R&D Tax Credit Case Analysis: Little Sandy Coal, Process of Experimentation, and Section 41 Substantiation

The Statutory Architecture of I.R.C. Section 41

Technical work does not qualify merely because it occurs in an engineering or manufacturing business. Section 41 requires analysis of the taxpayer’s activities for each relevant business component. A business component can be a product, process, computer software, technique, formula, or invention intended for sale, lease, or license, or for use in the taxpayer’s trade or business. The credit also requires eligible expense categories and a properly calculated base or alternative simplified credit.

The Four-Part Test for Qualified Research Activities

The familiar four-part framework combines the research-expenditure requirement, technological information requirement, development or improvement of a business component for a permitted purpose, and process of experimentation requirement. Each must be satisfied. The statutory cross-reference for research expenditures changed with the enactment of Section 174A; historical cases must be read under the law applicable to their tax years.

The credit analysis and the identification of qualified research expenses (QREs) are related but distinct. Satisfying an activity test does not automatically make every cost associated with a project creditable.

Detailed Analysis of Little Sandy Coal v. Commissioner

Little Sandy Coal Company, Inc., the parent of Corn Island Shipyard, Inc., claimed a research credit for the design and construction of 11 first-in-class vessels for its tax year ending in June 2014. The parties selected two representative vessels for trial: the Apex 720 Tanker and the Detyens Dry Dock.

The Core Conflict: Activities Versus Physical Components

The Tax Court rejected an attempt to establish the experimentation requirement by emphasizing the extent to which vessels or their physical features were new. The appellate court likewise found that newness and unsupported activity allocations did not demonstrate the required process. A redesigned feature may identify where to investigate qualification; its physical size or share of a vessel does not measure the research activities devoted to experimentation.

This distinction is useful for project reviews. A business should identify what employees actually did, the uncertainties addressed, and the alternatives evaluated. It should not assume that a large proportion of changed parts means a corresponding proportion of qualifying activities or costs.

The Calculation of the Substantially All Fraction

Treas. Reg. § 1.41-4(a)(6) measures the proportion of research activities constituting elements of a process of experimentation. The denominator is not automatically all project hours or all construction costs. The remaining research activities must satisfy the applicable research-expenditure requirement and must not be otherwise excluded by Section 41(d)(4). Routine commercial production does not become research merely by being included in an allocation.

Supply costs: Materials are not themselves activities. Supply expenses must be evaluated separately under Section 41(b); a high materials bill cannot substitute for evidence of experimentation. Where a cost-based method is used to measure activities, its design must remain consistent with the regulatory activity test.

Direct support and supervision: The Seventh Circuit rejected the Tax Court’s approach of categorically including these activities in the denominator but excluding them from the numerator. Their treatment depends on the research they support or supervise. They can count toward the numerator when they constitute elements of the qualifying process. Their inclusion is not automatic merely because wages might fit a qualified-services category.

The test is not restricted to people with engineer or researcher job titles. Production personnel may participate in relevant experimental activities, including qualifying pilot-model work, when the evidence supports that characterization. Conversely, an engineer’s routine work does not qualify solely because of the employee’s profession.

The Shrinking-Back Rule

Under Treas. Reg. § 1.41-4(b)(2), the tests apply first to the business component. If that component fails, the analysis proceeds to its most significant subset of elements and then to successively smaller subsets until a qualifying subset is identified or the most basic element also fails. The rule does not instruct taxpayers to begin with the smallest possible component.

A separately documented hull feature or propulsion subsystem could be a candidate for analysis, but the rule does not establish that a particular hypothetical subsystem would qualify. Shrinking back requires evidence for the subset; it does not cure missing evidence through relabeling. It is also not an automatic instruction to claim all costs of a smaller project.

The Evolution of the Process of Experimentation

The regulatory standard requires more than a general account of problem-solving. It calls for identifying technical uncertainty, identifying one or more alternatives, and conducting a process to evaluate those alternatives. This is a longstanding requirement, not a new rule created by a Reynolds shipbuilding decision.

Identification of Technical Uncertainty

Uncertainty must exist at the beginning of the relevant research activities. The information available to the taxpayer must leave unresolved the capability, method, or appropriate design for developing or improving the component. Aesthetic preferences, commercial demand, pricing, and financing uncertainty do not establish technological uncertainty.

These examples identify possible uncertainties, not established qualifying projects. The actual activities must satisfy the other requirements and exclusions.

Systematic Evaluation of Alternatives

A useful technical record explains the proposed alternative, how it was evaluated, the results, and any resulting changes. Depending on the work, the evaluation may involve modeling, simulation, or systematic trial and error.

• Identify the uncertainty and the proposed technical alternative.
• Describe the tests, calculations, models, or simulations used to evaluate that alternative.
• Record the findings and explain how they affected the next design or development decision.

Routine calculations using already available information may not establish experimentation. However, the use of established scientific principles is not itself disqualifying: the regulations expressly contemplate reliance on engineering and other established sciences. The issue is the nature of the evaluation and the unresolved technical question, not whether the underlying science is new. A failed experiment is not mandatory, and success does not disqualify otherwise eligible research.

Implications for Software Development and Automotive Retail

The activity-based framework applies across industries, including software developed by retailers and dealerships. The following applications are illustrative; the shipbuilding litigation did not decide the eligibility of a dealership’s software.

Software Innovation in Used Car Dealerships

A dealership might develop software for financing workflows, system interoperability, or operational processing. Purchasing a platform, configuring standard features, entering data, or making purely cosmetic interface changes ordinarily does not establish a qualifying experimental process. Integration work requires an assessment of the actual technical uncertainty and evaluation performed; it should not be categorically included or excluded based on the label “API integration.”

A company’s development or use of software does not by itself establish qualified research. A supportable software claim should:

• Identify the specific software business component and unresolved computer-science or engineering problem.
• Describe the alternatives and the process used to evaluate them.
• Measure the relevant activities consistently and connect claimed expenses to qualified services or other permitted cost categories.
• Determine whether the internal-use software rules and any exceptions apply.

The High Threshold of Innovation for Internal Use Software (IUS)

Treas. Reg. § 1.41-4(c)(6) imposes additional requirements on covered internal-use software, generally software developed for general and administrative functions supporting the taxpayer’s business. Not all software used internally falls within this category, and the regulation contains exceptions. Covered software must satisfy the ordinary research requirements and the high threshold of innovation test.

Software developed to enable third-party interaction or to allow third parties to initiate functions or review data generally is not internal-use software under these rules. Software with both internal and third-party uses requires the separate dual-function analysis. A lender portal therefore requires a review of its intended functions; its name alone does not resolve the issue. These software rules arise from the regulations, not from a holding in Little Sandy Coal.

Contractual Risk and the “Funded Research” Exclusion

Section 41(d)(4)(H) and Treas. Reg. § 1.41-4A(d), incorporated through § 1.41-4(c)(9), exclude research to the extent funded by another person. Relevant questions include whether payments depend on successful research and whether the taxpayer retains substantial rights in the results. This is a separate eligibility analysis from the substantially all test.

Substantial Rights and Local Law

Review the entire agreement, including intellectual-property clauses, restrictions on reuse, incorporated documents, and governing law. Substantial rights need not be exclusive, but a taxpayer retaining no substantial rights generally cannot claim the research credit for the contracted research. Incidental experience gained from performing services is not, by itself, a substantial right in the research results.

Contract silence does