IRC Section 41(d)(1): Qualified Research Defined
Under 26 U.S. Code Section 41(d)(1), “qualified research” includes research meeting the applicable research-expenditure requirement and utilizing the physical or biological sciences, engineering, or computer science to discover technological information, wherein substantially all activities constitute a process of experimentation to eliminate technical uncertainty for a permitted business purpose. This statutory definition limits the federal Research and Development (R&D) tax credit to qualifying research rather than cosmetic modifications or routine work that lacks the required uncertainty and evaluative process.
The application of this foundational statute underpins both the federal tax credit and state-level incentive programs, most notably the California Research and Development Tax Credit under Revenue and Taxation Code (R&TC) Section 23609. The statutory definition is deliberately stringent, requiring taxpayers to demonstrate that their development activities meet the applicable eligibility requirements. Instead, the activities must represent a deliberate, structured application of the scientific method aimed at creating or improving a business component. This comprehensive research study provides an examination of the core eligibility requirements set forth in 26 U.S.C. § 41(d)(1), focusing specifically on the interrelated mandates for discovering technological information and engaging in a verifiable process of experimentation. Furthermore, this analysis interprets these federal statutes through the lens of state-level frameworks, analyzing relevant United States congressional explanatory memoranda, state executive branch interpretations, state administrative interpretations, and pertinent federal case law to construct a holistic view of the modern innovation tax landscape.
The Federal Statutory Framework and the Four-Part Test
Under Section 41 of the Internal Revenue Code (IRC), an eligible taxpayer may claim a general business credit against income tax liability for increasing research activities. While the credit is intended to incentivize domestic innovation, the definition of what constitutes eligible “qualified research” is tightly controlled by Section 41(d)(1), which establishes a stringent “Four-Part Test” whose requirements are applied separately to each business component, with the substantially-all rule governing the experimentation element.
The first requirement, historically called the Section 174 Test, now refers in federal IRC Section 41(d)(1)(A) to domestic research or experimental (R&E) expenditures under IRC Section 174A. The costs must be incurred in connection with the taxpayer’s trade or business and must represent research and development costs in the experimental or laboratory sense, excluding expenses such as land acquisition, mineral exploration, or routine quality control. The applicable version of this requirement depends on the tax year and, for state credits, state conformity.
The second requirement is the Technological Information Test, which mandates that the research must be undertaken for the express purpose of discovering information that is “technological in nature.” The discovery of this information must be intended to eliminate objective uncertainty concerning the development or improvement of a business component.
The third requirement is the Business Component Test, frequently referred to as the Permitted Purpose test. The taxpayer must intend to apply the technological information discovered to develop a new or improved business component, statutorily defined as a product, process, computer software, technique, formula, or invention that is held for sale, lease, license, or used in a trade or business of the taxpayer. The research must specifically relate to a new or improved function, performance, reliability, or quality, and the statute explicitly prohibits research related to cosmetic, taste, or stylistic design factors.
The fourth requirement is the Process of Experimentation Test. The statute requires that substantially all (quantitatively defined by regulation as 80 percent or more) of the research activities must constitute elements of a process of experimentation for a qualified purpose.
| Statutory Requirement | Legal Reference | Core Objective | Key Exclusions |
|---|---|---|---|
| Research-Expenditure Eligibility | IRC § 41(d)(1)(A); IRC § 174A (current federal law); Treas. Reg. § 1.174-2 (historical framework) | Ensure costs are experimental or laboratory in nature and connected to a trade/business. | Mineral exploration, land acquisition, routine quality control, market research. |
| Technological Information | IRC § 41(d)(1)(B)(i); Treas. Reg. § 1.41-4(a)(3)–(4) | Ensure reliance on hard sciences (engineering, computer science, physical/biological sciences). | Social sciences, humanities, arts, economics, management studies. |
| Permitted Purpose | IRC § 41(d)(1)(B)(ii), (d)(3); Treas. Reg. § 1.41-4(a)(5)(ii) | Tie research to the functional improvement of a specific business component. | Aesthetics, style, taste, seasonal design factors, cosmetic changes. |
| Process of Experimentation | IRC § 41(d)(1)(C); Treas. Reg. § 1.41-4(a)(5) | Mandate a scientific, evaluative process to eliminate capability, method, or design uncertainty. | Unguided trial and error, duplication of existing components, routine production troubleshooting. |
Legislative History and Congressional Intent: The Tax Reform Act of 1986
The contemporary interpretation of 26 U.S.C. § 41(d)(1) cannot be understood without analyzing the legislative history of the Tax Reform Act of 1986 (TRA ’86). The original research tax credit, enacted in 1981, provided a subsidy for research spending in excess of a base amount but used a broader definition of qualified research. Prior to 1986, the definition of qualified research was viewed by Congress and the Treasury Department as too broad. The legislative history reveals profound congressional concern that businesses were claiming the credit for activities that had far more to do with routine product development, stylistic modifications, or market testing rather than genuine technological innovation. During the legislative hearings for TRA ’86, members of the Subcommittee on Oversight of the House Committee on Ways and Means, alongside Treasury officials, cited research credit claims submitted by fast food restaurants, fashion designers, and hair stylists as examples of activities that the credit was never intended to subsidize.
To narrow eligibility toward technological research, TRA ’86 introduced additional elements of the modern Four-Part Test: the explicit statutory requirements that the information discovered must be technological in nature and that the activities must fundamentally involve a process of experimentation. The Joint Committee on Taxation, in its General Explanation of the Tax Reform Act of 1986 (widely recognized as the “Blue Book”), provided an explanatory account for these sweeping changes. The Joint Committee emphasized that the revised credit was intended for research that fundamentally relies on principles of the physical or biological sciences, engineering, or computer science.
Furthermore, the Joint Committee’s explanatory memorandum clarified that the term “process of experimentation” meant a structured, evaluative process. The explanatory discussion describes formulating and testing hypotheses designed to achieve a result where the means of achieving that result are uncertain at the outset. It further describes the systematic testing and analyzing of those hypotheses—through mechanisms such as computational modeling, physical simulation, or a sequential, systematic trial-and-error methodology—and the subsequent refining or discarding of the hypotheses based on empirical data. This legislative history confirms that the 1986 modifications were explicitly driven by Congressional desires to target the credit toward industries involving high technology or its direct application in developing functionally superior products and manufacturing processes, excluding research undertaken for cosmetic and other nonqualifying purposes.
To illustrate the credit’s calculation, the Congressional Research Service used hypothetical firms rather than a sampled corporate panel. In its established-firm example, 1984 qualified research expenses were $5 million against $100 million in gross receipts, and qualified research expenses reached $18 million in 1989. A separate start-up-firm example used $105 million in qualified research expenses for 2003. These sample figures illustrate credit mechanics; they do not establish historical national research spending or demonstrate that the 1986 changes caused industry spending to rise.
Discovering Technological Information and Eliminating Uncertainty
The statutory mandate to discover technological information is a dual-pronged requirement: it governs both the underlying nature of the information sought and the specific purpose for which that information is utilized within the commercial enterprise.
Treasury Regulation § 1.41-4(a)(4) codifies the legislative intent of the TRA ’86 Blue Book, dictating that information is deemed technological in nature solely if the process of experimentation used to discover it fundamentally relies on the established principles of the physical or biological sciences, engineering, or computer science. This regulatory definition operates as a strict negative covenant, explicitly forbidding research based on the social sciences, arts, humanities, or economics from qualifying for the credit. However, the regulations also provide an important clarification for taxpayers: a company does not need to expand the global boundaries of human knowledge or make a profound academic discovery to meet this test. A taxpayer is fully permitted to employ existing, commercially available technologies and may rely on well-established scientific principles, provided that the application of those principles to their highly specific business component involves the resolution of technical uncertainty.
The concept of uncertainty is inextricably linked to the discovery of technological information. Treasury Regulation § 1.41-4(a)(3) dictates that research is undertaken for the purpose of discovering information if, and only if, it is intended to eliminate technical uncertainty concerning the development or improvement of a business component. The regulations state that objective uncertainty exists if the information available to the taxpayer at the commencement of the project does not clearly establish the capability to develop the component, the specific method required for developing it, or the appropriate design architecture of the final component. If the information available to the taxpayer already establishes the capability, method, and appropriate design, the required uncertainty is absent. Existing technology or industry knowledge alone does not categorically disqualify research.
The Core Engine: Engaging in a Process of Experimentation
The “Process of Experimentation” test under 26 U.S.C. § 41(d)(1)(C) can be a significant hurdle for taxpayers and is scrutinized by the Internal Revenue Service and state revenue departments during examinations. It requires a structured, empirical, and evaluative approach to resolving the technical uncertainties identified at the project’s inception.
Treasury Regulation § 1.41-4(a)(5) describes an evaluative process. First, there must be an identification of the specific uncertainty concerning the development or improvement of the business component. Second, the taxpayer must identify one or more alternatives intended to eliminate that precise uncertainty. Finally, the taxpayer must identify and conduct a process of evaluating those alternatives through scientifically valid means. The process must evaluate alternatives to resolve the identified technical uncertainty. It is profoundly insufficient to simply encounter a production problem, guess at a solution, implement the fix, and move on. There must be a methodical framework for testing potential solutions, measuring the empirical results against engineering or scientific baselines, and iterating upon those results as the research proceeds. The regulations notably concede that a process of experimentation can be undertaken even if there is no uncertainty concerning the taxpayer’s ultimate capability or method of achieving the desired result, provided that the appropriate design of the desired result remains uncertain at the beginning of the research activities.
A critical quantitative component of this test is the “Substantially All” requirement. IRC Section 41(d)(1)(C) requires that substantially all of the research activities constitute elements of a process of experimentation for a qualified purpose. Treasury Regulation § 1.41-4(a)(6) provides a quantitative standard, defining “substantially all” as 80 percent or more of a taxpayer’s research activities, measured on a cost or other consistently applied reasonable basis. If less than 80 percent of the relevant research activities for a business component meet this standard, the component fails this requirement at that level, subject to the shrinking-back rule.
To prevent the total loss of credits for complex projects that blend routine and experimental work, the Treasury Regulations provide the “shrinking-back” rule. If a taxpayer cannot satisfy the four-part test for an entire macro-level project, they are permitted to shrink back the analysis to a smaller, more focused subset of the project. For example, if a massive software deployment contains significant amounts of routine data migration (which does not qualify) alongside the development of a novel, highly complex predictive algorithm (which may qualify), the taxpayer applies the requirements to the most significant subset of elements and continues shrinking back as necessary. By isolating the experimental core, the taxpayer may satisfy the 80 percent threshold for that specific, shrunken sub-component, thereby salvaging the associated qualified research expenses.
Federal Judicial Interpretations: Drawing the Boundary of Experimentation
Federal courts have repeatedly interpreted 26 U.S.C. § 41(d)(1) to establish strict judicial boundaries regarding what constitutes a process of experimentation, consistently distinguishing true scientific inquiry from routine business operations and aesthetic design.
In Union Carbide Corp. and Subsidiaries v. Commissioner, T.C. Memo. 2009-50, the United States Tax Court explained the standard regarding the interpretation of the process of experimentation. The court held that the process of experimentation test inherently requires a structured method of discovering information and the strict application of the scientific method. The court emphasized that simple trial-and-error—defined as making unguided, sequential adjustments until a product merely works—does not qualify for the credit. Instead, the taxpayer must engage in “systematic trial and error,” which legally implies a higher standard of scientific methodology, analysis of results, and hypothesis testing. While a taxpayer does not necessarily need to test multiple distinct alternatives simultaneously, they should conduct a series of experiments if testing only a single alternative to constitute a valid process.
This standard was applied in the context of the fashion industry in Leon Max v. Commissioner, T.C. Memo. 2021-37, relying heavily on the precedents set in Siemer Milling Co. v. Commissioner. In this case, the Tax Court evaluated a prominent clothing designer’s attempt to claim the R&D credit for the development of new garments. The taxpayer argued that fitting, draping, and testing fabric for shrinkage constituted a process of experimentation. The Tax Court disagreed, finding that these activities were fundamentally related to style, taste, and seasonal design—which are explicitly excluded under IRC § 41(d)(3)(B). Furthermore, the court noted that the uncertainties the company faced, such as how to align prints or alter patterns for plus-size lines, did not require discovering information that was technological in nature. The court relied on the Merriam-Webster definition of engineering, noting that the employees utilized existing tailoring skills and basic garment construction techniques rather than applying engineering or hard science principles. The court explicitly found that activities such as washing fabric to test for shrinkage were quality control tests that did not establish qualifying experimentation on the facts presented.
In 2023, the Seventh Circuit Court of Appeals affirmed the denial of the research credit in Little Sandy Coal Co., Inc. v. Commissioner because the taxpayer failed to substantiate the substantially-all requirement under IRC Section 41(d)(1)(C). The court scrutinized how the company’s production and nonproduction activities related to experimentation. However, it rejected the Tax Court’s categorical exclusion of pilot-model production activities from the numerator of the 80 percent calculation. Such activities may constitute elements of experimentation when supported by the facts; the taxpayer’s evidentiary failure, rather than a blanket prohibition on production or supervisory work, was decisive.
State-Level Frameworks: California’s R&TC Section 23609 as the Anchor
While the federal statutes provide the baseline definition of qualified research, individual states adopt, modify, and strictly enforce these rules within their own sovereign borders. The California Research and Development Tax Credit, governed by Revenue and Taxation Code (R&TC) Section 23609, serves as the premier example of how a state adapts 26 U.S.C. § 41(d)(1) to drive localized economic policy. California largely conforms to the federal definitions, meaning the state requires taxpayers to pass the exact same Four-Part Test—including the rigorous requirement to discover technological information through a scientific process of experimentation. However, the state imposes several critical modifications that fundamentally alter how the credit is calculated, substantiated, and claimed.
The most profound modification is the imposition of strict geographical limitations. R&TC § 23609(c)(2) limits eligible research to research conducted within California. This geographical nexus requires meticulous accounting. For W-2 wages to be eligible, they must be paid to employees for qualified services physically performed within California. If an employee splits their time between a laboratory in San Francisco and a testing facility in Nevada, their qualified wages must be allocated under the applicable rules using adequate records. Similarly, any research performed by third-party contractors is subject to the same geographic restriction; the taxpayer bears the burden of proving the physical location of their subcontractors. In the modern era of remote computing, this nexus extends to infrastructure: if a taxpayer claims computer rental or cloud hosting costs as a qualified research expense, the taxpayer must establish that the costs are eligible computer-use expenses attributable to California research; server location and the nature of the services require fact-specific review.
California further diverges from the federal code in its calculation of the credit’s base amount. Under the federal regular-credit method, the base amount relies on a taxpayer’s historical gross receipts to measure the incremental increase in research intensity. R&TC § 23609(i) modifies the definition of gross receipts for the California computation. For state purposes, gross receipts are restricted solely to sales of property held primarily for sale to customers that is delivered or shipped to a purchaser within California, using a destination-based rule. This specific statutory modification completely excludes receipts from services, rents, royalties, operating leases, and sales shipped out of state. This mathematical alteration dramatically changes the fixed-base percentage, often resulting in a base amount that looks nothing like the federal calculation.
California also limits overlapping tax incentives. R&TC § 23609(c)(1) excludes from qualified research expenses amounts paid or incurred for tangible personal property eligible for the sales or use tax exemption under R&TC Section 6378. That provision concerns qualifying teleproduction or postproduction service equipment; it is not a general election between all R&D sales-tax exemptions and the research credit. The exclusion turns on eligibility for the specified exemption, not merely whether the taxpayer claims it.
Historically, California offered two distinct calculation methodologies: the reduced credit. The TCJA-era version used a different rule tied to the relationship between the credit and the deduction; taxpayers must apply the version governing the relevant year.
| Expense Characteristic | Section 41 QRE Eligibility | Federal R&E Treatment (Subject to Tax Year and Elections) |
|---|---|---|
| Domestic W-2 Research Wages | Eligible qualifying-service wages under the statutory wage definition | Domestic R&E generally deductible under § 174A for tax years beginning after 2024; earlier TCJA costs were subject to 5-year amortization, with transition relief available |
| Foreign Research Wages | Excluded | Foreign R&E generally amortized over 15 years |
| Research Supplies (Consumed) | Eligible if used in qualified research and not depreciable | Domestic qualifying costs generally deductible under § 174A; earlier TCJA treatment depended on the year |
| Depreciation on R&D Equipment | Excluded explicitly | Allocable depreciation may enter R&E costs; equipment acquisition itself is not an R&E deduction |
| Patent Attorney Fees | Excluded | Qualifying patent-procurement costs may enter R&E expenditures; domestic treatment depends on year and elections |
| Funded Research Costs | Excluded to the extent the research is funded under applicable rules | Depends on the contractual rights, financial risk, and applicable R&E rules |
The enacted One Big Beautiful Bill Act (OBBBA) created IRC Section 174A, generally allowing immediate expensing of domestic R&E expenditures for tax years beginning after December 31, 2024, while retaining the 15-year amortization rule for foreign research. State conformity remains selective. California updated its general IRC conformity date to January 1, 2025, but that update did not adopt the later OBBBA provisions.
California’s nonconformity does not require domestic R&E costs to be amortized over 5 years: California also did not adopt the TCJA’s mandatory research-expense capitalization changes and generally retains its earlier deduction framework. Taxpayers should therefore track federal and state treatment separately rather than assume that nonconformity eliminates immediate deductions. Separately, Texas Comptroller guidance clarified that depreciable-property expenses do not become Section 41 supply expenses merely because they qualify under Section 174. That guidance addresses the distinction between the expense and credit rules; it does not by itself demonstrate a broader multistate audit trend.
Practical Example: Software Development & Hardware Integration in California
To illustrate the dense interplay of IRC § 41(d)(1), the strict OTA Swat-Fame standard, and R&TC § 23609, consider the hypothetical AeroTech Systems, a company based in San Jose, California, developing a proprietary autonomous drone navigation module.
Phase 1: Conceptualization & Routine Engineering
AeroTech purchases commercial off-the-shelf (COTS) drone motors and assigns its engineers to write standard API integrations to link these motors to an existing open-source flight controller.
- Federal and State Treatment: On these assumed facts, these specific activities fail the Process of Experimentation test. Using standard APIs to link COTS hardware relies entirely on publicly available information and standard programming skills without any core technical uncertainty regarding capability or method. There is no scientific method applied, and no hypotheses are formulated.
Phase 2: Resolving Technical Uncertainty (The Scientific Method)
AeroTech attempts to develop a novel LiDAR data-processing algorithm to allow the drone to navigate through dense forest environments autonomously. There is extreme objective uncertainty regarding the capability and appropriate design of the algorithm because existing models require too much processing power, draining the battery too quickly.
- Hypothesis: The engineering team hypothesizes that a customized edge-computing data-culling algorithm can reduce the LiDAR data load by 40 percent without compromising obstacle detection speed.
- Experimentation: AeroTech sets up a systematic testing protocol. They run iterative computational simulations of the algorithm against baseline models, modifying code parameters in exact 5 percent increments, and strictly logging the processing load (CPU usage) and crash rates in a simulated digital forest matrix.
- Federal and State Treatment: These activities may satisfy the Technological Information and Process of Experimentation tests. The algorithm fundamentally relies on computer science. The evaluation of alternatives is highly systematic, documented via trial logs, and utilizes the scientific method to evaluate performance against objective engineering metrics, supporting eligibility under the standards applied by the California OTA, subject to substantiation and the remaining requirements.
Phase 3: Financial Qualification and Geography (California Rules)
Assume AeroTech spends $1,000,000 during the year, allocated as shown below. These are hypothetical amounts, and the 40 percent target and 5 percent testing increments above are illustrative assumptions. The eligible amounts assume that the stated wages and contract costs relate entirely to qualifying Phase 2 research, that the contract meets the applicable requirements, and that the cloud charges qualify federally as computer-use expenses. The example conservatively excludes the Nevada-server charges from California QREs pending a fact-specific location analysis.
| Expense Category | Amount | Location | Federal § 41 Eligibility | CA R&TC § 23609 Eligibility | Statutory Justification |
|---|---|---|---|---|---|
| W-2 Wages (Software Engineers) | $500,000 | San Jose, CA | $500,000 | $500,000 | Direct qualified services; physically performed in CA. |
| W-2 Wages (Algorithm Architect) | $200,000 | Austin, TX | $200,000 | $0 | Direct qualified services; but strictly excluded from CA due to out-of-state performance. |
| Contract Research (Data Testing) | $100,000 | San Diego, CA | $65,000 | $65,000 | Meets federal 65% statutory limit; physically performed within CA. |
| Cloud Computing Rental | $100,000 | Servers in NV | $100,000 | $0 | Assumed eligible federal computer-use expense; excluded from the California illustration pending location and service analysis |
| Depreciable Testing Server | $100,000 | San Jose, CA | $0 | $0 | Depreciable property is excluded from QRE supplies under § 41(b)(2)(C); R&E rules do not make its acquisition cost a credit-eligible supply |
| Total Project Spending and Assumed Eligible QREs | $1,000,000 | $865,000 | $565,000 | Illustrates geographic allocation and expense-category restrictions |
If AeroTech claims the California credit, it must substantiate the 80 percent experimentation requirement for each business component on a cost or other consistently applied reasonable basis. That test is not simply 80 percent of the $565,000 QRE total. Simulation logs, hypotheses, and code-version history can help establish what work occurred, alongside other credible evidence. For the 2025 tax year, AeroTech could elect the 3 percent Alternative Simplified Credit (ASC) on Form FTB 3523, subject to its base calculation and the 1.3 percent rule where applicable, or use the regular-credit method. California does not require the entire $1,000,000 to be amortized over 5 years under the TCJA; deduction or capitalization treatment must be determined by expense category under California law.
The Swanson Reed Advantage
Navigating IRC Section 41(d)(1) and decisions such as Swat-Fame can require both tax knowledge and technical expertise. Swanson Reed describes its team as including engineers, scientists, and tax professionals who work with taxpayers’ R&D teams. Such technical input may help identify technological uncertainty, document the process of experimentation, and evaluate differences between federal and state R&E treatment. Its effectiveness depends on the engagement and the underlying evidence; neither comparative superiority over other accounting firms nor a particular credit amount or audit outcome is assured.
This study reflects information reviewed as of October 13, 2026, and is provided for informational purposes only. Please seek legal or tax representation to understand how this applies to your own circumstances.
Frequently Asked Questions
What does IRC Section 41(d)(1) require for research to be treated as qualified research?
Section 41(d)(1) requires research to meet the applicable research-expenditure requirement, rely on the physical or biological sciences, engineering, or computer science to discover technological information, be intended to develop a new or improved business component for a permitted purpose, and involve a process of experimentation for substantially all (80 percent or more) of the research activities.
Does simple trial and error count as a process of experimentation?
No. In Union Carbide Corp. and Subsidiaries v. Commissioner, the Tax Court held that unguided trial and error does not qualify, and that taxpayers must instead engage in systematic trial and error involving scientific methodology, analysis of results, and hypothesis testing. The California Office of Tax Appeals applied the same standard in In re Swat-Fame, Inc.
How does the California research credit differ from the federal credit?
California largely conforms to the federal Four-Part Test but limits eligible research to work performed within California, restricts gross receipts to California destination sales for the base calculation, offers a 3 percent Alternative Simplified Credit for taxable years beginning on or after January 1, 2025, and allows unused credits to be carried forward indefinitely rather than for a maximum of 20 years.








