Paul L. Caron

Monday, August 2, 2021

Lesson From The Tax Court: COGS Are Not Deductions

Camp (2017)Today’s lesson is short, but not so sweet for the losing taxpayers in BRC Operating Company LLC, Bluescape Resource Company LLC, Tax Matters Partner v. Commissioner, T.C. Memo. 2021-59 (May 12, 2021) (Judge Pugh).  The taxpayer had claimed $160 million in Cost of Good Sold (COGS) for tax years 2008 and 2009.  But the taxpayer had no sales of goods in those years.  Judge Pugh teaches us a seemingly simple lesson:  you don’t get to claim COGS without any actual goods being sold.  Once again we learn how the concept COGS differs from the concept of deductions.  I last discussed this three years ago in Lesson From The Tax Court: Into The Weeds on COGS, TaxProf blog (June 25, 2018).

The shortness of the lesson, however, belies a metaphysical murkiness lurking underneath it.  Just what the heck is COGS, anyway?  On the one hand it’s not a deduction because it comes in the process of adjusting gross receipts to determine gross income.  On the other hand, it functions like a deduction, to account for the expense of producing income.  Judge Pugh appears to believe it is required by the Constitution.  That may be true if accountants had written the 16th Amendment.  But they didn’t.  So I’m not so sure there is any constitutional basis for the concept (pun intended).   I believe that murkiness is best explained by §1001.  Details below the fold.

Law: The Role of COGS in Computing Taxable Income
Congress imposes a tax only on “taxable income.” See §1 (individuals); §11 (corporations).  Section §63 defines “taxable income” as “gross income minus the deductions allowed by this chapter.”  Those deductions are listed in a series of statutes introduced by §161, which says “in computing taxable income under section 63, there shall be allowed as deductions the items specified in this part.”  The most famous of these statutes is that workhorse deduction statute, §162, which permits a deduction for “all the ordinary and necessary expenses paid or incurred during the taxable year in carrying on any trade or business.”

Since §63 defines taxable income as “gross income” minus deductions, we are thus led to the definition of gross income in §61, perhaps the most metaphysical tax statute in the entire Code (excepting, perhaps, §6211).  Section 61 defines “gross income” to be, mysteriously, “all income from whatever source derived, including (but not limited to)” a long, long list of sources of gross income.  Two of the items in the long, long list are: “gross income derived from business” and “gains derived from dealings in property.”  It’s the relationship of these two types of gross income that I think explains the concept of COGS.  But you cannot understand that without looking first at §1001.

When a taxpayer sells property, §1001 that says that a taxpayer has gain on the sale or exchange of property only to the extent that the amount realized exceeds the taxpayer’s basis in the property sold.  A trade or business may or may not involve dealings in property.  When it does, we call the items of property being sold “inventory.”  COGS is the method of allocating basis to that inventory—the items of property a business sells—to determine the “gains derived from dealings in property.”  See Treas. Reg. 1.61-3(a).  Those gains then become part of the “gross income derived from business.”  Notice that the amount realized from the sale of the property is not considered gross income.  Only the gains.  In the context of a business that sells property, the term “gross receipts” seems synonymous with the term “amounts realized.”

Let me try an example.  If I mess this up, please point it out in the comments.

Say my business is selling car parts. I stock Duralast radiators which I sell for $110 each.  I am selling property.  My stock of radiators is inventory.  Assume I can fairly allocate a cost of $100 to the radiator I am selling for $110.  That means I have an amount realized, or gross receipts, of $110 but my “gains derived from dealings in property” for the radiator is $10, per §1001.   My gross receipts do not become my gross income until after we account for my inventory.

Now multiply the example times hundreds or thousands of sales.  It is COGS accounting that allows me to allocate my total costs of acquiring or producing my inventory (the property I am selling) to each individual sale.  I thus need only report as gross income from my business the net gain (if any) allocable to each item or unit sold.

In contrast, if I operate a car repair shop and agree to fix your car’s cooling system, I may need to go buy a radiator and install it.  I may charge you $210, $110 for the part and $100 for the service.  Now the entire $210 is my “gross income derived from business” because my business is not selling car parts, it’s fixing cars (put aside businesses that sell both products and services like Battery Joe or Discount Tires).  Yes, I can still account for the cost of the radiator but now that accounting comes after I compute my gross income.  When I compute my taxable income §162 permits me to account for the $100 cost of the radiator by deducting that cost as an ordinary and necessary expense, so my taxable income ends up being net of the cost of the radiator, just as if I was a car parts store.

In short, both COGS and §162 effectuate the same recognition that it takes money to make money and Congress permits taxpayers to account for the expenses in producing income before imposing a tax on the net.  These concepts differ, however, in their timing, COGS is used to compute “gross income” for certain taxpayers, and §162 is used to compute “taxable income” for all taxpayers.

For those interested in more theory, you will have fun with Joe Dodge’s “The Netting of Costs against Income Receipts (Including Damage Recoveries) Produced by Such Costs, without Barring Congress from Disallowing Such Costs,” 27 Va. Tax Rev. 297 (2007) (advocating a computational netting approach to the treatment of contingent attorneys fees). My takeaway from that article is that there is no principled distinction between COGS and §162.  They are simply different approaches to the same problem: how to account for the money it takes to make money so that only the net income gets taxed.

The tax accounting statutes reinforce this takeaway.  Section 451 et seq. gives the rules income accounting.  Section 461 et seq. gives the rules for deduction accounting.  But neither sequence gives the rules for inventory accounting.  Nope.  You find those rules in §471 et seq.  And while you are there you find that some expenditures have a chameleon-like quality in that they are partly “deductions” and partly “COGS.”

Take salaries, for example.  Treas. Reg. 1.471-11(b) tells you that some salaries can be accounted for as part of the COGS calculation of gross income while other salaries must be taken as deductions.  For most businesses, perhaps it does not make much of a tax difference whether a particular employee’s salary is taken as a §162 deduction or goes into COGS.  But for some businesses it might. See e.g. §280E.

Facts and Parties Arguments
BRC was 100% owned by Bluescape for the tax years in dispute and was a disregarded entity.  So like Judge Pugh, I will refer to the taxpayer here as Bluescape.  It appears to have been an accrual method taxpayer.

In 2008 and 2009 Bluescape spent $180 million acquiring leases to explore for, mine, and produce natural gas on hundreds of thousands of acres in West Virginia, Pennsylvania, Ohio, and Kentucky.  Bluescape did not, however, actually drill in those years.  Nor did it pay someone to drill.  Nor did it acquire drilling property.  Nor did it sell any gas in those years.  In fact, Bluescape reported zero gross receipts in 2008 and only $140,000 in gross receipts in 2009 (but not from sale of gas).  However, Bluescape did claim $160 million in COGS.  That amount appears to have reflected Bluescape’s anticipated costs of drilling and extracting the gas it intended to sell—just as soon as it could find and extract it.  The opinion is not clear on why Bluescape was contending it had incurred an obligation to drill sufficient to satisfy the all-events test, but I think everyone was assuming that was the case.

The parties framed the issue before the Tax Court as whether the economic performance requirements in §461(h)(1) applied to Bluescape’s anticipated drilling costs.  The IRS said yes, because of Treas. Reg. 1.61-3(a).  That provision says that COGS “should be determined in accordance with the method of accounting consistently used by the taxpayer. Thus, for example, an amount cannot be taken into account in the computation of cost of goods sold any earlier than the taxable year in which economic performance occurs with respect to the amount (see § 1.446-1(c)(1)(ii)).”  That latter regulation says that COGS are to be accounted for just as any other liability and that means it cannot be accounted for until economic performance has occurred.  In Bluescape’s case, economic performance would seem to occur only once the actual drilling property or drilling services were provided to it.  See Treas. Reg. 1.461-3(d)(7), Example 4 (“Services or Property Provided To the Taxpayer).

Bluescape said no, because COGS was part of its determination of gross income in the first place.  Bluescape claimed that the economic performance rules should not apply to COGS but only to §162 deductions.

The IRS’s back-up position was that even when viewing COGS as part of the computation of gross income, §446 still requires a taxpayer’s accounting to clearly reflect income.  Allowing COGS here would distort Bluescape’s income for 2008 and 2009.

Judge Pugh was not happy with the way the parties had framed the case.  She asked a different question.

Lesson: No Sales?  No COGS!
Judge Pugh preferred to frame the question this way: can a taxpayer claim a COGS in a year when there were no gross receipts from the sale of goods?  She notes that the parties did not dispute whether the estimated drilling costs should be included in costs of good sold; they dispute only when.” Op. at 10 (emphasis in original).  To answer that question she consulted the case law and concluded that the cases “illustrate what is obvious from the phrase ‘cost of goods sold’ itself: ‘goods sold’ are generally a prerequisite to recognizing cost of goods sold.  *** Cost of goods sold does not exist in a vacuum, as a stand alone deduction in the Code, but serves as an offset against gross receipts.”  Op. at 14. She then concludes that “because Bluescape had no gross receipts from the sale of natural gas for the years in issue, the estimated drilling costs reported as [COGS] are not allowable as a cost of the goods sold offset to gross receipts.” Op. at 21.

End of lesson.  I told you it was short.

But if your coffee cut is not yet empty, I invite you to linger on a loose end.

Comment: Is Allowance for COGS Constitutionally Required?
While not necessary for the resolution of the case, Judge Pugh appears to view COGS as required by the Constitution.  She asserts that Congress may not constitutionally tax the gross receipts of a producer or reseller, citing to Doyle v. Mitchell Bros. Co., 247 U.S. 179 (1918).  Op. at 8.  She then explains that taxing gross receipts before accounting for the cost of goods sold would be an unconstitutional taxing “capital” and the “cost of goods sold offset against gross receipts ensures that there is not a tax on the return of capital.”  Op. at 9.

I confess this is confusing to me.  There are excellent reasons for the COGS allowance, just as there are for various other Code sections—such as §165 and §167—that permit taxpayers to exclude from income that which represents a return of basis.  I’m just not sure the reasons are constitutional ones.  I submit that the reason for COGS is statutory: the linkage is to §1001.  I think there is both a legal problem and an accounting problem with saying a COGS calculation is constitutionally mandated.  And if one does concluded that COGS is constitutionally required, that bodes ill for the constitutionality of statutes like §280E.

  1. The Legal Problem

As a matter of law, I doubt Doyle is great authority for the proposition that COGS is constitutionally required, chiefly because it was not a case about COGS, but also because it rests on an outdated concept of gross income.

First, the actual issue in Doyle was not about COGS.  It was about whether appreciation in timber property that had accrued before 1913 could be taxed as income when the timber was cut and sold after 1913.  The government wanted to tax the gross receipts of the timber sales as gross income.  The Court said no, because the government could not tax income before 1913.  Thus “the object is to distinguish capital previously existing from income taxable under the act.” The value of the cut timber attributable to pre-1913 appreciation could not be taxed.  Only the appreciated value of the timberlands after the 1913 was subject to the new income tax laws. Thus, the pre-1913 value had to be accounted for: it was that which ought not to be taxed.  In saying that, it is true that the Court likened that appreciation to something called “capital.”  But that was not important to its analysis.  Whether accounting for the pre-1913 value was in the form of COGS or otherwise did not really matter.  Wrote Justice Mahlon Pitney: “It may be observed that it is a mere question of methods, not affecting the result, whether the amount necessary to be withdrawn in order to preserve capital intact should be deducted from gross receipts in the process of ascertaining gross income or should be deducted from gross income in the form of a depreciation account in the process of determining net income.”  Id. at 188.

Second, Doyle used accounting concepts for its view of what constituted gross income, specifically the trust accounting concepts of stocks and flows.  Roughly speaking, stocks are what you find on the balance sheet and flows are what you find on an income statement.  If you think of stocks as capital, then “gross income” is what flows from that.  See generally, Joe Dodge, “The Story of Glenshaw Glass,” Tax Stories 2d (Paul Caron, ed.) (2009).

You see this capital/income distinction in Doyle“Whatever difficulty there may be about a precise and scientific definition of "income," it imports, as used here, something entirely distinct from principal or capital either as a subject of taxation or as a measure of the tax, conveying, rather, the idea of gain or increase arising from corporate activities. *** Income may be defined as the gain derived from capital, from labor, or from both combined.” Id. at 185 (internal citation and quotes omitted).

Other early Supreme Court cases also adopted the accountant’s view of the matter.  The most famous was Eisner v. Macomber, 252 U.S. 189 (1920), which restated Doyle’s idea that income flowed from either capital or labor.  That meant (people thought) that Congress could not tax receipt of windfalls or punitive damage awards as gross income.  Macomber also said that Congress could not constitutionally tax gross income until it was “realized.”  Macomber relied on the same accounting conception of income as did Doyle to determine the legal definition of gross income.

The Supreme Court has, however, since changed its view about the legal definition of income.  For example, the source doctrine—which the Macomber Court thought in 1920 was constitutionally mandated—was abandoned by the Court in Commissioner v. Glenshaw Glass Co., 348 U.S. 426 (1955).  That was because the Court shifted the legal meaning of gross income from an accounting concept to an economic concept: income was not a flow a source of capital or labor, but was instead any accretion of wealth, fully realized, over which the taxpayer has dominion and control.  Similarly the realization requirement was demoted from a constitutional to statutory requirement in a series of Supreme Court cases that came after Macomber.  I talk about all this in gruesome detail in Bryan Camp, "Taxation of Electronic Gaming," 77 Wash. & Lee Law Rev., 661, 702-732 (2020)(discussing history of the legal doctrine of “gross income”).  For a contrasting view, suggesting that realization is still constitutionally required, see Henry Ordower, “Revisiting Realization: Accretion Taxation, The Constitution, Macomber, and Mark to Market,” 13 V. Tax Rev. 1 (1993).

In short, I am confused by Judge Pugh’s cite to Doyle.  Whatever the Doyle Court may have thought was constitutionally mandated in 1918 is not really solid precedent now, even though Doyle is certainly solid precedent for the proposition that Congress cannot tax income earned or received before 1913!  But that’s not a problem for most of us.   

  1. The Accounting Problem

My other confusion is how to reconcile Judge Pugh’s apparent position that Congress cannot constitutionally tax gross receipts (hence necessarily implying that COGS is constitutionally mandated) with the settled law that the deductions from gross income for depreciation are mere matters of legislative grace.

I see both COGS and depreciation deductions as expressions of the same policy: they account for the money it takes to make money.  They just do so at different points in the computational process.  Put another way, both COGS rules and depreciation rules answer the question of what part of gross receipts constitute a return of the expenditures made to produce them—whether those expenditures are labeled “capital” or otherwise—and what constitute a return of profit.  The answer to that question is Congress’s to implement.  If one asserts that the policy is constitutionally mandated, that creates two problems for me: what counts as capital; and by when must it be accounted for so that it is not unconstitutionally taxed.

First, I don’t really know what “capital” means.  The dicta in Doyle that Judge Pugh relies upon was just the Court’s concern that a tax on “capital” might be a direct tax like a tax on real property might be.  But the concept of a capital expenditure is a much more amorphous concept that real property.  I am just not sure there is any satisfactory single top-down definition of the term.  Check it out for yourself in this Wikipedia entry

I prefer giving the term a bottom-up definition, the same as I do for basis: “that which ought not to be taxed.”  The definition is defiantly circular because I think there are many reasons why a particular receipt of wealth ought not to be taxed.  My problem is that I cannot say any of those reasons are constitutionally based.  The Constitution contains no prohibitions on what Congress can tax.  It contains only one restriction: a “direct tax” must be apportioned among the states.  Erik Jensen has a nice explanation of this in his article “Marijuana Businesses, Section 80E, and the 16th Amendment," 168 Tax Notes Fed. 1643 (Aug. 31, 2020).

Second, the accounting problem is to ensure that what gets reported as taxable income is a proper reflection of the economic increase in wealth.  Thus the question, as Judge Pugh points out, is really when to account for the money it takes to make money.  To expand on Justice Pitney’s point in Doyle, one could do that in a COGS calculation, or a §162 expense calculation, or a §167/168 depreciation calculation or a §195 start-up expenses amortization calculation.  If one takes seriously the notion that COGS is constitutionally required to protect “capital” (whatever the heck that is), then one needs to identify the constitutional timing requirement.  Otherwise, I do not see how other provisions in the Code that protect “capital” from taxation are less constitutionally required.  If it is, then Congress could probably not constitutionally impose a “basis last” rule, thus requiring businesses to account for capital only when the business is sold or closed, as one fellow academic pointed out to me.

For example, assume a business grows and sells flowers and spends $700 to buy a grow light for its business.  Most of us would say unthinkingly that the $700 was a capital expense, giving the taxpayer $700 basis in the grow light, even if purchased with borrowed funds.

How do we account for that $700?  If it is part of COGS, then the business would reduce its gross receipts by the properly apportioned $700.  But if not accounted for by COGS, the taxpayer would still be able to deduct the $700 against the income it helped produce.  Sure, the expenditure might not be deducted under §162 but might instead be capitalized and recovered through depreciation.  And, sure, like some mad post-modern playwright intent on breaking the Fourth Wall, Congress spastically tears down the distinction between expensing and capital recovery.  See e.g. §179 or §168(k).  But the point is that some post-COGS deductions are explicitly designed to protect “capital” from taxation in the same way COGS does.

Now, let's change our business from a flower grower to a marijuana grower.  Section 280E disallows any and all deductions for marijuana growers, including depreciation deductions.  But if COGS is constitutionally mandated to protect “capital,” then it would seem that the various statutory deductions that are also designed to protect capital should also be constitutionally required.  That is the argument marijuana grows have been pressing for years.  They keep losing because the courts keep distinguishing COGS from deductions.  Here’s an example from Alpenglow Botanicals, LLC v. United States, 894 F.3d 1187 (10th Cir. 2018):

“Although there can be similarity between...cost of goods sold and ordinary and necessary business expenses...the cost of goods sold relates to acquisition or creation of the taxpayer's product, while ordinary and necessary business expenses are those incurred in the operation of day-to-day business activities. The cost of goods sold is a well-recognized exclusion from the calculation of gross income, while ordinary and necessary business expenses are deductions. Indeed, while the Tax Code has statutorily excluded certain expenses from the calculation of gross income, only the cost of goods sold is mandatorily excluded by the very definition of ‘gross income’ even in the absence of specific statutory authority for such exclusion.”

Id. at 1200 (emphasis supplied, internal quotes and citations omitted).

To me, the Apenglow court misses the timing point and misses the statutory basis for COGS.  The difference between COGS and deductions is not the difference between exclusion and deduction, it’s a difference of timing: when may taxpayers account for the expenditures—including the capital expenditures—required to produce the income received.  The reason for the difference in timing is not a constitutional reason, but a statutory one.  COGS is linked to §1001’s statutory formula for determining gains from the sale of property.  Thus, when your business consists of dealings in property (whether selling car parts, nursery supplies, or plants) it is §1001 that supports the COGS calculation as part of determining “gross income” within the scope of §61.  COGS is just another expression of that classic choice: basis first, basis last, or basis ratably.  Under COGS, taxpayers get to ratably recover their costs of acquiring or producing inventory.

In allowing taxpayers a ratable recovery of basis, COGS (as an expression of §1001) is no different than other statutory provisions (e.g. depreciation, installment sales, life insurance payouts, annuity payouts) that also deal with the same problem of how to split income between that which is to be taxed and that which is not to be taxed.  Just because Congress chooses the ratably approach does not mean another timing choice would be unconstitutional.  Even if return of capital must not be taxed as income, denial of COGS is no more unconstitutional than denial of depreciation.  That is because denial now just kicks the capital-recovery can down the road.  The ultimate protection of “capital” from taxation is time.  The taxpayer will increase basis in the business until the proper time for accounting, whenever that comes.  I think that could be as late as the disposition or cessation of the business, although some of my academic colleagues disagree. 

So that is why I am doubtful that COGS is constitutionally required.  As usual, I welcome comments to point out flaws in my reasoning or gaps in my knowledge.  It’s part of learning.

Bryan Camp is the George Mahon Professor of Law at Texas Tech University School of Law.  He invites readers to return each Monday to TaxProf Blog for another Lesson From The Tax Court.

Bryan Camp, New Cases, Scholarship, Tax, Tax Scholarship | Permalink


“To preclude possible challenges on constitutional grounds, the adjustment to gross receipts with respect to effective costs of goods sold is not affected by this provision of the Act.” That’s in the Senate Finance Committee report on 280E in TEFRA,, page 309.

Maybe staff stuck that in (I was on the Joint Tax staff when 280E was enacted, but didn’t work on it) not because of some iffy constitutional concern, but because not allowing a deduction for COGS would crazily and nonsensically penalize non-vertically integrated operations to the point of no return. Without a COGS deduction, the tax would pyramid on every sale along the supply chain. The grower couldn’t deduct his costs, and the wholesaler couldn’t deduct what he paid the grower, the retailer couldn’t deduct what he paid the wholesaler, and so on. Only vertically integrated operations could survive.

Posted by: Patrick Oglesby | Aug 8, 2021 6:37:08 AM