U.S. research tax incentives distinguish qualifying business research from personal expenditure and routine activity. Mayrath v. Commissioner, 41 T.C. 582 (1964), affirmed at 357 F.2d 209 (5th Cir. 1966), illustrates that distinction under the historical research expenditure deduction. It predates the research credit and should not be treated as the source of every modern Section 41 requirement. This study examines its holding alongside later legislation, regulations, and cases.

The Genesis of Section 174 and the Trade or Business Prerequisite

Congress enacted Section 174 in 1954 to ease the tax treatment of research and experimental expenditures. Under the historical statute, qualifying expenditures connected with a taxpayer’s trade or business could generally be deducted currently or, subject to the statutory conditions and election, amortized over at least 60 months. These historical provisions must be distinguished from the rules applicable to later tax years.

Martin Mayrath, an inventor associated with farm machinery, claimed research deductions for construction of his family’s Dallas residence. He argued that its unusual construction features supported experimental treatment. The Tax Court found inadequate evidence of a business connection and qualifying experimental expenditures. The Fifth Circuit affirmed the denial on the business-connection ground, emphasizing the evidence of personal use and the absence of a demonstrated profit-oriented housing venture.

The completed residence cost $287,474.11. The deductions claimed for 1956 and 1957 totaled $174,797.77. That total resulted from annual allocation calculations using estimated conventional construction costs; it was not simply the final excess construction cost, which the taxpayer calculated as $170,954.11.

Mayrath does not establish that research must relate to an already operating business in the taxpayer’s existing industry. In Snow v. Commissioner, 416 U.S. 500 (1974), the Supreme Court recognized the broader reach of the historical Section 174 business-connection language for research undertaken before an operating business began. A bona fide business connection remains distinct from a personal project.

The Judicial Construction of Investigative Activity and Product Development

The Tax Court discussed whether Mayrath’s construction costs were investigative expenditures associated with developing a product or model. Government testimony and the taxpayer’s cost allocations supported the court’s skepticism. The appellate court expressly declined to resolve the experimental-nature question because its business-connection holding disposed of the claim.

These historical observations do not create a modern requirement to advance an abstract theoretical concept. Nor do they categorically exclude trial and error. The practical lesson is narrower: unusual features, substantial expense, or an inventor’s involvement cannot alone establish qualifying research.

The 1994 Regulatory Shift and Technical Uncertainty

The 1994 revision of Treasury Regulation Section 1.174-2 clarified the uncertainty-based definition of research and experimental expenditures. The relevant inquiry concerns whether available information establishes the capability, method, or appropriate design for developing or improving a product. The nature of the activity matters more than the project’s eventual success or commercial novelty.

The source’s suggestion that modern litigation universally revives a superseded Mayrath standard overstates the evidence. Phoenix Design Group, Inc. v. Commissioner, T.C. Memo. 2024-113, was decided on December 23, 2024; it should not be described merely as a pending pretrial dispute. The Tax Court found that the three trial projects did not establish qualified research. That outcome illustrates the need for specific evidence of uncertainty and experimentation, rather than establishing a general requirement to discover information new to an entire industry.

The Four-Part Test and the Research Credit

The research expenditure deduction and the research credit are separate provisions. Deductible research spending does not automatically generate a credit. Following the 2025 statutory amendments, Section 41(d)(1)(A) refers to domestic research or experimental expenditures under Section 174A. Older cases and regulations use the historical Section 174 reference and must be read in their applicable tax-year context.

• Research expenditure requirement: The research must satisfy the applicable domestic research expenditure standard, including its connection with the taxpayer’s trade or business.
• Technological information requirement: The research must rely fundamentally on physical or biological sciences, engineering, or computer science.
• Business component and qualified purpose: The information must be intended to help develop or improve a product, process, computer software, technique, formula, or invention used in the business or held for sale, lease, or license. The relevant purpose concerns function, performance, reliability, or quality.
• Process of experimentation: At least 80% of the relevant research activities, measured on a cost or other consistently applied reasonable basis, must constitute elements of experimentation for a qualified purpose.

The analysis also must address statutory exclusions, such as foreign or funded research, and the separate rules defining eligible wages, supplies, and contract research expenses. Meeting the activity tests does not make every project cost a qualified research expense.

The Substantially-All Requirement: Insights from Little Sandy Coal

In Little Sandy Coal Co. v. Commissioner, T.C. Memo. 2021-15, affirmed at 62 F.4th 287 (7th Cir. 2023), the courts considered research claims arising from shipbuilding. The litigation examined a tanker and a dry dock as representative projects. The taxpayer’s reliance on the proportion of new or redesigned physical features did not establish the required proportion of experimental activities.

The Seventh Circuit affirmed the denial for insufficient proof but rejected the Tax Court’s categorical exclusion of direct support and supervision from the numerator. Such activities may count when they themselves satisfy the relevant research and experimentation requirements. Their job labels alone neither qualify nor disqualify them.

The 80% calculation concerns research activities and permits costs or another consistently applied reasonable measure; it does not mandate labor hours as the exclusive measure. The activity-based calculation also differs from the separate substantially-all rule for employee wage allocation. Passing one does not automatically satisfy the other.

Methodical Evaluation: Siemer Milling and Union Carbide

Siemer Milling Co. v. Commissioner, T.C. Memo. 2019-37, and Union Carbide Corp. v. Commissioner, T.C. Memo. 2009-50, emphasize evidence of an evaluative process rather than an unsupported assertion that a business was innovating. Describing a difficult project or an improved product does not explain which alternatives were tested or how the testing addressed uncertainty.

However, systematic trial and error is expressly recognized by Treasury Regulation Section 1.41-4(a)(5), alongside modeling and simulation. A formal laboratory setting or a particular written hypothesis template is not universally required. The evidence should establish the uncertainty, the alternatives, the evaluative work performed, and its technical purpose. Contemporaneous records can make that showing considerably stronger.

Human Capital and Reasonable Compensation

Suder v. Commissioner, T.C. Memo. 2014-201, concerned research credits associated with telephone equipment and software developed by Estech Systems, Inc. The court accepted much of the research qualification and time-allocation evidence but reduced the chief executive’s compensation included in qualified research expenses to reasonable amounts.

The historical Section 174(e) reasonableness limitation addressed the amount paid under the circumstances. It did not authorize rejecting research merely because an examiner considered the project an unreasonable business choice. For executive wages, taxpayers should substantiate both qualifying services and the compensation amount. Seniority, ownership, patents, or a product-development title alone cannot establish that all compensation qualifies.

Software Development and the Discovery Test Controversy

Norwest Corp. v. Commissioner, 110 T.C. 454 (1998), addressed software research under the legal framework applicable to its tax years. Its discussion of discovery and internal-use software should not be presented as a universal statement of current law.

Treasury Regulation Section 1.41-4(a)(3)(ii) expressly provides that taxpayers need not expand or refine the common knowledge of skilled professionals. Existing technology can be used in qualifying research. The question is whether the taxpayer faces relevant technical uncertainty and performs qualifying experimentation, rather than whether the work produces an industry-wide scientific advance.

Certain software developed primarily for internal use faces additional requirements under Section 1.41-4(c)(6), including innovation, significant economic risk, and commercial availability criteria. Exceptions and rules for software serving third parties can change that analysis. Routine installation, maintenance, or configuration is not automatically research, but software should be evaluated by its actual activities and applicable classification.

The Shrink-Back Rule and Identification of Research

Treasury Regulation Section 1.41-4(b)(2) provides the shrinking-back rule. If the requirements are not met for the overall business component, the analysis moves to the most significant subset of its elements and continues, where appropriate, until a qualifying subset is identified or the most basic element is reached.

Shrinking back does not mean claiming an arbitrary experimental percentage of a failed project. Each retained subset must satisfy the applicable requirements, and its expenses must be supported. The rule also does not require taxpayers to define every business component at the smallest possible level from the outset. The unidentifiable Idaho dispute in the source cannot support a more restrictive general rule.

Substantiation and Record Retention

Treasury Regulation Section 1.41-4(d) requires records detailed enough to substantiate the credit. It does not impose one mandatory recordkeeping format or a universal rule that only contemporaneous documents can be considered. Retrospective interviews and reasonable allocations may contribute evidence, but unsupported estimates and generic project narratives leave material gaps.

The patent safe harbor in Section 1.41-4(a)(3)(iii) addresses the technological-discovery requirement for qualifying issued patents. A patent application or invention disclosure does not itself receive that treatment. Even an issued patent does not establish every credit requirement or substantiate all claimed costs.

There is no universal 40-year retention rule for research credits. Historical base-period information can remain relevant to the regular credit calculation, including information from 1984–1988 for some taxpayers. The source’s attribution of a general 40-year rule to United States v. Quebe should therefore be removed. Under Treasury Regulation Section 1.6001-1(e), records must be retained as long as their contents may become material to administering federal tax law; the relevant period depends on the taxpayer’s facts.

Economic and Strategic Implications of Legislative Changes

The Tax Cuts and Jobs Act generally required capitalization of specified research expenditures for tax years beginning in 2022 through 2024, with five-year amortization for domestic research and 15-year amortization for foreign research, using a midpoint convention. A statement that domestic research must still always be amortized over five years is outdated.

Public Law 119-21, enacted July 4, 2025, added Section 174A. For tax years beginning after December 31, 2024, it generally permits current deductions for domestic research or experimental expenditures, with an elective capitalization alternative. Foreign research remains subject to 15-year amortization under Section 174. Transition provisions address unamortized domestic costs from 2022–2024 and certain eligible small businesses; elections and deadlines require separate review. Revenue Procedure 2025-28 provides implementation procedures.

Section 280C coordinates research deductions and credits. Under the current rules, the domestic research deduction generally is reduced for the credit unless a reduced-credit election applies. At a 21% maximum corporate rate, that election generally produces 79% of the otherwise determined credit. This is an election calculation, not a universal measure of every taxpayer’s net tax benefit, and the law applicable to the particular year must be used.

Future Outlook: Implications for U.S. Corporate R&D Strategy

Moving Beyond First-Article Novelty

Document why capability, method, or design was uncertain and how alternatives were evaluated. A prototype is evidence of an object produced, not automatic proof of experimentation. Conversely, using established materials or scientific principles does not by itself disqualify development work.

Time Tracking and Activity Mapping

Connect technical activities with employees, supplies, contracts, and accounting records. Use a supportable measurement basis for the substantially-all calculation. Assess support and supervision by what was performed, consistent with the appellate analysis in Little Sandy Coal.

The Role of Technical Witnesses

Use people with firsthand knowledge to explain the uncertainty, technical alternatives, tests, and design decisions. Their explanations should connect to available records and the costs claimed. A formal expert designation is not a substitute for detailed factual evidence.

Preparing for Examination

Maintain support for the claim’s scope and calculation. The source’s assertion that IRS campaigns universally require examination of every project and reject statistical sampling is unsupported. Revenue Procedure 2011-42 provides guidance on statistical samples and estimates. Sampling still requires a defensible design and substantiation of the sampled activities; it cannot turn nonqualifying work into qualified research.