Row Crop Farming R&D: Guidance the JG Boswell Revenue Case Provides to Taxpayers

When a large California farming operation claimed more than $17 million in qualified research expenses (QREs), the IRS pushed back hard. The resulting Tax Court order in J.G. Boswell Co. & Subsidiaries v. Commissioner of Internal Revenue (No. 2408-19, decided July 12, 2022) is essential reading for any agricultural business claiming the research credit under Section 41. The case turns on a single question: was the farm researching a better product, or just a cheaper way to grow the same one?

‍

‍

The Background

‍

J.G. Boswell Company operates a large farming business in California's central valley, growing cotton, tomatoes, alfalfa hay, and safflower, and producing products including cotton lint and tomato paste. The company ran a structured agronomic research program.

‍

As the court described it:

"As part of an ongoing effort to improve its business through agronomic research, petitioner conducted 33 research trials in 2013 and 22 research trials in 2014 on about 7% of its farmland (research acres), in general by growing crops on a test plot and a control plot. Petitioner's researchers grew one of petitioner's standard crops on the control plot, using the standard, nonexperimental method petitioner used on the other 93% of its farmland (production acres). On the test plot the researchers tried an experimental method, which they evaluated by comparing the output of the test plot to the output of the control plot. Petitioner claims the researchers hoped each experimental method would improve the quality or yield of the crop or reduce economic or environmental costs of production relative to the standard method."

On its 2014 return, Boswell claimed $17,062,147 of QREs and a corresponding Section 41(a) credit of $1,735,424, generally including all costs of cultivating and evaluating the research acre crops. The IRS allowed a credit of only $117,722 based on QREs of $885,132, the amount it said Boswell could prove it spent on the research acres beyond what it would have spent farming that same land normally. The result was a deficiency of $1,102,464.

‍

‍

The Legal Question: Product vs. Process

‍

The dispute centers on Section 41(d)(2)(C), which treats the product a taxpayer produces, and the process used to produce it as separate business components. Relying on Union Carbide Corp. & Subs. v. Commissioner, the IRS argued that all of Boswell's trials were process research, meaning the company could only claim the incremental costs of experimentation, not the baseline costs of growing crops it would have incurred anyway.

‍

The court agreed that the Union Carbide reasoning applies to process-only research. If a trial seeks only to produce the same crop at lower cost or higher yield, the ordinary costs of cultivation do not become QREs just because an experiment happened on that field.

‍

Boswell argued that even its cost-cutting trials required researchers to confirm the product was not harmed. The court was not persuaded:

"Petitioner points out that even where its research sought only to cut production costs, researchers had to verify that its experimental method would not have 'a negative effect on the resulting plant and derived product.' To the extent petitioner thereby implies that all 55 research trials pertained to the product business component, this argument proves too much."

The court continued:

"Here again, allowing petitioner to treat as QREs the usual costs of cultivating the research acres just because any change to petitioner's production processes could affect its crops would give no effect to section 41(d)(2)(C)."

‍

Where the Taxpayer Won Ground

‍

The order was not a loss for the farm. The court drew a critical line in the taxpayer's favor:

"Section 41(d)(2)(C) does not apply to any research trials that sought to improve one of petitioner's products."

The court pointed to Boswell's cotton seed trial as a possible example. In the court's words:

"Petitioner tried planting new seed varieties, a revised production process that produced not only more cotton than the standard process, but higher quality cotton that could be sold at a higher price per unit."

To the extent the company set out to improve cotton quality, the court explained, "petitioner's cultivation of the cotton on the research acres is a new research project to develop improved cotton, a new business component."

‍

The court also rejected the IRS's broader theory that baseline production costs can never be QREs:

"To the extent respondent invites us to interpret Treasury Regulation § 1.41-2 to exclude from QREs all expenditures a taxpayer would have incurred had it used its standard production process, even if the research seeks to improve the product rather than the process alone, we decline the invitation."

Perhaps most significantly for agricultural taxpayers, the court treated experimental crops like prototypes. This passage is worth reading in full:

"Moreover, petitioner's cultivation of experimental products is analogous to the regulation's example of a machinist's fabrication of part of an 'experimental model,' which is direct support if the model is used in qualified research. The regulation does not define 'experimental model,' but the term connotes the definition of 'pilot model' in Treasury Regulation § 1.174-2(a)(4): 'any representation or model of a product that is produced to evaluate and resolve uncertainty concerning the product during the development or improvement of the product.' See Little Sandy Coal Co., Inc. v. Commissioner, T.C. Memo. 2021-15, at *30-31. The same regulation confirms that a trial production to 'evaluate and resolve uncertainty during the . . . improvement of the product' is a pilot model even if the taxpayer sells it. See Treas. Reg. § 1.174-2(a)(11), ex. 7. If petitioner demonstrates that its tests of the viability of improved crops are section 41(d) qualified research, therefore, the trial crops it produced were experimental models, and the associated costs are QREs to the extent permitted by section 41(b). Respondent cannot defeat any such argument simply by demonstrating that petitioner would have incurred the expenditures had it used its standard production processes on the research acres."

In other words, a test crop grown to evaluate an improved product can be a pilot model, and the full cost of producing it can qualify, even if the crop is ultimately sold.

‍

‍

The Outcome

‍

The court denied both the IRS's motion for summary judgment and Boswell's cross-motion for partial summary judgment. The case will require further factual development to determine which of the 55 trials sought to improve a production process alone, which sought to improve a product, and the QREs attributable to each. The court also made clear that the taxpayer bears the burden of establishing its disallowed QREs.

‍

‍

Takeaways for Agricultural Businesses

‍

The Boswell order offers three practical lessons for farms and agribusinesses claiming the research credit.

‍

First, the purpose of each trial matters enormously. Research aimed at improving the crop itself, such as quality, can support a full-cost QRE claim. Research aimed only at cutting costs or boosting yield of an unchanged crop limits QREs to incremental experimental costs.

‍

Second, documentation of research objectives needs to be established at the outset. Because the taxpayer bears the burden of proof, contemporaneous records showing that a trial targeted product improvement can be the difference between claiming full cultivation costs and claiming only the increment.

‍

Third, selling your test crops does not disqualify them. Under the pilot model framework the court endorsed, trial production grown to resolve uncertainty about an improved product can qualify even when it goes to market.

Cases like Boswell underscore why precision matters in building and defending an R&D credit claim. Distinguishing product research from process research is not an academic exercise. In this case, it was the difference between a $1.7 million credit and a $117,722 one.