R&D Tax Credit Update: Recent Court Decisions Shed Light on Key Issues

By Joe Stoddard, CPA, Salt Lake City; Ryan Coleman, MBA, J.D., LL.M., Philadelphia; and Tim Rankins, J.D., LL.M., Philadelphia

Editor: Greg A. Fairbanks, J.D., LL.M.

Credits Against Tax

The research and development (R&D) tax credit under Sec. 41 continues to be an important incentive for many companies; however, the incentive effect is often dampened by challenges in identifying, documenting, and defending the credit. These challenges are heightened by the subjective nature of the terms defining qualified research and related expenses, the lack of guidance on what constitutes proper substantiation, and contentious IRS examinations. Court decisions from recent months provide guidance on several important R&D tax credit issues including the qualification of certain supply expenses, proper substantiation and documentation, and the definition of gross receipts. This item explores these cases and the potential impacts on companies claiming the R&D tax credit.

Note: The R&D credit had expired as of Dec. 31, 2011, but the American Taxpayer Relief Act, H.R. 8, which was signed on Jan. 2, extended it through 2013 and introduced modifications that are too extensive to go into for the purposes of this item.

Union Carbide Corp.: Disallowance of Supply Costs

The Second Circuit affirmed the Tax Court’s decision in Union Carbide Corp., No. 11-2552 (2d Cir. 9/7/12), which disallowed certain supply costs used in process research activities as qualified research expenses (QREs) for purposes of the R&D tax credit. The Second Circuit’s decision is a blow to the manufacturing sector, which had hoped for a reversal of the original Tax Court ruling (Union Carbide, T.C. Memo. 2009-50) to allow a broader scope of process research costs to qualify for the R&D tax credit.

Overview: After negotiating an agreement with the IRS on R&D tax credits of $14 million and $4 million on its 1994 and 1995 tax returns, respectively, Union Carbide claimed additional credits arising from 106 projects conducted in the taxpayer’s manufacturing facilities during 1994 and 1995. The IRS contested these additional credits, spurring the original Tax Court litigation. The Tax Court analyzed the five largest projects to determine the eligibility of the additional claimed projects and costs.

The Tax Court held that two of the five projects constituted qualified research activities, but it disallowed all of the additional supply costs because they were incurred in producing goods for sale, not in conducting qualified research. In doing so, the Tax Court created a “primary purpose” test, where even if supply costs are essential for process research, they can never be qualified as process research expenses if they are incurred primarily for the production of products. As a result, Union Carbide was entitled to only $1,045 of additional QREs of the more than $200 million in costs at issue.

The appeal: The appeal focused primarily on whether the supply costs had been properly disallowed by the Tax Court. Union Carbide, now owned by Dow Chemical Co., argued that the plain language of Sec. 41 allows the credit for “any amount paid or incurred for supplies used in the conduct of qualified research” and that the production supplies used in process research met the requirement because they were used in the conduct of qualified process improvement research activities. The IRS argued that the “[s]upply costs are ‘indirect research expenditures’ if they would have been incurred regardless of any research activities,” even though it conceded that the production process research could not have been conducted without buying those supplies.

Union Carbide pointed to the plain meaning of the word “use” from Webster’s Third New International Dictionary to argue that “used in the conduct of qualified research” encompassed costs of all supplies regardless of whether the supplies were used to make a product for sale. The court disagreed, focusing instead on the whole phrase “used in the conduct of qualified research.” The court pointed to the title of Sec. 41, “Credit for Increasing Research Activities,” (emphasis added) which it said suggests that supplies purchased in the ordinary process of producing goods are not allowable for credits.

The Second Circuit agreed with both the IRS and the Tax Court, stating that the costs of supplies were “at best, indirect research costs” and not allowable under Regs. Sec. 1.41-2(b)(2). In doing so, the court agreed that the interpretation of such supplies as indirect costs was consistent with the purpose of the R&D tax credit. Further, in interpreting congressional intent, the Second Circuit found that “[a]ffording a credit for the costs of supplies that the taxpayer would have incurred regardless of any qualified research it was conducting simply creates an unintended windfall.” In a concurring opinion, Judge Rosemary Pooler opined that Congress may well have intended for these types of process supplies to qualify for the credit but “failed to write the statute in such precise terms.”

Impact: Manufacturing companies involved in process research activities will be directly affected by the Second Circuit’s decision because they are most likely to incur supply costs used in process research activities. In fact, according to the most recently published IRS R&D tax credit data (see “Table 1, Corporations Claiming a Credit for Increasing Research Activities on Form 6765”) manufacturing companies in 2009 claimed approximately 70% of the total dollar amount of R&D tax credits claimed, while qualified supplies accounted for approximately 20% of the total QREs for manufacturing companies.

In an amicus curiae brief, the National Association of Manufacturers, the American Chemistry Council, and the U.S. Chamber of Commerce argued that the Tax Court’s interpretation of the process research supply issue would stifle process research and that “Congress has not created distinctions between product and process research for purposes of which costs constitute QREs, and the courts should not insert them.” Further, the brief argued that “[i]f no credit is available for one of the most important cost components of process research, logic dictates that less process research will occur,” and U.S. manufacturing firms will fall behind foreign competitors in manufacturing process innovation.

Unfortunately, based on the Second Circuit’s decision, any supply costs claimed for process research that would have been incurred absent the research will likely be disallowed by the IRS during R&D tax credit examinations. Taxpayers claiming supply costs for process research should carefully document how the supplies were used in the process research and be prepared to demonstrate that the primary purpose of the costs was for the research activity. Sustaining qualification of process supplies as QREs will likely require detailed documentation of the incremental costs incurred beyond regular production costs.

Taxpayers should note that the Tax Court generally accepted the validity of employee recollections and data extrapolation to estimate QREs. Despite clear statements in the legislative history that eligibility for the R&D tax credit should not depend on meeting unreasonable recordkeeping requirements, the IRS has actively opposed the use of employee recollections and extrapolations from existing data to reasonably estimate the portion of expenses that may be properly classified as QREs. For projects that otherwise satisfy the qualification requirements, however, the Union Carbide decision generally rejects the IRS’s arguments and accepts the validity of such evidence.

Note: Union Carbide filed a petition for certiorari with the U.S. Supreme Court on Dec. 4, 2012, requesting the Court review the Circuit Court’s decision. The government’s response to the petition was due Feb. 4, 2013, but had not been submitted as of this writing.

Bayer Corp.: Statistical Sampling and Substantial Variance

Two rulings by the U.S. District Court for the Western District of Pennsylvania in the ongoing litigation of Bayer Corp., No. 09-cv-00351 (W.D. Pa. 2/6/12 and 9/20/12), underscore the importance of tying research expenses used to qualify for the R&D tax credit to identifiable “business components.” On Feb. 6, 2012, the court denied Bayer’s motion to use statistical sampling to assess the accuracy of the company’s R&D tax credit claim, holding that the identification of business components is a critical step in substantiating the credits. In the second ruling, on Sept. 20, 2012, the court denied the IRS’s motion for partial summary judgment because it disagreed with the IRS’s arguments that Bayer had substantially varied the factual basis for the R&D tax credits at issue.

Overview: Bayer claimed approximately $175 million of R&D tax credits from 1990 to 2006. To compute its credit, Bayer used an accounting system that was based on the nature of the activities performed rather than individual projects. Similar activities (e.g., clinical studies for pharmaceuticals or formulation development for crop protection agents) were grouped into more than 1,300 cost centers for which eligible QREs were compiled for the credit.

Sec. 41(d)(2) provides that the test for determining whether an expense was incurred in connection with qualified research is applied separately to each business component of the taxpayer. A business component is defined as “any product, process, computer software, technique, formula, or invention” that is either (1) held for sale, lease, or license or (2) used by the taxpayer in its trade or business (Sec. 41(d)(2)(B)). Bayer estimated it had developed more than 100,000 business components during the years at issue. While Bayer provided extensive information on its products, processes, formulas, etc., the company made no attempt to segregate costs by business component, which it argued its books and records do not, and are not required to, track individually.

In preparing its R&D tax credit study for some of the years at issue, Bayer delivered a draft of the report to the IRS to provide feedback before finalizing it. In subsequent meetings, the IRS asked numerous questions but did not object to the methodology used and did not request that additional types of information be collected. During the detailed audit that preceded this litigation, the IRS never requested Bayer to identify or list the business components to which the claimed QREs related.

Statistical sampling motion: During the case’s discovery phase, the IRS issued an interrogatory requesting Bayer to identify and describe “each new or improved business component Bayer contends it incurred qualified research expenses to develop during the credit years.” Bayer argued that the request was “overbroad and unduly burdensome without the adoption of a suitable sampling method.” The vast scope of the enterprise was demonstrated by the fact that Bayer had collected more than 1 billion pages of potentially relevant electronic records from just four of its 49 sites at issue. In addition, more than 3 million pages of documents had already been provided to the IRS.

In response, Bayer filed a motion with the court requesting a discovery plan based on statistical sampling to examine a subset of the cost centers and research activities conducted during the credit years. The results of the sample would then be extrapolated to all of Bayer’s claims without requiring it to introduce further evidence at trial. The IRS did not want to be bound by Bayer’s sampling approach and “vigorously opposed” being forced to accept statistical sampling during discovery or at trial.

Both Bayer and the IRS hired statistical sampling experts to work out an acceptable sampling plan to streamline the discovery process. Bayer proposed selecting 50 cost centers from a sample of eight research sites to examine for each of two years (i.e., for a total of 100 sample units). The IRS’s expert proposed using a pilot sample that he described as a “dress rehearsal” for a full sampling plan that would facilitate a settlement of this dispute. Bayer and the IRS could not agree on a sampling approach.

The court denied Bayer’s motion, noting that it would eliminate Bayer’s burden of proof regarding all QREs claimed at 41 of the 49 research sites during the credit years, as well as the QREs not selected for analysis in the eight sample research sites. The court stated that it could “find no authority for the extraordinary relief sought by Bayer” and further stated that such relief would “constitute a reward to Bayer for failing to keep evidence regarding research expenses in ‘sufficiently usable form and detail’” (quoting Regs. Sec. 1.41-4(d)).

Motion for partial summary judgment: After the sampling motion was denied, the court directed Bayer to fully respond to a similar but much narrower interrogatory to provide information on the business components developed during the years at issue. The IRS then filed the motion for partial summary judgment on the grounds that in gathering and providing information to respond to the interrogatory, Bayer was substantially varying the factual basis for its refund claim (citing Lockheed Martin Corp., 210 F.3d 1366 (Fed. Cir. 2000)).

The “substantial variance” rule bars a taxpayer from presenting claims in a tax refund suit that substantially vary the legal theories or factual bases set forth in the refund claim. One purpose of the substantial variance rule is to limit litigation to those grounds that the IRS had an opportunity to consider and is willing to defend. In siding with Bayer, the court ruled that

Bayer has been compelled to engage in Herculean efforts to comply with the Government’s demand in this litigation for a list of the business components to which the claimed QRE credits relate. . . . [T]he Government will not be permitted to demand a list of business components for the first time in this Court and then object based on the substantial variance rule to Bayer’s need to gather significant, additional evidence to comply with the demand when the QRE credits underlying Bayer’s refund claim have not changed. [ Bayer , 9/20/12 slip op. at 16.]

Impact: The Bayer statistical sample ruling highlights the need for taxpayers to identify business components and produce documentation that creates nexus between the business components and qualified costs. In denying the IRS’s motion for partial summary judgment, the court acknowledged the challenges that taxpayers face in producing business component listings and related documentation to meet IRS demands. The decision further demonstrates the IRS’s shifting focus to business-component-based (i.e., project-based) R&D documentation.

Shami: Substantiation of Eligible Expenses

A recent Tax Court decision, Shami , T.C. Memo. 2012-78, provides another example of the importance of retaining sufficient records to substantiate that expenses are eligible for the R&D tax credit, as required under Regs. Sec. 1.41-4(d).

Overview: Farouk Systems Inc. (Farouk) developed hair, skin, and nail products. Farouk performed product development activities across several departments for which it claimed R&D tax credits for 2003, 2004, and 2005. At the direction of the court, the parties agreed to limit the focus to the two top executives, founder and CEO Farouk Shami and Executive Vice President John McCall. At issue was whether their wages were eligible to be included as QREs.

For 2003 and 2004, Farouk claimed that both Shami and McCall spent 80% of their time performing qualified services, and for 2005 that Shami spent 35% of his time performing qualified services. Shami’s wages were approximately $9 million and $8 million for 2003 and 2004, respectively, while McCall earned approximately $6 million in 2003 and $2 million in 2004. Neither Shami nor McCall had any formal education or training in science or engineering.

Decision: The IRS contended, and the court agreed, that Farouk failed to adequately substantiate the wage allocations of both Shami and McCall. The court cited Tax Court Rule 142(a) for the proposition that “[t]he Commissioner’s determinations in a deficiency notice are presumed correct, and taxpayers bear the burden of proving otherwise” (Shami at *5). Farouk relied solely on oral testimony to substantiate the amount of time spent by both Shami and McCall on qualified services. The court said, “Petitioners failed to provide any documentation that establishes how much time, if any, Mr. Shami or Mr. McCall spent performing research and development services” (id. at *9). The court also noted that several witnesses contradicted Shami’s testimony, and no witnesses corroborated McCall’s testimony. The court found the oral testimony of both Shami and McCall to be self-serving and unreliable. Further, the court held that the inadequate substantiation prevented any amount of their wages from qualifying for the R&D tax credit, as Farouk failed to provide documentation substantiating how much time, if any, the two employees spent performing qualified activities.

Farouk contended that the Cohan rule (Cohan, 39 F.2d 540 (2d Cir. 1930)) should permit the court to estimate the appropriate allocation of wages between qualified and nonqualified services. The Cohan rule has been used in other R&D tax credit cases to estimate QREs in the absence of documentation of the exact amount of expenses. The court held that since Farouk provided no documentation to establish how much time, if any, these two executives spent on qualified research, it was not able to apply the Cohan rule. The court noted that “[f]or the Cohan rule to apply . . . a reasonable basis must exist on which this Court can make an estimate. . . . Petitioners failed to satisfy the Court that there is sufficient evidence to estimate the appropriate allocation of wages between qualified services and nonqualified services for Mr. Shami and Mr. McCall” (Shami at *10).

Impact: The Tax Court’s disallowance of the wages as qualified research expenses reinforces the need for taxpayers to adequately document the extent of their employees’ involvement in qualified services for purposes of the R&D tax credit. Without documentation supporting the nature of the claimed activities as well as the amount of time incurred, estimates will not be permitted due to the lack of a reasonable basis for making the estimate.

Davenport: Process of Experimentation Requirement

A recent decision by the U.S. District Court for the Northern District of Texas, Davenport , No. 3:09-cv-02455-L (N.D. Tex. 9/14/12), held that costs of modifying a commercially available software package did not qualify for the R&D tax credit.

Overview: Members of the Davenport family owned Burly Corp., which manufactured metal roofing and stand-alone metal buildings through its subsidiary, Mueller Supply Co. In 2006, the Davenports filed refund claims for the 2002 and 2003 tax years reflecting R&D credits for expenses incurred in developing its OneWorld software system. The system managed, automated, and integrated all aspects of Mueller’s business, including automation and integration of manufacturing, design, sales, accounting, and shipping. The system was created using the commercially available J.D. Edwards OneWorld enterprise resource planning (ERP) software application suite. Mueller also purchased a license from IBM for an ERP bridge, which consists of base code that is used as a starting point, and contracted with IBM and other contractors in modifying the software for its specific needs. The IRS sought the return of $292,096 in what it claimed were erroneously issued refund checks to the Davenports for the 2003 tax year, while the Davenports claimed the IRS had wrongfully disallowed its R&D tax credits for the 2002 tax year for refunds totaling $196,898.

The IRS took a “kitchen sink” approach to disqualifying the Mueller OneWorld project by arguing the activities (1) did not satisfy the Sec. 174, technological information, or process-of-experimentation tests; (2) fell within the exclusion for internal-use software; and (3) were specifically excluded as research done after commercial production, adaptation, or duplication of existing business components and as routine testing or inspection for quality control. Further, the government contended that all contract research QREs claimed were ineligible because the Davenports failed to produce any supporting agreements. The government also argued the Davenports used the wrong fixed-base percentage to calculate the credit, and under the correct fixed-base percentage, they would not be entitled to a credit. The Davenports disputed all of the government’s claims.

In granting the government’s motion for summary judgment, the court focused solely on whether a process of experimentation occurred.

Decision: In its decision, the court cited Union Carbide repeatedly in explaining that to constitute qualified research, a valid process of experimentation must have a result that was uncertain at the beginning of the research. The Davenports acknowledged the definition of a process of experimentation includes evaluating alternatives via a scientific method, where a hypothesis is developed and tested in a scientific manner, results are analyzed, and the hypothesis is either discarded or refined. The Davenports argued the government’s contention that no process of experimentation occurred was conclusory and factually inaccurate and oversimplified the efforts undertaken. In doing so, the Davenports relied on the holding in Trinity Industries, Inc., 691 F. Supp. 2d 688 (N.D. Tex. 2010), to argue there was a genuine dispute of material facts.

The court disagreed and held that the evidence established that the testing did not rise to the definition of a process of experimentation as stated in Union Carbide and acknowledged by the Davenports. The court stated the testing was not performed to test and refine a hypothesis to evaluate the strengths and weaknesses or to determine other alternatives in realizing the Davenports’ requirements. The court noted that integration and script testing was performed after IBM developed or customized the Mueller OneWorld system rather than at the outset of the project.

In reaching this finding, the court relied heavily on two IBM proposals submitted by the government. The proposals discussed the phases of the project performed: preparation, redesign/design, configuration, and deployment. The court highlighted that if the system had not operated as desired, IBM would have created a “fix” to address the issue or confirm the system was working properly, and Mueller would have needed to adjust its business methods accordingly. Among the reasons the project did not satisfy the process-of-experimentation test was that testing was done merely to confirm or validate that the system operated as intended, rather than to identify design alternatives, the court said.

Impact: The court’s ruling highlights the need for companies claiming the R&D tax credit to pay particular attention to documenting how the activities claimed as qualified research constitute a valid process of experimentation and gathering evidence. It further highlights the need to evaluate whether design alternatives are being considered or if the activities simply involve configuration, testing, and validation, where no uncertainty exists at the outset of the project.

Hewlett-Packard Co.: Gross Receipts

The Tax Court granted the IRS’s motion for partial summary judgment in ruling that Hewlett-Packard (HP) was required to include nonsales income (i.e., dividends, interest, rents, and other income) in its R&D tax credit computations (Hewlett-Packard Co., 139 T.C. No. 8 (2012)). Additionally, the IRS conceded to, and the court granted, HP’s motion for partial summary judgment allowing it to exclude intercompany gross receipts received from controlled foreign corporations (CFCs) for purposes of computing its R&D tax credit claims.

Overview: During 1999–2001, HP computed its R&D tax credit using the alternative incremental research credit (AIRC) method. The AIRC is based on statutorily defined percentages of a taxpayer’s average annual gross receipts compared with its QREs. On the original returns for the years at issue, HP included only sales receipts, less returns and allowances, from Form 1120, U.S. Corporation Income Tax Return , line 1(c) in its R&D tax credit calculations. In 2003, HP filed amended returns for 1999–2001 to further reduce its gross receipts figures by intercompany receipts received from its CFCs.

The IRS issued final regulations in 2001 (T.D. 8930) clarifying that calculations of the R&D tax credit require a broad definition of the term “gross receipts” (see Regs. Sec. 1.41-3(c)). However, the regulations are applicable only to tax years beginning after Jan. 3, 2001. HP argued that the IRS’s position was an impermissible retroactive application of the regulations. The IRS contended that Treasury’s embracing of a broad definition of gross receipts equally applied to pre-effective-date tax years.

Decision: The court agreed with the IRS and noted that “[n]owhere in the Code has the isolated term ‘gross receipts’ been construed as narrowly as HP suggests. On the contrary, an examination of the Federal income tax laws reveals that Congress widely embraces the notion of a broad, inclusive definition for the term” (Hewlett-Packard, slip op. at 17–18).

Also at issue was HP’s motion for partial summary judgment to exclude from gross income for calculating the credit its intercompany gross receipts from its CFCs. This issue was addressed in 2010 in Procter & Gamble Co. , 733 F. Supp. 2d 857 (S.D. Ohio 2010). The only issue addressed in that case was “whether a taxpayer must include the results of its intercompany transactions within its ‘Gross Receipts’ for the purposes of determining the amount of its research credit” under Sec. 41.

During the IRS’s audit of Procter & Gamble’s (P&G’s) 2001–2005 tax years, the IRS Office of Chief Counsel altered its position regarding the treatment of intercompany sales as gross receipts with respect to foreign subsidiaries. (See Chief Counsel Advice (CCA) 200233011, which authorized excluding intercompany sales to CFCs from gross receipts, and CCA 200620023, in which the Chief Counsel reversed its position with respect to intercompany transfers with CFCs.) In accordance with the new position, the IRS team auditing P&G reversed its position to no longer allow P&G to exclude its intercompany sales to its CFCs in computing its R&D tax credit claims for the years at issue.

The court relied heavily on both the plain language of the statutes and regulations and the legislative history of the statutes in holding that P&G correctly excluded receipts from its intercompany sales to its CFCs in computing the base amount and fixed-base percentage portion of the R&D tax credit computation. Sec. 41(f)(1)(A)(i) states that “all members of the same controlled group of corporations shall be treated as a single taxpayer” for the purpose of computing the R&D tax credit. This statute therefore requires all intercompany transactions to be disregarded for R&D tax credit purposes, the court said. The IRS unsuccessfully argued that this provision applies only to QREs and not to gross receipts. Further, the IRS attempted to draw a distinction between domestic transfers and foreign transfers. The court rejected both arguments and held that they were without merit as a matter of law.

Therefore, in Hewlett-Packard, the IRS did not object to HP’s motion for partial summary judgment to exclude intercompany gross receipts from its CFCs, and summary judgment was granted to HP on this issue as well.

Impact: The Tax Court’s decision reinforcing the broad definition of gross receipts should have a limited impact on taxpayers due to the final regulations that have been in effect since 2001. Further, it appears the IRS will no longer contest the exclusion of intercompany gross receipts from CFCs in computing the base amount.


While the R&D tax credit continues to be a contentious issue between taxpayers and the IRS, these recent court decisions provide guidance on subjective issues for companies wrestling with the amount and type of substantiation required to support their R&D tax credit claims. Taxpayers should consult with their tax adviser to consider the implications of these cases in identifying, computing, and documenting their R&D tax credit claims.


Greg Fairbanks is a tax senior manager with Grant Thornton LLP in Washington, D.C.

For additional information about these items, contact Mr. Fairbanks at 202-521-1503 or greg.fairbanks@us.gt.com.

Unless otherwise noted, contributors are members of or associated with Grant Thornton LLP.

Tax Insider Articles


Business meal deductions after the TCJA

This article discusses the history of the deduction of business meal expenses and the new rules under the TCJA and the regulations and provides a framework for documenting and substantiating the deduction.


Quirks spurred by COVID-19 tax relief

This article discusses some procedural and administrative quirks that have emerged with the new tax legislative, regulatory, and procedural guidance related to COVID-19.