The Internal Revenue Code (I.R.C.) [§ 41] research and development (R&D) tax credit remains one of the most litigated and complex provisions within the United States tax system, primarily due to its reliance on highly fact-intensive inquiries and rigorous documentation standards. Recent judicial developments, characterized by the Tax Court’s rulings in 2021 and 2024, have established a new paradigm for how taxpayers must substantiate their claims. Central to this evolution is the “substantially-all” test, a quantitative threshold requiring that at least 80% of a taxpayer’s research activities for a given business component constitute elements of a process of experimentation. The case of Leon Max, Inc. v. Commissioner—and the thematic parallels drawn from the contemporary case of Little Sandy Coal Co. v. Commissioner—illustrates a significant shift toward a narrower, more rigid interpretation of what constitutes qualified research.

Statutory Foundations and the Four-Part Test Architecture

The federal R&D tax credit is designed to be incremental, encouraging businesses to expand their technological capabilities beyond routine operations. To qualify for the credit, research activities must meet the “four-part test” established under § 41(d). This framework serves as the gatekeeper for all qualified research expenditures (QREs), including employee wages, supplies, and contract research costs. See also how the four-part test works in simple terms.

The Section 174 Test: Experimental or Laboratory Sense

The first requirement mandates that expenditures be eligible for deduction under I.R.C. [§ 174]. This necessitates that the activities be “research and development” in the experimental or laboratory sense, specifically aimed at resolving technological uncertainty. Uncertainty exists when the information available to the taxpayer at the outset of the project does not objectively establish the capability or method for developing or improving the product, or the appropriate design of the product. Recent rulings, such as Phoenix Design Group, Inc. v. Commissioner, emphasize that this uncertainty must be technological rather than business-related or economic. Compare the difference between the Section 41 credit and Section 174 expenses and elimination of uncertainty.

The Technological Information Test: Reliance on Hard Sciences

The second part of the test requires that the research be undertaken to discover information that is “technological in nature“. This is defined as research that fundamentally relies on principles of physical or biological sciences, engineering, or computer science. Judicial interpretations have increasingly scrutinized activities that rely on “common knowledge” or professional expertise that is standard within an industry. In Leon Max, Inc. v. Commissioner, the court concluded that designers and patternmakers who used their standard training to adjust garment fit were not discovering technological information, as their work was rooted in non-technical style and taste rather than hard science. Apparel taxpayers should also review apparel and fashion industry R&D tax credits.

The Business Component Test: Functional Improvement

Third, the taxpayer must intend to use the discovered information to develop a new or improved “business component”. A business component includes any product, process, software, technique, formula, or invention held for sale, lease, or license, or used by the taxpayer in its trade or business. The improvement must relate to the function, performance, reliability, or quality of the component.

The Process of Experimentation Test: Systematic Evaluation

The final, and perhaps most contentious, requirement is the “process of experimentation” test. This requires that “substantially all” (80% or more) of the research activities constitute elements of a process of experimentation. This process must involve a systematic evaluation of design alternatives, such as modeling, simulation, or iterative testing, to resolve the identified technological uncertainty. For applied guidance, see how to document a process of experimentation and how to calculate the substantially all rule.

Four-part test elements under I.R.C. § 41(d)

Analysis of the Substantially-All Test and Its Mathematical Application

The “substantially-all” rule has been transformed by recent case law from a qualitative guideline into a strict mathematical requirement. In Little Sandy Coal Co. v. Commissioner, the court clarified that the test is applied to activities, not to the physical parts of a business component. This distinction is critical for taxpayers in industries such as shipbuilding, manufacturing, and construction, where large-scale prototypes are common. Related reading: do manufacturing process improvements qualify for R&D credits.

The Tax Court in Little Sandy Coal established a fraction to determine if the 80% threshold is met.

In this calculation, the Tax Court excluded wages paid for “direct supervision” and “direct support” from the numerator, even though these are qualified research expenses (QREs) under § 41(b). The reasoning was that while supervision and support support the research, they do not themselves “constitute elements of a process of experimentation.” On appeal, the Seventh Circuit disagreed with this categorical exclusion, holding instead that direct support and direct supervision expenses should be included in both the numerator and denominator whenever they are properly deductible research expenses under § 174. Even so, the appellate court affirmed the outcome, since the taxpayer still could not offer a principled way to determine what portion of its activities constituted elements of a process of experimentation—a ruling that caused a “sea change” in how the IRS approaches audits. See what triggers an R&D tax credit audit and how to defend an R&D tax credit claim in an audit.

The “Novelty” Fallacy in the Substantially-All Test

A common argument rejected by the courts in Leon Max and Little Sandy Coal is the “novelty” argument—the idea that because a product is a first-of-its-kind prototype, all activities associated with its creation are experimental. The Tax Court ruled that “novelty” of design has no probative value in satisfying the test. Even if a taxpayer creates a complex, novel ship, they must still prove that 80% of the labor hours associated with that ship were spent on systematic testing or modeling to resolve technical uncertainties, rather than routine fabrication.

Impact of Leon Max v. Commissioner and the Garment Industry Precedent

The litigation involving the apparel industry in Leon Max, Inc. v. Commissioner provides a cautionary tale regarding the definition of “discovery”. The taxpayer sought to claim credits for the design of high-end garments, arguing that the iterative process of patternmaking and sample creation was a form of experimentation.

The Tax Court’s “scorched-earth” opinion in this area focused on several key failures:

1. The Professional Knowledge Bar: The court noted that patternmakers and sample makers were using “common knowledge” standard in their field. To qualify for the credit, the research must seek to discover information that is not already known to a professional in that field.
2. Quality Control vs. Experimentation: The court found that much of the garment testing was “quality control testing” designed to meet internal standards or industry norms, rather than an experimental process designed to resolve a technological unknown.
3. Style and Seasonality: Activities related to “style, taste, and seasonal design factors” are expressly non-qualified under I.R.C. § 41(d)(3)(B). The court ruled that even the non-disqualified activities (the technical construction of the garments) did not undergo a systematic process of experimentation.

These findings have deep implications for all creative and engineering-adjacent industries. If an activity can be characterized as a professional using their existing toolkit to reach a standard outcome, it likely fails the “discovery” requirement of § 41, regardless of how many iterations it takes to finish the project. See also what expenses are excluded from R&D tax credits.

The Funded Research Exclusion and Economic Risk

Another significant hurdle clarified in the 2021-2024 jurisprudence is the “funded research” exclusion under § 41(d)(4)(H). Research is considered funded if a third party pays for it and the taxpayer does not bear the financial risk of the research’s failure, or if the taxpayer does not retain “substantial rights” in the research.

In Tangel v. Commissioner, the court disallowed research credits because the agreement between the researcher and the customer did not allow the researcher to retain substantial rights. Even if the researcher bears the risk of not getting paid if the research fails, the research is still considered “funded” if the customer owns all the intellectual property and the researcher cannot use the results without further payment or permission. Contractors and outsourced development raise related issues; see claiming R&D credits when development is outsourced and the 65 percent contractor expense rule.

The Fixed-Price Contract Trap

Many taxpayers believe that a fixed-price contract inherently signals economic risk because the taxpayer must absorb any cost overruns. However, in Meyer, Borgman & Johnson, Inc. v. Commissioner, the court ruled that a fixed-price contract is not enough. The contract must explicitly tie payment to the success of the research. If the contract provides for payment upon reaching certain milestones that are merely progress-based rather than achievement-based (resolving a technical hurdle), the IRS may view the research as funded. Architecture and engineering firms should also review whether architects can claim R&D tax credits and R&D tax credits for architecture and engineering firms.

Funded research outcomes in leading contract cases

The Shrinking-Back Rule as a Remedial Measure

When a taxpayer fails the substantially-all test at the project level, they often turn to the “shrinking-back” rule. This rule allows the four-part test to be applied at a more granular level—to a sub-component, a sub-process, or a specific system—until a qualifying “business component” is identified.

In Phoenix Design Group, the court acknowledged the shrinking-back rule but ultimately could not apply it. The reason for this failure was the same as the project-level failure: inadequate documentation. To successfully shrink back, a taxpayer must have contemporaneous records that isolate the costs and activities of the specific sub-component. For example, if a shipbuilder cannot qualify the entire ship because fabrication costs are too high, they might shrink back to a novel hull design. However, if they cannot show exactly which hours were spent by which engineers on that hull specifically (and not on the rest of the ship), the claim will still be denied. See what is a contemporaneous record for R&D tax credits and what documents are needed for an R&D tax credit study.

Documentation Standards and the 2021 IRS Memorandum

The Internal Revenue Service has signaled a new level of aggressiveness in challenging R&D credit claims, particularly refund claims. In October 2021, the IRS Office of Chief Counsel published a memorandum that significantly raised the bar for what must be included in a valid refund claim. See the impact of Chief Counsel Memorandum 20214101F and latest IRS guidance on R&D tax refund claims.

According to the Memorandum, a valid claim must:

1. Identify Business Components: List every specific product or process the credit relates to for that year.
2. Identify Research Activities: Describe exactly what was done for each component.
3. Identify Individuals: Name every employee involved in the research for each component.
4. Identify Information Sought: State precisely what technological uncertainty each person was trying to resolve.

This memorandum essentially requires taxpayers to provide “audit-ready” documentation at the time they file their claim, rather than waiting for an exam. Failure to provide this information can lead to the claim being rejected as “deficient” without a merits review. Practical drafting resources include drafting a project narrative and an example of a technical uncertainty description for the IRS.

Refund-claim documentation items emphasized by the 2021 IRS memorandum

Implications for Future R&D Tax Credit Applications

The cumulative effect of Leon Max, Little Sandy Coal, Phoenix Design Group, and the 2021 IRS Memorandum is a “High Burden of Proof” for all future R&D credit applications. Tax professionals must now advise their clients that arbitrary estimates of time spent on R&D will no longer suffice. See whether time estimates can be reconstructed if hours were not tracked.

The Necessity of Contemporaneous Records

The courts have repeatedly stressed that retrospective “R&D studies” conducted years after the fact are less persuasive than records created “as they happen”. Taxpayers should implement time-tracking systems that allow employees to tag their hours not just to a project, but to specific experimental tasks within that project. Related FAQs: do I need timesheets for R&D tax credits, how to track R&D hours, and best software for tracking R&D time and expenses.

Redrafting Supply and Service Agreements

To avoid the “funded research” trap, companies that perform work for clients must review their contracts. It is no longer enough to assume that a fixed-price contract protects the credit. Contracts should explicitly state that payment is contingent on the successful resolution of technological uncertainties and that the researcher retains a “substantial right” to the developed technology.

Statistical Sampling and the 80% Threshold

In large-scale R&D environments, the IRS allows for “statistical sampling” under Rev. Proc. 2011-42. However, the 2021 Memorandum creates a conflict: if a taxpayer uses sampling, they only have granular information for the units in the sample, which might not satisfy the “identify all individuals” requirement for the entire claim. This remains a gray area that may lead to further litigation.

Economic Consequences of Non-Compliance: Accuracy-Related Penalties

The shift in judicial temperament is also evident in the sustained application of accuracy-related penalties under § 6662. In the past, taxpayers who made a “good faith” attempt to calculate their credit were often spared penalties even if the credit was disallowed.

Recent cases show a different trend:

• In Phoenix Design Group, the Tax Court sustained a 20% penalty because the taxpayer’s work was deemed to be “routine engineering” rather than research.
• In Little Sandy Coal, the penalty was upheld because the taxpayer did not offer a “principled way” to determine what portion of activities were experimental.
• In valuation-related cases involving conservation easements, gross valuation misstatement penalties of up to 40% have been imposed when a claimed value substantially exceeded the property’s actual value.

These penalties reinforce the message that poor documentation and aggressive “estimations” of R&D activity are no longer just a risk to the credit itself, but a direct threat to the company’s bottom line. See whether a taxpayer can be fined for a bad claim, what happens if an R&D tax credit is denied, and the success rate of R&D tax credit audits.

Future Outlook: Amortization and the Credit Intersection

Starting in 2022, I.R.C. § 174 requires that all R&D expenditures (including those underlying the § 41 credit) be amortized over five years (fifteen years for foreign research) rather than being d