Developing a consistent approach to the sec. 41 research credit.

The IRS issued two new research credit audit guides during 2005: one assists examiners in auditing research credits reported by aerospace companies, while the other assists them in applying the "process of experimentation" requirement to cases involving software development. These guidelines have implications beyond their intended scope. They recommend having examiners adopt positions on issues that relate to unsettled areas of the law. As such, they may foretell which positions the IRS will take in formal guidance it intends to issue during 2006. Moreover, they will give taxpayers a sense of the issues likely to be raised during an examination and the type of documentation that will be demanded by examiners to substantiate a credit. The practical implications of the shifts in policy reflected in these audit guidelines will be addressed here.

Aerospace Industry ATG 

The "Aerospace Industry-Audit Techniques Guide," dated January 2005, begins by asking a basic question:

Why is there a need for an Aerospace Research Credit Audit Technique Guide (ATG)? ***Because the ATG zeroes in on industry specific matters, it gives background and analysis in addressing research credit issues specific to the Aerospace industry and therefore provides direction to the Aerospace examiner not covered in other audit technique guides. ("Aerospace Industry--Audit Technique Guide--January 2005," p. 1, 2005 TNT 146-147)

While the ATG focuses on research credit issues prevalent in the aerospace industry, it covers many topics that have broad applicability to other business sectors. First, it addresses the Sec. 41(d)(4)(H) exclusion for "funded research." Research performed on behalf of another person is "funded," and thus ineligible for the credit, to the extent the contractor is entitled to payment regardless of whether the research is successful or whether the contractor does not retain substantial rights in the research; see Regs. Sec. 1.41-4A(d); Fairchild Industries, Inc., 71 F3d 868 (Fed. Cir. 1995); and Lockheed Martin Corporation, 210 F3d 1366 (Fed. Cir. 2000).

In Lockheed Martin, the court determined that a defense contractor had substantial rights to the research because the contractor was allowed, subject to certain restrictions, to use the research results to develop future products for sale or license to customers. In some cases, however, the contractor's rights to the research results were limited to future work performed for the same customer that originally paid for the research. The ATG notes that whether the contractor in this instance retains "substantial rights" is unresolved and currently being litigated in a Tax Court case. Also, according to the ATG, contracts often have different payment terms for each of the line-item deliverables due under the contract. Thus, it advises examiners that they need to resolve the "funding" issues based on the terms of each line item--which is consistent with the fact that contractors often perform a mix of funded and unfunded research activities under a single contract; see Regs. Sec. 1.41-4A(d)(3), providing that contracts can be partially funded.

The ATG further addresses the so-called "cascading credit" argument raised in Lockheed Martin, 49 Fed. C1. 241 (2001). The cascading credit potentially occurs when the prime contractor treats a prototype component developed by a subcontractor as a "supply" used in the contractor's research effort, while the subcontractor treats its development of the same component part as qualified research. In the case, the government argued that the prime contractor and subcontractor cannot both generate a credit for costs allocable to the same development activities. While this issue was raised in the litigation, the ATG correctly concludes, "[t]here is nothing ... specifically denying a prime contractor and subcontractor from both claiming the research credit with respect to the same item."

Second, the ATG advises examiners of "a trend to include a myriad of non-qualified costs in the research credit computation by claiming such costs are supplies." Of particular concern is whether taxpayers are claiming the costs incurred for constructing "property of a character subject to allowance for depreciation" (Sec. 41(b)(2)(C) and Regs. Sec. 1.174-2(b)). As noted in the ATG, this issue is particularly prevalent with respect to prototypes: "The design and testing of the prototypes may be considered qualified research, while the construction of the prototypes may be considered the acquisition of depreciable property." Whether prototype construction costs (i.e., costs of component materials and construction labor) qualify for the credit is tricky because most prototypes are durable and subject to wear and tear and, thus, have characteristics of depreciable property. However, there is a key difference: a prototype ordinarily is not developed to be sold or placed in service, but rather to be tested to determine if the design meets that taxpayer's basic functional and economic needs. The ATG recognizes this distinction, concluding that "[b]ecause the prototype was produced to be consumed in the research process, it is not considered property of a character subject to the allowance for depreciation." The ATG advises examiners that components that "are pre-sold while the item undergoes further testing" are "of a character which is subject to allowance for depreciation." Thus, examiners will consider any contracts for the sale of items developed as "prototypes" (even though not commercially available) in determining whether the costs incurred in their construction may be treated as qualified research expenses.

It is unclear whether the ATG is recommending that examiners limit allowable prototype construction costs to just the prototypes actually destroyed or substantially "consumed" during testing. At one point, the guide suggests that examiners should focus on whether the prototype, at the time it was constructed, was intended to be used only in the research project. For example, if 10 prototype milts are constructed solely for testing, but a component meets the taxpayer's basic functional and economic needs after testing only five units, the cost of constructing all 10 should qualify as a "supply" expense--regardless of whether the taxpayer scraps (or finds some other incidental use for) the five units that were never tested; the construction expenses for these 10 units were incurred "for supplies used in the conduct of qualified research" (Sec. 41(b)(2)(A)(ii)).Whether each prototype unit is actually destroyed or consumed in the research project should not control. This seems consistent with the IRS's long-standing position that the ultimate sale of a prototype unit does not change the character of the development expenses for Sec. 174 purposes; see Rev. Rul. 85-186 (no recapture of Sec. 174 deductions required when prototype is sold). The ATG may nonetheless suggest to some examiners that they should simply disallow prototype construction expenses in all cases in which a prototype was not actually destroyed during the research project. This may not be appropriate, however, when the baseline unit was created solely for testing, yet remains intact at the end of the testing process.

Although there are some concerns over how examiners will apply the ATG position on prototypes, it appears to reflect a useful shift in the IRS's position on this issue. Prior guidance suggested that the only relevant factor was the prototype's intrinsic nature (i.e., would it be "property of a character subject to allowance for depreciation" in anyone's hands), regardless of how it was used ultimately in the research project; see FSA 200013017 (12/23/99) (prototype computer chips) and TAM 199927001 (3/4/99) (plastic injection molds). In contrast, the ATG considers whether the prototype was constructed for use in the research effort as relevant in determining its "character" to the particular taxpayer. The ATG nonetheless notes that this area of the law is evolving and that formal guidance is expected on this issue.

Lastly, the ATG addresses the difficult issue of when a qualified research activity begins and ends. With respect to the former, the ATG suggests that, in developing a bid and proposal (B&P) for a government contract, most of the expenses incurred by the contractor are administrative and would not qualify for the credit. This is often not the case. In fact, initial conceptual development usually takes place prior to the submission of the B&R Additionally, significant technical development (often pure laboratory work) is necessary to respond to follow-up questions relating to the Request for Proposal. Indeed, according to Regs. Sec. 1.41-4A(d)(4), research costs are incurred to prepare the B&R as this provision addresses when such costs should be treated as "funded" and, thus, disallowed on this separate ground.

The ATG also indicates that research undertaken to develop a new aircraft is generally completed prior to FAA approval. It notes that the FAA "'tests' are done to prove and substantiate that the design meets certain government standards and criteria," which occurs after the research is completed. This position suggests that the research is completed before the aircraft "is ready for commercial sale or use," which is the general standard for determining when research is completed; see Regs. Sec. 1.41-4(c)(2).This position also conflicts with the regulatory provision providing that new drug development activities continue until completing the clinical testing needed to meet the Food and Drug Administration's requirements; see Regs. Sec. 1.41-4(c)(2)(iv). Moreover, if an aircraft's design fails to meet FAA requirements, it may be scrapped and, under the rationale of the "supply" cost section of the ATG, the related construction costs would qualify for the credit because they were "consumed" in the research effort. Thus, the IRS's reasoning suggests a difference in how to treat successful and unsuccessful research. Yet, under Regs. Sec. 1.41-4(a)(3)(ii), the outcome of the research effort is irrelevant in determining the qualified research scope.

The aerospace industry ACT is available online at www.irs.gov/ businesses/corporations/article/0,,id= 137957,00.html.

Software

For almost two decades, a hotly debated issue has been whether some taxpayers are claiming research credits for a broader range of software development activities than Congress intended. This is evident from the number of cases the IRS has litigated, as well as the numerous policies instituted by the Examinations Division to address this issue. Initially, the IRS focus was on internal use software--which is generally ineligible for the credit unless "highly innovative"; see Sec. 41(d)(4)(E) and Advance Notice of Proposed Rulemaking, REG-153656-03 (issued 12/23/03).

The IRS recently issued "Audit Guidelines on the Application of the Process of Experimentation for all Software"; see www.irs.gov/businesses/ article/0,,id=138989,00.html. The audit guideline (AG) reflects broader concerns over research credits claimed for all types of software development, including the development of new software products for sale, as well as products with embedded software.

The purpose of the AG is to target audit resources on the issues most likely to result in adjustments. The AG attempts to accomplish this in several ways. First, it breaks down software development into two general categories: (1) activities occurring during the lifecycle of software development (i.e., configuration, maintenance, etc.) and (2) typical software development projects (e.g., new applications, system software, etc.). It then analyzes the level of qualified research typically undertaken in these types of activities/projects, by using a three-tier approach to categorize software development:

* High risk: these activities usually fail to constitute qualified research under Sec. 41 (d).

* Moderate risk: these activities often fail to constitute qualified research under Sec. 41(d).

* Low risk: these activities rarely fail to constitute qualified research under Sec. 41 (d).

This taxonomy is supposed to help examiners focus audit resources on high-risk issues and avoid burdening taxpayers with unwarranted issues that are unlikely to result in an adjustment. The problem is that the AG suggests that only unique or novel software development will qualify for the credit. This follows from the Fact that almost all of the software that would fit within the AG's low- or medium-risk category would be highly innovative. Yet, Congress intended the so-called "high-threshold of innovation standard" to be applied only to internal-use software; see H Rep't 841, at II-73 (reprinted in the Cumulative Bulletin, 1986-3 (Vol. 4) 1, 73). Thus, examiners may end up focusing on whether the commercial software developed by the taxpayer has innovative features, in determining whether there is compliance with the Sec. 41(d)(1)(C) "process of experimentation" requirement, despite the fact that Congress intended software offered for sale not to be subject to a special "innovativeness" requirement.

The AG also directs examiners to focus on whether taxpayers are adequately substantiating a qualifying "process of experimentation." It indicates that taxpayers often presume that evidence of the technical risk addressed through development efforts is sufficient proof of a process of experimentation. Moreover, the AG notes that some taxpayers simply rely on the fact that software development is an inherently experimental process--which is consistent with the IRS's historic stance on this issue; see Rev. Procs. 69-21 and 2000-50. Nevertheless, the AG specifies that taxpayers must substantiate the steps taken in their research effort, to prove that they engaged in a qualifying process of experimentation.

This feature is important because it may prompt examiners to ask for substantiation that had not been requested in prior examination cycles. This is particularly troublesome because there is no business purpose for creating the very documentation that the AG is encouraging examiners to request. Software developers do not generally keep notebooks documenting the methodology and techniques they use at each stage of the development process. As a result, to document the steps taken in the experimentation process, taxpayers have to interview developers. With the high turnover of software developers in many companies, this may prove to be a challenge when the examiner is asking for details about software development that occurred several years ago.

Third, the AG is intended to be an educational tool providing examiners with a sense of the lifecycle of software development, as well as the activities typically undertaken during each phase. It also describes the methodologies (e.g., waterfall, iterative, rapid application development) typically applied by software developers. The problem, however, is that it does not provide examiners with any sense of the process (steps) generally followed in a software development project that involves a qualifying process of experimentation.

It is simply unclear whether the AG is suggesting that all software development has experimental elements, but does not always involve a qualified process of experimentation, because it does not necessarily address complex or "cutting-edge" computing issues. While one part of the AG directs examiners to focus on whether taxpayers have substantiated the steps taken to develop software, it potentially could be interpreted (in practice) as requiring the taxpayer to show that its software development techniques or applications were unique or novel relative to other software products commercially available at that time. Such an interpretation would not, however, be consistent with Sec. 41's statutory scheme, regulations or legislative history.

Conclusion

Although the IRS is attempting to promote consistent audit practices, in the research credit context this is particularly difficult because industrial research and development is so varied. The IRS is nonetheless directly confronting this challenge by issuing audit guidelines and it deserves to be commended. The aerospace ATG addresses several technical issues that have troubled examiners and taxpayers, particularly whether costs incurred in constructing prototypes are qualified research expenses. The software AG, on the other hand, focuses on identifying the type of software development activities that warrant a thorough audit. It also directs examiners to demand proof of the steps taken that constitute a qualifying "process of experimentation." Unfortunately, the AG is not sufficiently clear about how to distinguish a qualified process of experimentation in the software development context from other routine software-related activities. This could result in more expansive audits, which is the opposite of what the AG is supposed to accomplish. Ultimately, whether these guidelines represent a step in the right direction will depend on how they are applied by IRS examiners.

MICHAEL GOLDBAS, J.D., HARTFORD, CT