The Statutory Framework of IRC Section 41 and the Research Credit

The R&D tax credit provides a credit against tax for qualifying research costs, subject to calculation and utilization limits. Qualified research must satisfy cumulative requirements applied to each business component. Failure at the overall component level does not necessarily dispose of every smaller component; the shrinking-back rule may require further analysis.

The Four-Part Test for Qualified Research

The following table summarizes the requirements. The Section 174 reference describes the law applicable to the PDG research years; following the 2025 legislation, Section 41(d)(1)(A) refers to Section 174A. Taxpayers must retain records sufficient to substantiate eligibility and expenses, but the rules do not prescribe a universal project diary or mandatory time-tracking software.

The substantially-all threshold is 80%, measured by cost or another consistently applied reasonable basis for the relevant research activities. It is not a blanket requirement that 80% of an entire firm’s work be experimental. The shrinking-back rule applies the tests to successively smaller significant subsets of a business component when necessary.

Quantitative Calculation Methodologies

Qualified research expenses (QREs) generally include qualifying wages, supplies, and an eligible portion of contract research costs, with additional statutory rules for each category. Taxpayers may use the regular credit or elect the alternative simplified credit (ASC), subject to the applicable rules. The ASC requires less remote historical information than the regular credit, but current and preceding-year QREs still need support.

Where the taxpayer has QREs in each of the three preceding tax years, the basic ASC calculation is:

ASC = 14% × the excess, if any, of current-year QREs over 50% of average QREs for the three preceding tax years.

If the taxpayer has no QREs in any one of the three preceding tax years, the alternative rate is 6% of current-year QREs. This rule is not limited to newly formed businesses. These are basic rates before any reduced-credit election under Section 280C and other applicable adjustments or limitations. Expense classification and wage allocations must be defensible.

Detailed Case Analysis: Phoenix Design Group, Inc. v. Commissioner

PDG was a multidisciplinary engineering consulting firm designing mechanical, electrical, plumbing, and fire protection (MEPF) systems, including systems for hospitals and university facilities. Its research credit study identified more than 200 potentially qualifying projects. The parties tried three projects, but the court expressly stated that its findings were not binding on the remaining projects. It expected the decision to help the parties resolve those other claims.

Sampled Project One: Gerald Champion Military Psychiatric Unit

Project No. 13008.00 involved MEPF design for a medical facility. PDG pointed to design complexities and revisions as evidence of uncertainty. The court required an explanation connecting the work to technical uncertainty and a qualifying evaluative process. Revised drawings alone did not establish what the engineers investigated or how they evaluated alternatives.

A design can remain changeable without all of its elements remaining technically uncertain. Client changes, coordination, or code requirements may explain revisions but do not by themselves establish research. The question is what particular uncertainty existed and what activities PDG performed to resolve it.

Sampled Project Two: Vanderbilt University Engineering and Science Building (VU ESB)

The Vanderbilt project involved technically demanding building systems. PDG sought qualification at both the overall system and smaller-component levels. The court found that the evidence did not adequately explain the investigative activities or experimentation underlying the solutions. Technical difficulty and multiple drawings did not fill that evidentiary gap.

Using established engineering principles or standard software is not itself disqualifying. The regulations permit reliance on existing technology and do not require discovery beyond the common knowledge of skilled professionals. The problem is a failure to establish qualifying uncertainty and an evaluative process, rather than the mere use of familiar tools. Information supplied by clients or other contractors also must be distinguished from research actually performed by the taxpayer.

Sampled Project Three: BHNM Oxford Project

The BHNM Oxford project likewise exposed weaknesses in linking project records to claimed research. Broad time descriptions and explanations identifying a problem and eventual solution did not sufficiently explain the intervening research activities. PDG needed evidence showing which systems or subcomponents were involved and how the claimed activities met the statutory tests.

This is not a finding that a particular timesheet label automatically defeats a claim. Generic entries can be difficult to use when the rest of the evidence does not establish the nature and allocation of the work. Consistent technical records and credible explanations are therefore important.

The Judicial Rejection of the AIA Design Process as Automatic R&D Qualification

PDG relied on a six-stage design process associated with standard architectural and engineering project delivery. It argued that design refinement and revisiting earlier stages resembled experimentation. The court rejected the inference that placing activities inside such a workflow was sufficient proof of qualified research.

The practical distinction is between describing project administration and proving a technical evaluation. Modeling, simulation, or systematic trial and error may support qualification when tied to uncertainty and a permitted purpose. Neither the phase name nor a formal scientific-method label decides eligibility.

This table applies the opinion’s reasoning to the phase descriptions; it is not a categorical judicial ruling that every activity within a listed phase is excluded.

The Substantially-All Rule and the Shrinking-Back Failure

A taxpayer must support the substantially-all calculation for each business component. General statements that a project involved innovation do not establish the required share of experimental activities. Evidence should permit a reviewer to distinguish qualifying work from ordinary design, drafting, coordination, and other activities.

In PDG, the court did apply shrinking back to mechanical, electrical, and plumbing systems. It rejected shrinking back by employee or design phase because those categories did not identify the required subsets of the business component. Inadequate evidence limited further analysis. The opinion therefore should not be described as refusing to apply the rule altogether.

Records organized by meaningful system or subcomponent can help support this analysis. Separate time codes may be useful, but they are not the only permissible evidence. A defensible allocation must connect the underlying work, the relevant component, and the claimed costs.

Comparative Analysis with Funded Research Cases: Smith and System Technologies

The funded-research exclusion under Section 41(d)(4)(H) raises a separate issue from the nature of the activities. Under Treasury Regulation § 1.41-4A(d), contractual payment risk and substantial rights in the research results are central. Research may be funded to the extent of payments; partial eligibility can remain in appropriate circumstances. Describing every third-party payment as a complete disqualification is too broad.

The early summary-judgment proceedings in Smith and System Technologies were limited procedural decisions, rather than final awards of all claimed credits. In System Technologies, the January 3, 2025 order considered Indiana law and the obligation to refund payments if the promised product was not delivered. That supported denial of the IRS motion on funding.

Smith subsequently produced a merits opinion, T.C. Memo. 2026-50, on June 16, 2026. The court found that payments under the six contracts were not contingent on successful research. Substantial rights remained under four contracts, leaving possible partial credits under the regulation; the precise amounts, if any, were not determined. The court also accepted the total 2008 partner compensation as reasonable. The earlier procedural victory therefore cannot be presented as an unqualified final taxpayer win.

Contractual eligibility and technical eligibility require separate support. A favorable funding analysis does not establish the four-part test, and technically qualifying activities do not resolve funding restrictions.

Implications for Future R&D Tax Credit Applications

PDG reinforces the importance of connecting technical evidence with the tax claim. Administrative filing requirements are a separate consideration and should not be attributed to the opinion without support.

IRS Review of Research Credit Refund Claims

IRS guidance requires specified information for research credit refund claims on amended returns, including business components, activities and expense information. Since June 18, 2024, the IRS has waived the filing-stage requirement to identify the individuals performing each activity and the information each individual sought to discover; those details may still be requested in examination. The applicable instructions should be checked for the claim being submitted.

The IRS describes review by campus personnel, examiners, subject-matter experts, and counsel. The source document’s assertion that a new automated Classifier rejects claims without human review is unsupported. Filing sufficiency and substantive eligibility are distinct, and neither a polished narrative nor software-generated documentation guarantees acceptance.

The Accuracy-Related Penalty and Deterrence

Section 6662 can impose a 20% accuracy-related penalty on applicable underpayments. PDG’s penalty result followed the parties’ express stipulation: if none of the trial projects qualified, penalties would apply in credit years with deficiencies. The court did not independently decide the broad reasonable-cause and good-faith theory attributed to it in the source document.

Businesses should assess penalty exposure and the quality of any adviser-supported claim. The decision does not impose a blanket ban on retrospective interviews or establish that every reasonable estimate is impermissible. Such evidence still needs a sound factual basis and consistency with the records.

Intersection with Sections 174 and 174A

For tax years beginning in 2022 through 2024, the TCJA generally required capitalization and amortization of specified research or experimental expenditures over five years for domestic research and fifteen years for foreign research. That historical regime should not be presented as the unchanged domestic rule for 2025 onward.

Public Law 119-21 added Section 174A, generally allowing immediate deduction of domestic research or experimental expenditures for tax years beginning after December 31, 2024, with an election to capitalize and amortize eligible domestic costs. Foreign research generally remains subject to fifteen-year amortization under Section 174. Transition provisions address unamortized domestic costs and eligible small-business retroactive relief; Revenue Procedure 2025-28 provides implementing procedures.

Deduction treatment and credit eligibility remain separate. An expenditure may fall within the research expenditure rules without satisfying all Section 41 requirements. The proper treatment depends on the applicable year, elections, expenses, and credit coordination rules.

The Procedural Dimension: CDP Hearings and the Collection Statute

The source document mentions Hall v. Commissioner, Docket No. 23342-22L, in connection with collection procedure. It should not be conflated with the PDG research-credit case. The following collection principles arise from the statutory rules and IRS guidance, not from an R&D holding attributed to Hall.

The ordinary collection period generally runs for ten years after assessment, subject to suspension and extension provisions. A timely collection due process (CDP) hearing request can suspend that period. A credit disallowance is not itself a CDP proceeding, and deficiency litigation must be distinguished from review of collection action.