three-project sample automatically binding on the remaining population.

PDG did not establish qualified research for the three projects examined. The opinion also sustained accuracy-related penalties under the parties’ stipulation. It should not be described as a separate factual adjudication of every project in the original population, or as holding that any unsuccessful credit claim automatically attracts a penalty.

The Section 174 Test: Technical Uncertainty and Design

For the years considered in Phoenix Design Group, the credit incorporated the research-and-experimental expenditure standard under Section 174. The relevant question was whether available information established the capability, method, or appropriate design of the product or component. An unfinished design is not, by itself, proof of the required uncertainty.

Investigative Activity and Calculation

PDG relied on engineering calculations and design decisions to explain how it addressed project questions. A completed calculation or drawing identifies an outcome, but it may not reveal what information was missing or how the engineers investigated alternatives. The evidentiary gap matters when determining whether the work was research rather than the application of known inputs to a familiar solution.

Existing engineering principles can support qualifying research. The use of established science, commercially available tools, or standard calculations does not categorically disqualify a project. The claim must establish how those tools were used in investigative work directed at an actual capability, method, or design uncertainty.

Business Uncertainty and Technical Uncertainty

Waiting for a customer’s equipment choice, an architect’s footprint, or a final budget can delay a project without creating a qualifying technical uncertainty. Conversely, an unresolved design problem arising from performance constraints may warrant investigation. The analysis should identify which kind of uncertainty existed and who resolved it.

A problem affecting one subsystem does not automatically establish uncertainty for an entire building-system design. Component boundaries and the activities attributed to each component require support. The source’s association of hybrid operating rooms with the Vanderbilt laboratory project should not be relied on without project-specific evidence.

The Process of Experimentation and Standard Design Workflows

Section 41 requires more than a description of ordinary design phases. The taxpayer must identify an evaluative process directed at resolving technical uncertainty, and substantiate the activities included in that process.

Standard Workflows Do Not Establish Experimentation by Themselves

PDG’s reliance on American Institute of Architects design phases did not establish that its employee activities constituted a qualifying process of experimentation. Labels such as schematic design, design development, and construction documents describe how work is organized. They do not explain the alternatives evaluated or how a particular technical problem was investigated.

Modeling, simulation, and systematic trial and error can demonstrate experimentation. A formal laboratory protocol, a failed prototype, or a documented sequence of repeated failures is not invariably required. Successful research can qualify, while revisions made solely to accommodate customer preferences may not.

Activities and the Substantially-All Requirement

Treasury Regulation Section 1.41-4(a)(6) sets an 80% threshold for research activities constituting elements of a process of experimentation, measured by cost or another consistently applied reasonable basis. The test applies separately to each business component. It should not be simplified into an inflexible requirement that 80% of every employee’s total project hours be experimental.

Little Sandy Coal Co. v. Commissioner illustrates why the underlying activities and their measurement matter. An assertion that a product is largely new does not supply an activity-based calculation. The component-level experimentation test must also be distinguished from the separate wage rule for an employee whose services are substantially all qualified services.

Balint, Theft-Loss Guidance, and the Limits of the Analogy

T.C. Memo. 2023-118 is Balint v. Commissioner, not Windus v. Commissioner. Balint concerned income inclusion from disputed retirement-account and life-insurance withdrawals made by a spouse acting under a power of attorney. It is not an R&D credit decision and does not establish a Section 41 substantiation standard.

Basis, Income Inclusion, and Research Expenses

The source also conflated Balint with IRS Chief Counsel Advice 202511015, which discusses theft losses involving scams. That advice addresses Section 165, including loss basis, profit motive, and recovery prospects. It is not a judicial holding and may not be used or cited as precedent.

A rule limiting a theft deduction for fictitious investment earnings does not create a rule requiring research expenditure amounts to have first been included in the claimant’s gross income. Section 41 instead addresses qualifying expenses paid or incurred, their statutory categories, and their relationship to qualified research. Payroll, accounting records, invoices, and supported allocations should establish the amounts claimed.

Profit Motive and the Trade or Business Requirement

The IRS scam advice distinguishes transactions entered into for profit from personal transactions under the law applicable to the situations it analyzes. Those theft-loss distinctions cannot be imported directly into the research credit.

Section 41 has its own trade-or-business and business-component requirements, including a limited rule for certain startup in-house research expenses. A commercial objective alone does not establish qualification, just as an activity’s personal character cannot be cured by labeling it research. The substantive credit requirements must be applied to the actual work.

Fiduciary Misappropriation and Local Law

In Balint, the court considered Florida law and the agent’s authority when determining whether disputed withdrawals were attributable to the taxpayer. Its holding concerned the particular withdrawals and economic benefit, not a blanket exemption for all spousal withdrawals.

The limited parallel to funded-research disputes is that legal rights may depend on applicable nonfederal law. Contract terms and governing law can affect payment obligations and ownership rights. Balint itself does not determine those questions for an R&D credit claim.

Documentation and the Weight of Evidence

Research-credit documentation should explain both the technical activities and the expenses associated with them. A broad project description can provide context, but it is less useful when it leaves the actual investigation, personnel roles, or cost allocation unexplained.

Retrospective Narratives Require Support

The source’s statement that undocumented real-time activity does not exist for tax purposes is too absolute. Treasury Regulation Section 1.41-4(d) requires usable records with enough detail to substantiate eligibility; it does not prescribe a universal contemporaneous task-log format. Later interviews and reconstructions require a credible evidentiary foundation and should be checked against available project and financial records.

The practical concern raised by disputes over interview-based studies, including the Kyocera discussion cited in the source, is the reliability of the resulting evidence. Such disputes should not be presented as establishing that all retrospective interviews or estimates are legally inadmissible. Preserve original evidence and clearly distinguish recorded facts from later recollections.

Mapping Activities to Section 41

A useful project file identifies the technical question, the alternatives considered, how they were evaluated, and the connection to claimed costs. The evidence may include design versions, test results, calculations, technical correspondence, and explanations from the personnel who performed the work.

For wages, distinguish employees conducting qualified research from those directly supervising or directly supporting it. Job titles and general administrative involvement are insufficient substitutes for evidence of the services performed. Reconcile the calculation to payroll and document the basis for allocations.

The following comparison preserves the source’s table headings. Its final column describes recommended substantiation practices, not a new legally mandatory system created by Phoenix Design Group.

The Shrinking-Back Rule and Its Limits

When an entire business component does not meet the qualification requirements, the shrinking-back rule permits examination of its most significant subset of elements, continuing to smaller subsets as appropriate. This is a method for applying the eligibility test, not a guarantee that some portion of every project qualifies.

The source overstates the distinction between complete projects and employee tasks. The relevant focus is the supported component or subset and the research activities associated with it. Renaming a design phase or selecting isolated employee hours does not demonstrate that a qualifying subset exists.

PDG’s evidence did not establish qualifying research for the component or subcomponent work considered. As a practical matter, firms should retain information that permits meaningful analysis at more than one level, such as an integrated system and its relevant subsystems. A narrower claim still needs evidence of qualification and related expenses.

Funded Research and Financial Risk

Section 41(d)(4)(H) excludes research to the extent funded by another person. Treasury Regulation Section 1.41-4A(d), incorporated into the applicable research-credit rules, examines payment contingency and substantial rights. Research performed without retaining substantial rights is fully funded. Where substantial rights are retained, noncontingent payments can fund research to the extent of those payments; qualifying excess expenses require a separate analysis.

Lessons from Meyer, Borgman & Johnson

In Meyer, Borgman & Johnson, Inc. v. Commissioner, the Eighth Circuit upheld the funded-research determination. The problem was not simply that the agreements used the term design services. General duties to provide competent, code-compliant work did not establish payment contingent on successful research.

Review payment, acceptance, rejection, termination, refund, and intellectual-property terms together. A fixed fee or ordinary exposure to cost overruns does not automatically establish the research risk required by the regulations.

Smith and System Technologies: Procedural and Substantive Outcomes

The earlier Smith summary-judgment ruling did not finally establish credit entitlement. The later opinion, Smith v. Commissioner, T.C. Memo. 2026-50, found that payments under six contracts were not contingent on research success. Substantial rights were retained under four contracts, permitting potential partial credits under the funded-research regulations; the precise amounts remained unresolved in that opinion.

In System Technologies, Inc. v. Commissioner, docket No. 12211-21, the January 3, 2025 order denied the IRS’s motion for partial summary judgment. Indiana law required refunds for nondelivery, and the warranty terms did not eliminate that remedy. The court treated ultimate payment as contingent on successful research. This resolved the funding argument presented, rather than every requirement for the claimed credit.

Implications for Future R&D Tax Credit Applications

The practical lessons are to establish the qualifying activities, support the calculation, and evaluate contracts on their own terms. The decisions do not create a single new standard that replaces Section 41 and its regulations.

Testing Technical Assumptions

Engineering, architecture, and software work should be assessed by the activities performed. Complexity, novelty, professional skill, and a successful deliverable do not independently prove qualification. Explain the actual investigation and separate it from routine implementation, customer preference changes, and other excluded activities.

Integrating Tax and Project Management

Capturing useful evidence during the project can reduce dependence on memory. Project-management software may help, but no particular platform is required. Choose a process that technical staff can maintain, preserves original work product, and allows the finance team to reconcile costs to the claim.

Strategic Use of Sampling

Define the project population and document selection and extrapolation methods. A small group of convenient examples may not represent the whole. Where sampling will be used in an examination or litigation, seek an agreement that explains its scope and consequences. Maintain the ability to provide relevant information about projects outside the sample.

Legislative and Regulatory Developments

Public Law 119-21, commonly called the One Big Beautiful Bill Act, was enacted July 4, 2025. It added Section 174A, generally permitting current deductions for domestic research or experimental expenditures paid or incurred in taxable years beginning after December 31, 2024. Foreign research expenditures remain subject to 15-year amortization under Section 174. Transition rules and elections require separate attention, as explained in IRS Revenue Procedure 2025-28.

The deduction changes do not automatically make expenditures eligible for the Section 41 credit. Apply the law for the tax year at issue and coordinate the deduction and credit rules. The source’s assertion that the IRS is systematically targeting smaller technical firms because of weak documentation is not established by the cited decisions and should not be treated as a documented enforcement policy.