Union Carbide Corp. & Subsidiaries v. Commissioner

The legal framework governing the Credit for Increasing Research Activities, as established under Section 41 of the Internal Revenue Code (IRC), represents one of the most complex and litigated areas of federal tax law. At the center of this complexity lies the landmark case of Union Carbide Corp. & Subsidiaries v. Commissioner, a dispute that fundamentally redefined the eligibility of supply costs and the standards for documenting the process of experimentation in large-scale industrial environments. This case, originating from the 1994 and 1995 tax years and culminating in a significant Second Circuit affirmation in 2012, has forced a recalibration of how American manufacturers approach Research and Development (R&D) tax incentives. The judicial determination that only incremental supply costs—those incurred over and above routine production expenses—qualify for the credit has effectively ended an era where entire commercial production runs could be leveraged to generate massive tax offsets.

The Statutory Genesis and the Congressional Intent of Section 41

To appreciate the impact of Union Carbide, one must first examine the statutory architecture of the research credit. Enacted as part of the Economic Recovery Tax Act of 1981, the credit was designed to stimulate technological advancement and reverse the decline in American industrial competitiveness. Unlike a standard deduction, which merely reduces taxable income, the Section 41 credit provides a dollar-for-dollar reduction in tax liability, making it a highly scrutinized "tax expenditure".

The primary hurdle for any taxpayer is the "four-part test," a set of criteria that must be satisfied for every individual "business component"—a term that encompasses products, processes, computer software, techniques, formulas, or inventions held for sale, lease, or license. The Union Carbide decision focused heavily on the intersection of the Section 174 Test and the Process of Experimentation Test, particularly in the context of improving existing manufacturing processes.

The Union Carbide litigation primarily contested the interpretation of "supplies used in the conduct of qualified research" under Section 41(b)(2)(A)(ii). While the taxpayer argued for a broad linguistic interpretation, the Commissioner of Internal Revenue advocated for a functional, economic standard that excluded routine production inputs. This tension between the "plain meaning" of the word "use" and the "incremental" purpose of the research credit forms the core of the case’s legal legacy.

The Hahnville Projects: A Technical Review of the Disputed Research

The controversy arose at Union Carbide’s petrochemical production facilities in Hahnville, Louisiana. Union Carbide Corporation (UCC) sought to claim additional research credits for 106 manufacturing process improvement projects, representing millions of dollars in claimed Qualified Research Expenses (QREs). The Tax Court, presided over by Judge Goeke, selected five representative projects to determine the eligibility of the wider claim. Three of these projects—the Amoco Anticoking project, the UCAT-J project, and the Sodium Borohydride project—became the primary subjects of the ensuing appellate discourse.

The Amoco Anticoking project attempted to resolve technical uncertainty regarding whether a specific pretreatment could diminish the creation of coke in furnaces. Despite the technical complexity, the court found that the production process yielded a normal amount of ethylene, and because UCC sold this product, the raw materials were already destined for commercial use. Similarly, the UCAT-J project, which ran 19 times before being discontinued due to high levels of "off-grade" product, was deemed to be qualified research in its intent but failed to justify the inclusion of the base raw material costs as research expenses.

The "Used in the Conduct" Doctrine: Linguistic vs. Functional Statutory Interpretation

The defining legal question in Union Carbide was the interpretation of the phrase "used in the conduct of qualified research". UCC argued that the dictionary definition of "use"—to put into action or service, employ, or utilize—should govern. Under this view, if a molecule of ethylene was present in a furnace while an experiment was being conducted, that molecule was "used" in the research.

The Second Circuit Court of Appeals, however, rejected this isolationist reading of the word "used". The court emphasized that a statute must be read in the context of its specific purpose and the broader statutory scheme. It concluded that "in the conduct of qualified research" specifies the type of use that is creditable. The court reasoned that providing a credit for supplies that would have been purchased anyway to produce goods for sale would result in an "unintended windfall" for the taxpayer. This created the "incremental" or "additional" supply cost standard, which dictates that only those costs which would not have been incurred but for the research activity are eligible for the credit.

The court’s alignment with the Commissioner’s "indirect cost" argument relied on Treasury Regulation § 1.41-2(b)(2), which excludes indirect research expenditures such as general and administrative expenses or overhead. By categorizing routine raw material costs as "indirect" research costs when those materials are also used for commercial production, the court effectively built a wall between COGS (Cost of Goods Sold) and QREs.

The Process of Experimentation (POE) and the Scientific Method Requirement

A critical failure for UCC occurred in the Sodium Borohydride project, which the court determined did not satisfy the Process of Experimentation test. To meet the POE standard, a taxpayer must demonstrate that "substantially all" (defined as 80% or more) of the research activities constitute a process designed to evaluate one or more alternatives to achieve a result where the capability, method, or design is uncertain at the outset.

The court found that UCC’s activities in the Sodium Borohydride project amounted to "simple validation testing" rather than a true experimental process. The project involved adding a chemical to a production run to see if it reduced unwanted byproducts, but UCC failed to perform rigorous post-testing analysis or comparisons of the data collected. This led to the judicial conclusion that a systematic inquiry—one involving a methodical plan to test, analyze, refine, and retest a hypothesis—is the bedrock of qualified research.

Elements of a True Process of Experimentation (POE)

• Identification of Technical Uncertainty: The taxpayer must define a specific uncertainty regarding the capability, method, or appropriate design of a business component.
• Evaluation of Alternatives: The process must involve identifying and evaluating multiple alternatives intended to eliminate that uncertainty.
• Iterative Testing and Analysis: The taxpayer must use a structured, scientific approach (often called a "design spiral") to test and refine the chosen solution.
• Follow-up and Refinement: Meaningful analysis of data must lead to a refinement of the initial hypothesis or design.

The Union Carbide ruling clarifies that "simple trial and error" to validate a change is insufficient without the surrounding scientific structure and documentation. This has significant implications for future R&D applications, particularly in fields like software engineering or architectural design where "design iterations" are often mistaken for "experimentation".

The Shrink-Back Rule: A Narrowing Path for Complex Systems

The "shrink-back rule" is a regulatory mechanism that allows a taxpayer to salvage a research credit when a high-level business component fails the four-part test. Under this rule, if the overall product or process does not qualify, the test is applied to the next most significant subset of elements. This continues until a qualifying subset is found or the most basic element is reached.

In Union Carbide, the taxpayer attempted to treat the entire production process as the business component. However, the court emphasized the dichotomy established by Section 41(d)(2)(C), which requires that any plant process or technique for commercial production be treated as a separate business component from the product being produced. This separation meant that while the "process" might involve research, the "product" (the chemicals sold) did not. Because UCC could not provide a principled way to allocate costs specifically to the experimental sub-processes, the shrink-back rule could not be used to justify the inclusion of raw material costs.

The "substantially all" requirement—meaning 80% or more of research activities must be experimental—is the gatekeeper for the shrink-back rule. In Little Sandy Coal, the Seventh Circuit affirmed that if a taxpayer cannot differentiate between the "research" and "production" elements of a pilot model, the presence of production activities in the denominator of the fraction can easily pull the percentage below the 80% threshold, disqualifying the entire project.

Substantiation Standards: From the Cohan Rule to Modern Documentation Mandates

Historically, taxpayers relied on the "Cohan Rule"—derived from Cohan v. Commissioner—which allows courts to estimate a taxpayer's expenses if it is clear that some expenses were incurred, even if the exact amount is not proven. The Tax Court in Union Carbide initially provided some "good news" for taxpayers by holding that the Cohan Rule could be used to estimate R&D credits and that oral testimony from employees was admissible evidence.

However, the subsequent decade has seen a dramatic shift toward more stringent documentation standards. Rulings like Phoenix Design Group (2023) have emphasized that while estimates are a framework, they cannot replace "fundamental existence" proven through solid documentation. The IRS now routinely disallows credits based primarily on ad hoc methodologies or "prepackaged" studies prepared by consultants that rely heavily on retrospective interviews rather than contemporaneous project logs.

The recent Tax Court decisions in Meyer, Borgman & Johnson and Phoenix Design Group signal that the industry bar for "adequate records" has been raised. Sparse hour logs and unexplained drawings are no longer sufficient to substantiate the "uncertainty" or the "process of experimentation".

The IRS Response: Directives, Form 6765, and the Path to 2026

The legacy of Union Carbide has been formalized through a series of IRS administrative actions designed to increase transparency and consistency in R&D audits. Two major developments—the ASC 730 Directive and the overhaul of Form 6765—now dictate the compliance landscape for large and mid-market taxpayers.

The ASC 730 LB&I Directive

To reduce the administrative burden on Large Business & International (LB&I) taxpayers (those with assets ≥ $10 million), the IRS issued a directive in 2017 (revised in 2020) that allows taxpayers to use financial statement R&D expenses (ASC 730) as a proxy for QREs. If a taxpayer satisfies the directive’s requirements, the IRS will generally accept the "Adjusted ASC 730" amount as sufficient evidence of QREs for the credit year.

This directive serves as an "administrative solution" to the substantiation problems seen in Union Carbide by aligning tax reporting with the rigorous internal standards used for GAAP-compliant audited financial statements. However, it requires a series of downward adjustments to remove non-qualifying costs like stock-based compensation, overhead, and—crucially—production-scale supply costs that do not meet the Union Carbide incremental standard.

The Evolution of Form 6765: Section G

Beginning in 2025 (optional) and 2026 (mandatory for large filers), the IRS is introducing "Section G" to Form 6765. This new section transforms the form from a purely quantitative summary into a qualitative disclosure document.

Section G requires taxpayers to list the top 50 business components (by QRE amount) or those accounting for 80% of total QREs. For each component, the taxpayer must describe the information sought for discovery and provide a granular breakdown of qualified wages into direct performance, direct supervision, and direct support. This aligns the return filing with the documentation previously only requested during intense audits, essentially "front-loading" the burden of proof.

Industry Implications: The Economic Impact on Manufacturing and Innovation

The Union Carbide decision has had a "dampening effect" on the incentive for process research in the manufacturing sector. The National Association of Manufacturers and the U.S. Chamber of Commerce argued that the "incremental" standard ignores the reality of modern industrial R&D, where experimentation must occur at scale to be valid. In industries like chemical refining or aeronautics, small-scale laboratory tests often fail to capture the variables (pressure, heat, fluid dynamics) present in full-scale production.

By denying a credit for the materials placed at risk during these at-scale tests, the court has effectively increased the financial barrier to entry for large-scale innovation. For example, if a company tests a new catalyst in a multi-million dollar production run and that run produces "off-grade scrap," the loss of that inventory is an economic cost of the research. Yet, under the Union Carbide standard, the cost of that inventory remains non-creditable because it was "intended" for sale, regardless of the experimental risk.

Economic Challenges for Production-Scale Research

• Risk of Loss: The supplies consumed in plant-scale testing are "trivial in comparison to the supplies that must be placed at risk of loss" if the experiment fails.
• Capital Intensity: Large-scale industries face higher hurdles for the "substantially all" test because the high cost of production wages and materials can easily dwarf the "experimental" costs in the credit calculation.
• Data Granularity: The requirement to isolate "additional" costs demands advanced cost-accounting systems that many mid-market manufacturers do not possess.

Comparative Case Analysis: The Narrowing Scope of "Qualified Research"

The decade following Union Carbide has seen a steady narrowing of what the judiciary considers "qualified research." Cases like Siemer Milling and Suder v. Commissioner have reinforced the need for a "methodical plan" over simple trial and error. In Siemer Milling, the court found that slight adjustments to machines (like replacing bearings) were "ordinary construction work" rather than experimentation, even though they relied on engineering principles.

Similarly, the Little Sandy Coal case (2023) highlighted the "all-or-nothing" danger of the 80% rule. The taxpayer attempted to claim credits for the construction of a pilot barge, but because the construction activities themselves (welding, painting, assembly) were not "experimental," they were included in the denominator of the POE fraction. This pushed the experimental percentage below 80%, resulting in a total disallowance of the barge as a business component.

These trends suggest that the IRS and the courts are moving toward a highly structured, "scientific sense" definition of research that leaves little room for the "applied design" and "technical adaptation" common in the engineering and architectural sectors.

Future Outlook: Navigating the R&D Landscape in 2026 and Beyond

As businesses look toward the 2025 and 2026 tax years, the lessons of Union Carbide are more relevant than ever. The IRS's extension of the "refund claim perfection" period through January 10, 2027, provides a small window for taxpayers to refine their documentation before final determinations are made. However, this "perfection" requires providing the very details Union Carbide was unable to produce: specific business components, activity descriptions, and detailed QRE breakdowns.

Strategic Recommendations for Tax Professionals

• Tagging at Source: Companies must begin tagging QREs at the source (payroll, general ledger, accounts payable) to specific business components to meet the requirements of Form 6765 Section G.
• Formalizing Hypotheses: Documentation should move beyond "test logs" to include formal hypothesis statements, analysis of alternatives, and documented conclusions for every major project.
• Isolation of Incremental Costs: For at-scale testing, accounting systems must be able to isolate the "extra" costs (e.g., specialized catalysts, additional energy, wasted raw materials) that would not exist in a standard production run.
• Contractual Review: For architectural and engineering firms, contracts must be reviewed to ensure they are "unfunded"—meaning the taxpayer bears the financial risk and retains substantial rights to the research.

Conclusion: The Permanent Legacy of Union Carbide

The Union Carbide Corp. & Subsidiaries v. Commissioner case has left an indelible mark on the American tax landscape. It served as the judicial "stop-gap" that prevented the research credit from becoming a general manufacturing subsidy. By enforcing an "incremental" standard for supply costs and a "scientific method" standard for the process of experimentation, the courts have ensured that the Section 41 credit remains focused on its original purpose: incentivizing new and improved technological breakthroughs rather than subsidizing the routine costs of commercial production.

For future applicants, the case serves as a stark reminder: technical complexity is not a substitute for experimental structure, and physical necessity is not a substitute for incremental cost. As the IRS moves toward a more transparent, project-level reporting system via the new Form 6765, the ability to substantiate the "uncertainty" and the "experimentation" with contemporaneous, granular data will be the only way to protect these valuable tax incentives from disallowance. The "windfall" era is over; the era of the "documented experiment" has truly begun.