R&D Tax Credit Deep Dive: Little Sandy Coal Company, Inc. v. Commissioner

March 10, 2021 – In Little Sandy Coal Company, Inc. v. Commissioner,[1] the Tax Court sustained the IRS’s denial of a research credit claimed by a company for expenses incurred by its shipbuilding subsidiary because the company failed to show that at least 80% of the subsidiary’s research activities constituted elements of a process of experimentation (POE substantially-all test) under I.R.C. section 41.

The Taxpayer’s subsidiary, Corn Island Shipyard, Inc. (“CIS”), built custom barges and specialized marine structures. CIS claimed a research credit for building 11 vessels. The Taxpayer and the IRS agreed to treat two projects, the Apex Tanker and Dry Dock projects, as representative samples of the types of activities undertaken by the Taxpayer.

While the primary issue was whether the Taxpayer had substantiated the POE substantially-all test, this case is significant for several reasons.

Primary Takeaways:

  1. POE Substantially-All Test:
    • Activity-based test – That at least 80% of a taxpayer’s research must constitute elements of a process of experimentation applies to activities—not to physical components of the product being developed or improved. Therefore, the requirement is not satisfied simply because at least 80% of the product’s elements are new or differ from those of products the taxpayer previously developed.


The substantially-all test requires an analysis of the ratio of (i) research activities conducted in regard to a business component that constitute elements of a process of experimentation to (ii) all research activities in regard to the business component that satisfy section 41(d)(1)(A) that are not covered by one of the section 41(d)(4) exclusions from qualified research.


    • Direct supervision and direct support activities are not included in calculating the numerator of the POE substantially-all test –  These activities do not constitute elements of a process of experimentation and cannot be included in calculating the numerator of the POE substantially-all test. Relying on I.R.C. section 41(b)(2)(B), the court concludes that while these activities are “qualified services,” they do not constitute “qualified research” requiring a process of experimentation.


  1. Pilot Model


    • Must truly be a prototype to be eligible as a section 174 pilot model –  The taxpayer’s purpose in building the model must be to evaluate and resolve uncertainty during development or improvement of the product and test the appropriate design. If a taxpayer builds a product simply to fulfill its contractual obligations to its customer, the taxpayer cannot claim the cost as a qualified supply expense.


    • Distinguishing Trinity – The Tax Court rejects the holding in Trinity Indus., Inc. v. United States,[2]a U.S. District Court decision, in which the court ruled the POE substantially-all test was met for two of the six vessels at issue based largely on its conclusion that aspects of those vessels were novel compared to vessels previously developed by the taxpayer:


“[I]f the court did not make a line-by-line determination of those otherwise qualifying research expenditures that involved a process of experimentation, we do not understand how the court concluded that the 80% test was met.”[3]


    • “Shrink-Back Rule” – CIS chose an “all-or-nothing strategy” by claiming the overall project was the business component, and choosing not to apply the four-part test at the subcomponent level. Thus, the Tax Court was prevented from applying the Shrink-Back rule, resulting in the disqualification of all activities and related costs.


TCC Insights:

We consider this case to be significant. This is the second Tax Court win for the IRS in less than two years. See TCC’s article on Siemer Milling v. Commissioner.[4] While some commentators have rightly pointed out a memo opinion is technically not precedential, the vast majority of opinions are memo opinions and it is rare for a non-Tax Court opinion to reject the reasoning of a memo opinion.

POE Substantially-All Test

Substantiating this test has instantly become more challenging. In our experience, many studies do not employ a methodology that assesses how much time is spent on activities constituting a process of experimentation.

Even if records are available, it now becomes more challenging for taxpayers to reach the 80% threshold for qualification. While the R&D tax credit is specifically targeted to encourage innovation in product, process, and software development (i.e., research in the commercial field, not just in the lab or academic environment), if the credit is limited solely to projects where 80% of the activities involve experimentation, many commercial projects likely would not meet this threshold.

Excluding direct support and direct supervision activities from the numerator while including them in the denominator of the POE substantially-all test calculation simply ignores the commercial realties of the marketplace —  i.e., it is bad business for a for-profit company to have its product development team spend 80% of their time on experimentation.

Research in the commercial environment usually involve numerous activities outside of experimenting, and the core R&D group is often supported by other qualifying groups not involved in the actual experimentation. Examples include the activities of a model shop building prototypes at the direction of R&D personnel, analysts assisting programmers define requirements and testing new software applications, production workers supporting the new process development work of manufacturing engineers, etc. If these support activities are only included in the denominator, then reaching the 80% threshold will be challenging for any projects with significant involvement outside of a company’s core R&D departments.

While we understand the Court’s position and its statutory underpinnings, this result is not a foregone conclusion. And we certainly do not believe this is what Congress intended. The statutory language does not require “experimentation,” it simply requires the activities being performed constitute “elements of a process of experimentation.”

It would be reasonable to conclude that certain activities involved in the product development process, for example, activities geared at conceptualizing new product designs, and building and testing prototypes at the direction of engineers or programmers, constitute elements of a process of experimentation even if these groups are not conducting the actual experiments to resolve a technical uncertainty. This interpretation is consistent with the statute and regulations, and is certainly more aligned with Congressional intent. We hope the courts will have an opportunity to revisit this issue.

Substantiation of POE Substantially-All Test

The Tax Court stated, in dicta:

“[I]f the court did not make a line-by-line determination of those otherwise qualifying research expenditures that involved a process of experimentation, we do not understand how the court concluded that the 80% test was met.”[5]

This statement was not necessary to decide the case and does not establish new law.

Eligibility for the research credit is not intended to be contingent on meeting unreasonable recordkeeping requirements.[6]  There are no specific requirements with respect to the form or nature of the records or the time for creating the records. Cross referencing the general recordkeeping requirements under I.R.C. section 6001, the current regulations simply provide a taxpayer claiming research credits must retain records in a sufficiently usable form and detail to substantiate that the expenditures claimed are eligible for the credit.[7]  The court’s statement also seems to contradict  the well-established principle adopted by both Treasury and the IRS that in the absence of any R&D specific recordkeeping requirements that “…taxpayers must be provided reasonable flexibility in the manner in which they substantiate their research credits.”[8]

Thus, while the Court in Trinity may have adopted a standard too loose in applying the POE substantially-all test, the Tax Court’s statement above reflects an unduly high standard not supportable under the statute and regulations. Consequently, taxpayers who could not provide “line by line” records could still meet the POE substantially all requirement if their methodology for doing so was reasonable.

Pilot Model

Upon reading the evidence presented to support the two sample projects, one might question why this case was litigated, and certainly question why the taxpayer would agree to use these two projects as a representative sample of the others. Yet, in the taxpayer’s defense, the Trinity case did offer some hope when it concluded the substantially-all test was met with respect to two of the six sample ships, and allowed all the associated costs.

Since Trinity, TCC has noticed an increase in taxpayers claiming supplies as qualified research expenditures under the section 174 pilot model regulations. When Trinity was decided, TCC like some other commentators was skeptical of the Court’s analysis in allowing the two projects. Some of our concerns were eliminated when Treasury subsequently promulgated the section 174 pilot model regulations. Nonetheless, we believe the Tax Court’s pilot model analysis in Little Sandy Coal Co. is more aligned with section 174 and section 41, and we at TCC have developed a methodology for applying and substantiating any costs claimed under the pilot model regulations.

Learn more about the Research & Development Tax Credit (R&D).

[1] T.C. Memo. 2021-15, Dkt. No. 17431-17.
[2] 757 F.3d 400, 404-405 (5th Cir. 2014), aff’g 691 F. Supp. 2d 688 (N.D. Tex. 2010).
[3] Little Sandy Coal Company @ 28.
[4] Siemer Milling v. Commissioner, T.C. Memo 2019-37 (2019).
[5] Id.
[6] See  H. Conf. Rept. 106-478, at 132 (1999).
[7] Treas. Regs. section 1.41-4(d).
[8] REG-112991-01.

Learn more about the Research & Development Tax Credit (R&D Tax Credit).

Peter Mehta, J.D., LL.M, is TCC’s Managing Director and National Service Leader for our R&D Tax Credit practice. With over 20 years of tax experience with the Big 4 and a law firm, Peter has significant expertise with R&D tax credit matters and federal and state practice and procedure issues.

More Tax Credit Insights

Free assessment

Talk To a Tax Credit Expert Today

Connect with a partner relationship manager to explore our turn-key solutions and integrations.

Get a free assessment