SWIFT v. COMMISSIONER OF INTERNAL REVENUE (2025)

Dr. Bernard T. Swift and his wife took tax deductions for insurance premium payments Swift's medical practice made to captive insurance companies. The Internal Revenue Service issued notices of deficiency, disallowing the deductions and imposing penalties against the Swifts. The tax court sustained the determination of deficiencies and penalties, finding the payments were not really for insurance. We AFFIRM.

I. Factual & Procedural Background

This case arises from a micro-captive transaction: “an insurance agreement between a parent company and a ‘captive’ insurer under its control.” CIC Servs., LLC v. Internal Revenue Serv., 593 U.S. 209, 213, 141 S.Ct. 1582, 209 L.Ed.2d 615 (2021). Micro-captive transactions have the potential to come with tax advantages. Under section 831(b) of the Internal Revenue Code, a captive insurer with annual premiums of less than $1.2 million (a “micro-captive”) pays no taxes on premium income.¹ See id. And the insured party can deduct what it pays in premiums as a business expense under section § 162(a). Id. “The result is that the money does not get taxed at all.” Id. Of course, these tax benefits should only accrue if the money is “really for insurance.” Id. With that context, we turn to the largely undisputed facts.

A. The Insurance Arrangement

Petitioner-Appellant Dr. Bernard T. Swift is the founder and sole proprietor of an urgent care center with 18 locations, called Texas MedClinic, and two other smaller entities focused on sports rehabilitation and dermatology (together, the “Clinic”). In 2004, Swift began exploring the possibility of creating a captive insurance company. His efforts led him to Celia Clark, a lawyer who specialized in forming and maintaining captive insurance companies.

With Clark's help, Swift incorporated his first captive (Castlegate Insurance Co, Ltd.) in 2004, and it operated until 2009, reporting premiums just under the $1.2 million cap. Although the tax court provided more detail on the formation, operation, and legitimacy of Castlegate, this dispute concerns only the tax years 2012–2015, so we jump ahead to the formation of the two captives relevant here.

In 2010, Swift, with Clark's assistance, formed two new captives: Castlerock Insurance Co., Ltd. and Stonegate Insurance Co., Ltd. (“the Captives”), incorporated in the Federation of Saint Christopher and Nevis. Each was owned by a trust benefiting one of Swift's two children, for which Swift and his wife were trustees. The Captives' business plans referenced concerns about obtaining medical malpractice coverage at a reasonable cost and obtaining appropriate levels of terrorism risk insurance. Around the time the Captives were formed, Swift emailed Clark's office that he would “be looking for other risks to insure” to “get closer to maxing out the premiums.”

1. The Captives' Direct Policies

The Captives issued various lines of insurance to the Clinic that fall into two categories: medical malpractice coverage and nonmedical coverage. The medical malpractice coverage was a “tail” policy, meaning that it covered claims related to professional services rendered at the Clinic before the policy started, but reported after the policy started. The tail policies covered the Clinic's physicians back to their start date at the Clinic. The Clinic's physicians signed a form each year acknowledging the coverage, and were responsible for their relative portion of the deductible and any losses exceeding policy limits on claims made against them.

To determine pricing for the medical malpractice policies, Swift engaged an actuary with KPMG. KPMG considered the premiums for comparable policies, reporting lag factors, physician specialties, and the period of exposure. In addition to “pure premiums,” KPMG provided a range of percentages, known as “loads,” intended to capture administrative expenses, profit, and contingencies. Swift chose a load percentage of 20% to arrive at the final premiums. Premiums for the medical malpractice coverage ranged from about a third to just over half of the total premium payments made to the Captives each year.

The remaining premium payments went toward the nonmedical coverage, including policies for administrative actions, business income, business risk indemnity, computer operations and data, employment practices, litigation expense, cost of defense, terrorism, and political violence. Clark worked with another actuary, Allen Rosenbach, to price these nonmedical policies. Rosenbach provided premium estimates for each line of coverage, explaining they were “based on a survey of rating plans obtained from regulatory filings submitted by commercial insurance carriers” and adjusted to “reflect our interpretations of the differences in the captive insurance policies.” Clark and her team often provided Rosenbach with a total maximum premium, and Rosenbach responded with an estimate close to that maximum.

2. The Captives' Participation in Reinsurance Pools

Relying on her interpretation of tax court cases and Internal Revenue Service (IRS) rulings, Clark advised that 30% of the Captives' premiums should come from unrelated businesses. To achieve this, the Captives participated in two reinsurance pools affiliated with Clark: Jade Reinsurance Group, Inc. (Jade) in 2012 and 2013, followed by Emerald International Reinsurance, Inc. (Emerald) in 2014 and 2015. Jade and Emerald consisted of around 100 captive insurance companies affiliated with Clark. Clark represented to Swift and the other participants that the pools were designed to comply with IRS rulings and to result in risk distribution well above 30%.

At a high level, each participating captive paid premiums to the pool, and in return, the pool reinsured a portion of each captive's risk. Those same captives also contracted with the pool to provide this reinsurance coverage, each covering a quota share portion of the pool's blended liability. The pool paid each of the captives for covering their quota share of the blended liability by releasing a percentage of the total premiums received back to the captive. To join, captives submitted an application with limited information such as the value of property insured and past claims, and other participants were given a week to determine whether to exclude applicants from their respective pools.

The pools provided three types of reinsurance policies referred to as Coverage Part A, B, and C. Part A reinsured a wide variety of insurance policies, Part B reinsured “General Cost of Defense Insurance” policies, and Part C reinsured “Terrorism or Political Violence” policies. For Part A policies, the pools set premiums as a percentage of the underlying premiums the captives charged for the direct policies: Jade varied this percentage based on the direct policies' per-claim limits while Emerald charged a flat 30.3%. For Part B policies, the pools charged a flat $7,000 premium. The pools set varying limits for the reinsurance coverage provided under Part A and Part B (for example, $10,000 per claim under Part B).

For Part C policies, the pools set premiums as a percentage of the underlying premiums, and agreed to reinsure that same percentage, but the captives chose the percentage. The tax court found this was consistent with Clark's view “that terrorism and political violence policies could be used as necessary to assure the desired risk distribution.” Swift v. Comm'r, T.C.M. (RIA) 2024-013, 2024 WL 378671, at *12 (Feb. 1, 2024). Clark assured participants that the total premiums ceded to the pools “will be at least 30% and will usually be substantially above that.”

The Captives' arrangement with the reinsurance pools looked similar in each of the four years at issue. The Captives paid premiums to the pool (and the pool reinsured the Captives in return) in an amount equal to around 30% of the total premiums the Captives received from the Clinic. Importantly, participation in the reinsurance pools was limited to the Captives' nonmedical policies; the Captives did not reinsure their medical malpractice tail policies.

* * *

From 2012 through 2015, the years at issue, the Clinic paid the Captives a total of $5.98 million in premiums. During this time, the Clinic submitted three claims to the Captives, all of which were approved and paid despite irregularities, resulting in total payment of $339,224. One of the three claims was made under the medical malpractice tail policies and paid out $13,212. The other two were made under the nonmedical policies, with the largest payment for $275,793 made under the “Administrative Actions Insurance Policy” to cover the pending IRS audit. The Captives invested most of the premiums received, causing Clark to express concern about the illiquidity of the Captives' portfolio.

For context, during the years at issue, the Clinic also paid commercial insurance carriers about $480,000 for both medical malpractice and nonmedical coverage. About $180,000 went toward commercial medical malpractice policies covering claims that both occurred and were reported during the one-year term (known as “claims-made” policies, as opposed to the “tail” policies the Captives issued). The remaining $300,000 went toward other lines of commercial insurance, including policies covering buildings, equipment, accounts receivable, fiduciary liability, workers' compensation, and terrorism, and a $10 million umbrella liability policy.

B. The IRS Examination and Tax Court Decision

Swift and his wife, Petitioner-Appellant Kathy L. Swift, claimed the premium payments made to the Captives as business expense deductions on their joint tax returns for the years 2012 through 2015. After conducting an examination, the IRS issued the Swifts notices of deficiency disallowing the deductions. The notices reflected deficiencies of $893,809, $596,855, $494,259, and $461,524 for each respective year. The notices also included 20% accuracy-related penalties under I.R.C. § 6662 for negligence or substantial understatement.²

The Swifts petitioned the tax court, challenging the IRS's determination of the deficiencies and penalties. After a trial, the tax court sustained both in a lengthy opinion. The Swifts timely appealed.

II. Analysis

On appeal, the Swifts challenge the tax court's decision to sustain the determination of deficiencies, arguing the tax court erred in determining that the arrangement with the Captives did not constitute insurance. Independent of the ruling on deficiencies, the Swifts contend the tax court erred in finding them liable for penalties.³ We address (A) the deficiencies, then (B) the penalties.

A. Deficiencies

“This Court applies the same standard of review to tax court decisions and district court decisions: Findings of fact are reviewed for clear error and issues of law are reviewed de novo.” Brinkley v. Comm'r, 808 F.3d 657, 664–65 (5th Cir. 2015). The “characterization of a transaction for tax purposes is a question of law subject to de novo review, but the particular facts from which that characterization is made are reviewed for clear error.” Chemtech Royalty Assocs., L.P. v. United States, 766 F.3d 453, 460 (5th Cir. 2014) (quoting Southgate Master Fund, L.L.C. ex rel. Montgomery Cap. Advisors, LLC v. United States, 659 F.3d 466, 480 (5th Cir. 2011)). Clear error exists if this court is left with the definite and firm conviction that a mistake has been made. Id.

To be deductible as business expenses under section 162(a) of the Internal Revenue Code, the insurance premium payments made to the Captives must have been “really for insurance.” See CIC Servs., 593 U.S. at 213, 141 S.Ct. 1582. While no part of the Code defines insurance, the Supreme Court has explained that “insurance involves risk-shifting and risk-distributing.” Helvering v. Le Gierse, 312 U.S. 531, 539, 61 S.Ct. 646, 85 L.Ed. 996 (1941); see also Rsrv. Mech. Corp. v. Comm'r, 34 F.4th 881, 903 (10th Cir. 2022) (explaining the Code lacks a definition). Relying on the Supreme Court's explanation of insurance in Le Gierse, the tax court generally looks to four criteria to determine whether a transaction constitutes insurance: “(1) risk-shifting; (2) risk-distribution; (3) insurance risk; and (4) whether an arrangement looks like commonly accepted notions of insurance.” Swift, 2024 WL 378671, at *16 (quoting Caylor Land & Dev., Inc. v. Comm'r, 121 T.C.M. (CCH) 1205, 2021 WL 915613, at *10 (Mar. 10, 2021)); see, e.g., Avrahami v. Comm'r, 149 T.C. 144, 177 (2017); Rent-A-Center, Inc. v. Comm'r, 142 T.C. 1, 13 (2014).

Here, the tax court found that the arrangement with the Captives did not constitute insurance because (1) it did not achieve risk distribution. In the alternative, the tax court concluded (2) that the arrangement did not constitute insurance in the commonly accepted sense. We begin with risk distribution.

1. Risk Distribution

Risk distribution is essential to insurance. Steere Tank Lines, Inc. v. United States, 577 F.2d 279, 280 (5th Cir. 1978). This court has explained that risk distribution “emphasizes the broader, social aspect of insurance as a method of dispelling the danger of the potential loss by spreading its cost throughout the group.” Ross v. Odom, 401 F.2d 464, 467 (5th Cir. 1968). It depends on “the statistical phenomenon known as the law of large numbers.” Clougherty Packing Co. v. Comm'r, 811 F.2d 1297, 1300 (9th Cir. 1987).

The law of large numbers tells us that “when there are a sufficiently large number of independent risks each having an annual probability of X%, there is an extraordinarily small likelihood that the percentage of insureds that suffer a loss during a year will deviate significantly from X%.” Rsrv. Mech., 34 F.4th at 904. Put differently, “[i]f a coin is tossed a million times, it is highly unlikely that the percentage of heads will differ appreciably from 50%.” Id. In this way, “insuring a large number of independent risks protects the insurer against financial calamity,” because the insurer can accurately predict losses for the group as a whole, and set premiums accordingly. Id. at 904–05. Therefore, risk distribution depends on the existence of a sufficient number of independent risks.

The Swifts contends that the Captives achieved risk distribution in two independent ways. First, they argue that the medical malpractice tail policies the Captives issued directly to the Clinic were sufficient to distribute risk. Second, even if these direct policies were not sufficient, the Swifts argue that the Captives' participation in the reinsurance pools distributed risk. We address each in turn.

a. The Captives' Direct Policies

The tax court concluded that the medical malpractice policies failed to distribute risk because there was not “a sufficient number of unrelated risks to allow the law of large numbers to predict losses.” Swift, 2024 WL 378671, at *17 (quoting Caylor Land, 2021 WL 915613, at *11). The tax court noted that in a given year, the Captives at most issued nine lines of coverage, covering three entities, spanning 28 locations and 530 workers. The court explained that this “risk exposure pales in comparison” with that it had deemed adequate “for the law of large numbers to apply.” Id.

On appeal, the Swifts argue that the tax court erred in determining there was an insufficient number of risks by counting the number of physicians (199), rather than the number of individual physician-patient interactions (millions). One of the Swifts' experts, David Sommer, explained that the “source of a medical malpractice lawsuit is a doctor-patient interaction” and therefore “the relevant exposure unit for this risk pool is a patient visit rather than an individual physician.” However, another of the Swifts' experts, Michael Angelina, and two of the government's experts, David Russell and Toni Mulder, looked to the number of physicians. Russell explained that a claim arises from “the alleged incompetence of the physician,” including activities beyond a doctor-patient interaction, and “the industry has coalesced around a measurement of exposure unit as a physician.”

The tax court sided with the government, reasoning that Angelina was the more persuasive of the Swifts' experts, the government's experts looked to the number of physicians, KPMG looked to the number of physicians when pricing the policies, and it was industry standard. The Swifts do not dispute this evidence on appeal, but merely reiterate their expert Sommer's view. Accordingly, the tax court did not err in finding, consistent with much of the record evidence, that a physician, rather than a doctor patient-interaction, is the relevant exposure unit.

In the alternative, the Swifts argue that their expert Sommer established there is sufficient risk distribution even if risk exposure is measured by the number of physicians, because a recognized formula shows the risk reduction achieved by a pool of 199 physicians is 92.9%.⁴ We are not convinced for three reasons. First, as the Swifts seem to concede in their reply brief, the formula does not appear in the various opinions they claim have relied upon it. See Ocean Drilling & Expl. Co. v. United States, 988 F.2d 1135, 1150 (Fed. Cir. 1993); Harper Grp. v. Comm'r, 96 T.C. 45, 55 (1991); AMERCO v. Comm'r, 96 T.C. 18, 33–34 (1991); Securitas Holdings, Inc. v. Comm'r, 108 T.C.M. (CCH) 490, 2014 WL 5470747, at *9–10 (Oct. 29, 2014).

Second, the government explains, and the Swifts do not dispute, that the formula establishes relative risk reduction, as compared to the risk faced by a hypothetical medical practice with only one physician. The government persuasively argues that relative risk reduction misses the point. “Distributing risk allows the insurer to reduce the possibility that a single costly claim will exceed the amount taken in as a premium and set aside for the payment of such a claim.” Clougherty, 811 F.2d at 1300. That possibility may be reduced with each added risk, but the question of whether it is sufficiently reduced, such that losses are predictable, remains. And the government's experts' report concluded that the Captives' expected losses were too unstable for them to have relied on the law of large numbers.

Finally, the number of independent risks ensured by the Captives is “at least a couple orders of magnitude smaller than the captives in cases where [the tax court] found sufficient distribution of risk.” Swift, 2024 WL 378671, at *18 (quoting Caylor Land, 2021 WL 915613, at *12); see Rent-A-Center, 142 T.C. at 24 (captive provided workers' compensation, automobile and general liability insurance for 14,300 to 19,740 employees, 7,143 to 8,027 vehicles, and 2,623 to 3,081 stores); Securitas Holdings, 2014 WL 5470747, at *9–10 (captive provided workers' compensation, automobile and general liability insurance to 25 to 45 entities in over 20 countries, covering over 200,000 employees and 2,250 vehicles).

Given the lack of support for the Swifts' argument that 199 exposures establishes sufficient risk distribution, the competing view offered by the government's experts, and the comparable cases from the tax court, we find the tax court did not err in finding that the Captives failed to establish risk distribution through the direct medical malpractice policies.

b. The Captives' Participation in Reinsurance Pools

In the alternative, the Swifts seek to establish risk distribution through the Captives' participation in reinsurance pools. At a high level, a reinsurance pool allows a group of insurers to transfer some portion of the risk each insures to a pool, while agreeing to insure some quota share of the pool's blended risk. See Rsrv. Mech., 34 F.4th at 886–87. In this way, a captive insurer that otherwise lacks risk distribution might achieve it by effectively trading the few, related risks it insures for the many, unrelated risks the pool insures. See Harper Grp. v. Comm'r, 979 F.2d 1341, 1342 (9th Cir. 1992) (explaining that an arrangement with a captive insurer constitutes insurance “if that captive does substantial unrelated insurance business”).

Here, the Captives participated in reinsurance pools designed to reallocate at least 30% of the Captives' business to insuring the pool's—as opposed to the Clinic's—risk. The 30% threshold appears to be rooted in Harper, which found a true insurance transaction where 29 to 33 percent of the captive's business came from insuring unrelated entities. 979 F.2d at 1342. For our purposes, we can assume that this threshold is correct. See Rsrv. Mech., 34 F.4th at 885. The issue here is not whether the Captives derived enough of their business from the reinsurance pools, but rather whether the arrangement with the reinsurance pools served to accomplish its intended purpose of risk distribution at all. See id. at 912 (examining whether arrangement with reinsurance pool constituted “a true insurance arrangement for the distribution of risk”).

The tax court found that it did not. In doing so, it referenced nine factors it uses to “determine whether a company is a bona fide insurer.”⁵ Swift, 2024 WL 378671, at *19. The tax court stressed that it did not consult the factors to determine whether the reinsurance pools met the formal definition of an insurance company, but rather whether their “products constituted insurance as necessary for [the Captives] to distribute risk.” Id.

On appeal, the Swifts take issue with the tax court's reliance on a multi-factor test, arguing it goes beyond the two-part test mandated by precedent—whether a transaction is risk shifting and risk distributing. See Le Gierse, 312 U.S. at 539, 61 S.Ct. 646; Ross, 401 F.2d at 467. We find the tax court's analysis fairly captures “whether the [reinsurance pool] arrangement was a true insurance arrangement for the distribution of risk.” Rsrv. Mech., 34 F.4th at 912. As a whole, this approach is consistent with precedent requiring that insurance involve risk distribution in substance, not form. See Arevalo v. Comm'r, 469 F.3d 436, 439 (5th Cir. 2006) (“The Supreme Court has repeatedly stressed that, in examining transactions for the purpose of determining their tax consequences, substance governs over form.”).

Of course, the Swifts are free to challenge the relevance or application of any given factor. The tax court focused on three: (1) whether there was a circular flow of funds; (2) whether the policies were arm's-length contracts; and (3) whether the entity charged actuarially determined premiums.

First, the tax court found that the amount the Captives received from the pool in premiums (for covering their share of the pool's blended liability) was close to 100% of what they paid the pool in premiums: 99.59%, 98.74%, 94.98%, and 98.99% in 2012–2015 respectively. The tax court reasoned that “[w]hile not quite a complete loop, this arrangement looks suspiciously like a circular flow of funds.” Swift, 2024 WL 378671, at *19 (quoting Avrahami, 149 T.C. at 186).

Second, the tax court found that the Captives did not enter into arm's-length contracts with the pools for reinsurance coverage. That the pools were “thinly capitalized” and could struggle to pay claims “raise[d] questions whether a reasonable business would enter into these contracts absent tax motivations.” Id. at *20. And “the contracts were set up to dissuade participants from using the pools as reinsurance” with meaningful deterrents, such as requiring the captive to pay a retained limit before making a claim and retaining “authority to exclude an insured making excessive claims from future pools.” Id.

Third, the tax court found that the Swifts failed to show how the premiums the reinsurance pools charged were derived “in light of the various risks purportedly being reinsured” and that the evidence instead showed that Clark and Rosenbach “were simply manipulating numbers to design a system where 30% of total premiums would be allocated to reinsurance before being retroceded back.” Id. at *21. Specifically, the tax court reasoned that the “tiny loss ratios suggest that premiums were much higher than what the risks called for, which calls into question whether these were actual insurance arrangements intended to distribute risk”⁶ ; charging uniform percentages to all captives “plainly fails to account for the specific risks presented by each of the agreements being reinsured through the pools”; and the pools' agreement to reinsure terrorism and political violence coverage at a percentage of the captives' choosing created “a flexible tool to adjust the reinsurance premiums to whatever level necessary to hit 30% risk distribution overall.” Id.

Based on the circular flow of funds, the lack of arm's length contracts, and the questionably determined premiums, the tax court concluded that the reinsurance policies “were not bona fide insurance arrangements” such that the Captives “could not use their reinsurance ․ to achieve the risk distribution that they lacked.” Id. The Swifts do not contest the facts underlying the tax court's analysis of each factor, but instead argue that the analysis improperly faulted Swift for making tax-motivated business decisions and erroneously focused on pricing. We reject this argument for the same reason we rejected the Swifts' objection to the multi-factor test: The tax court fairly considered the substance and pricing of the transactions to assess whether the pools really achieved risk distribution. See Rsrv. Mech., 34 F.4th at 911–12 (endorsing the tax court's reliance on many of the same factors to analyze a similar arrangement between a captive insurer and a reinsurance pool).

In the absence of any meaningful challenge to the tax court's findings, we agree with the court's conclusion that the arrangement between the Captives and the reinsurance pools was not “a true insurance arrangement for the distribution of risk.” See Rsrv. Mech., 34 F.4th at 912. Because neither the direct policies nor the reinsurance pools achieved the risk distribution that is essential to insurance, the tax court did not err in finding that the premium payments made to the Captives were not “really for insurance.” See CIC Servs., 593 U.S. at 213, 141 S.Ct. 1582.

2. Insurance in the Commonly Accepted Sense

Because we find that the tax court correctly found risk distribution lacking, we need not reach the Swifts' arguments challenging the tax court's alternative basis for sustaining the deficiencies: The Captives did not provide insurance in the commonly accepted sense. However, we note the tax court's findings—including the apparent lack of any business need for the Captives, the irregular handling of claims, the various oddities contained in the policies, the unreasonable, reverse-engineered premiums, and the availability of similar commercial coverage at a fraction of the cost—further bolster the conclusion that the arrangement did not constitute insurance.

* * *

In conclusion, we find that the tax court did not err in finding that the premium payments made to the Captives were not for insurance and could not be deducted as business expenses.

B. Penalties

The Swifts argue that regardless of whether they improperly deducted the payments made to the Captives, the tax court erred in upholding the 20% penalties for negligence or substantial understatement under I.R.C. § 6662. They challenge the penalties on two independent grounds: (1) the IRS's failure to comply with a statute requiring supervisory approval of the “initial determination” of the assessment of particular penalties, and (2) a defense of reasonable cause or substantial authority.

1. Supervisory Approval

The Swifts' first challenge to the penalties applies only to tax years 2012 and 2013. In a single letter addressing both years, IRS Agent Allen Sohrt notified the Swifts that he had completed his review, recommended disallowing the deduction of the premium payments, and enclosed a report including a 20% penalty. The letter also included a Form 870 “for an agreement of the tax and penalties at the Examination level.” Group Manager Cynthia Tam signed a Civil Penalty Approval form approving the penalties after the letter was sent, but before the issuance of a deficiency notice.

The Swifts assert that failure to secure supervisory approval before sending the letter violated I.R.C. § 6751(b). The tax court found that it did not. In doing so, the tax court applied the same interpretation of the statute the Swifts advocate for: Supervisory approval is required before formal communication of a penalty is sent to the taxpayer.⁷ Belair Woods, LLC v. Comm'r, 154 T.C. 1, 14–15 (2020). But it reasoned that the letter here lacks “the high degree of concreteness and formality that we associate with a determination for purposes of section 6751.” Swift, 2024 WL 378671, at *28. The government agrees with the tax court that supervisory approval was timely under the statute, but disagrees as to why. It advocates for the interpretation of the statute every circuit to address the issue has adopted: Supervisory approval is required before assessment, or if earlier, before the supervisor loses discretion.⁸

The Swifts present compelling reasons why the letter constitutes a formal communication under the tax court's interpretation.⁹ To resolve this dispute, we must interpret the statute to determine when supervisory approval was required, an issue of first impression in this circuit. We conduct statutory interpretation of the Internal Revenue Code de novo. PBBM-Rose Hill, Ltd. v. Comm'r, 900 F.3d 193, 200 (5th Cir. 2018).

Section 6751(b)(1) of the Internal Revenue Code provides:

No penalty under this title shall be assessed unless the initial determination of such assessment is personally approved (in writing) by the immediate supervisor of the individual making such determination or such higher level official as the Secretary may designate.

The text requires supervisory approval of the “initial determination of such assessment.” § 6751(b)(1) (emphasis added). The tax court interpreted “initial determination” to mean a formal communication notifying the taxpayer of a definite decision to assert penalties. Swift, 2024 WL 378671, at *27 (citing Belair Woods, 154 T.C. at 14–15). The problem with this interpretation “is that it has no basis in the text of the statute.” Laidlaw's Harley Davidson Sales, Inc. v. Comm'r, 29 F.4th 1066, 1072 (9th Cir. 2022). “The statute does not make any reference to the communication of a proposed penalty to the taxpayer, much less a ‘formal’ communication.” Id.; see also Kroner v. Comm'r, 48 F.4th 1272, 1277–78 (11th Cir. 2022) (finding “the term ‘initial determination of such assessment’ has nothing to do with communication”).

The statute further provides that no penalty “shall be assessed unless the initial determination of such assessment” is approved. § 6751(b)(1) (emphasis added). “Assessed” refers to “ ‘the formal recording of a taxpayer's tax liability on the tax rolls,’ which is ‘the last of a number of steps required before the IRS can collect’ a tax or penalty from a taxpayer.” Laidlaw's, 29 F.4th at 1071 (quoting Chai v. Comm'r, 851 F.3d 190, 218 (2d Cir. 2017)). And “unless” makes the assessment of penalties conditional upon supervisory approval, but it does not impose a requirement that supervisory approval be obtained at any particular time before assessment. See Kroner, 48 F.4th at 1278.

Still, “[i]f supervisory approval is to be required at all, it must be the case that the approval is obtained when the supervisor has the discretion to give or withhold it.” Chai, 851 F.3d at 220; see also Laidlaw's, 29 F.4th at 1071; Minemyer v. Comm'r, No. 21-9006, 2023 WL 314832, at *5 (10th Cir. Jan. 19, 2023); but see Kroner, 48 F.4th at 1279 n.1 (declining to “address this nuance because it [was] undisputed that the supervisor had discretion”). Taken together, we are convinced that “§ 6751(b)(1) requires written supervisory approval before the assessment of the penalty or, if earlier, before the relevant supervisor loses discretion whether to approve the penalty assessment.” Laidlaw's, 29 F.4th at 1074. That requirement was met here, because supervisor approval was obtained before the issuance of a deficiency notice.¹⁰

We recognize that this interpretation of the statute may be less effective at “prevent[ing] IRS agents from threatening unjustified penalties to encourage taxpayers to settle.” Chai, 851 F.3d at 219 (citing S. Rep. No. 105-174, at 65 (1998)); see also Laidlaw's, 29 F.4th at 1073. But it still “incentivizes supervisory involvement at an early stage in the negotiation process and disincentivizes agents from proposing improper penalties solely for the sake of negotiations.” Kroner, 48 F.4th at 1280. And by requiring approval before assessment, the statute also serves as “a check on the imposition of erroneous penalties.” Id. at 1279.

In sum, we agree that failure to obtain approval before sending the letter did not violate § 6751(b)(1), but for different reasons than those given by the tax court.¹¹

2. Reasonable Cause and Substantial Authority Defenses

The code provides “the taxpayer a complete defense to both the negligence and understatement penalties ․ if the taxpayer acted ‘in good faith’ and with ‘reasonable cause.’ ” Bemont Investments, L.L.C. ex rel. Tax Matters Partner v. United States, 679 F.3d 339, 346 (5th Cir. 2012) (quoting I.R.C. § 6664(c)(1)), abrogated on other grounds by United States v. Woods, 571 U.S. 31, 134 S.Ct. 557, 187 L.Ed.2d 472 (2013). The code also provides “a defense to the substantial understatement penalty ‘if there is or was substantial authority for such treatment.’ ” Id. (quoting I.R.C. § 6662(d)(2)(B)). On appeal, the Swifts argue the tax court erred in denying both defenses.

a. Reasonable Cause

“An accuracy-related penalty does not apply to any portion of a taxpayer's underpayment for which the taxpayer had ‘reasonable cause’ and acted in good faith.” Ray v. Comm'r, 13 F.4th 467, 482 (5th Cir. 2021) (citing I.R.C. § 6664(c)(1)). “The taxpayer bears the burden of proving entitlement to the reasonable cause and good faith defense.” Id. “The assessment of whether the taxpayer has met this burden is ‘made on a case-by-case basis, taking into account all pertinent facts and circumstances.’ ” Id. (quoting Treas. Reg. § 1.6664-4(b)(1)). “The most important factor is the extent of the taxpayer's effort to assess his proper liability in light of all the circumstances.” Id. “The tax court's determinations regarding the taxpayer's eligibility for a reasonable cause and good faith defense are factual findings reviewed for clear error.” Id. at 482–83.

“[A] taxpayer's reliance on the advice of a professional may constitute reasonable cause and good faith where the taxpayer proves by a preponderance of the evidence that: (1) the taxpayer reasonably believed that the professional upon whom the reliance was placed was an independent, competent adviser, without a conflict of interest, and with sufficient expertise to justify reliance; (2) the taxpayer provided all necessary and accurate information to the adviser; and (3) the taxpayer actually relied in good faith on the adviser's judgment.” Nevada Partners Fund, L.L.C. ex rel. Sapphire II, Inc. v. U.S. ex rel. I.R.S., 720 F.3d 594, 617 (5th Cir. 2013), vacated on other grounds, 571 U.S. 1119, 134 S.Ct. 903, 187 L.Ed.2d 767 (2014). Courts regularly find taxpayers cannot reasonably rely on advisors that act as promoters of a transaction and therefore possess an inherent conflict of interest. See, e.g., 106 Ltd. v. Comm'r, 684 F.3d 84, 90–91 (D.C. Cir. 2012) (explaining advisor's role “in promoting, implementing, and receiving fees from the ․ strategy is more than sufficient to support the tax court's finding that he was a promoter and therefore possessed an inherent conflict of interest”); Stobie Creek Invs. LLC v. United States, 608 F.3d 1366, 1382–83 (Fed. Cir. 2010).

The tax court found that the Swifts failed to show reasonable cause because Swift could not reasonably rely on advice from Clark, a primary promoter of the transaction. The Swifts do not directly challenge this reasoning on appeal, but argue that Swift also relied on other professionals such as Rosenbach, Swift's accountant, and KPMG. As the government notes, “there is no evidence showing that [Swift] ever communicated with Rosenbach; that [Swift's] accountant had any expertise in insurance; or that [Swift] discussed the captive arrangements with KPMG beyond the premiums for the tail policies.” Therefore the tax court did not clearly err in finding the defense did not apply.

b. Substantial Authority

“[I]n order for there to be substantial authority, the weight of the authorities supporting treatment of an item must be substantial in relation to the weight of those supporting contrary treatment.” Chemtech Royalty Assocs., L.P. v. United States, 823 F.3d 282, 290 (5th Cir. 2016) (citing Treas. Reg. § 1.6662–4(d)(3)(i)). “For purposes of the substantial-authority inquiry, an authority's weight ‘depends on its relevance and persuasiveness, and the type of document providing the authority.’ ” Id. (quoting Treas. Reg. § 1.6662–4(d)(3)(ii)). “An authority is not particularly relevant if it is materially distinguishable on its facts.” Id. (citing Treas. Reg. § 1.6662–4(d)(3)(ii)). “There may be substantial authority for the tax treatment of an item despite the absence of certain types of authority.” Treas. Reg. § 1.6662–4(d)(3)(ii). Where the facts are essentially undisputed, the applicability of the substantial authority defense is a question of law reviewed de novo. Chemtech Royalty, 823 F.3d at 287.

The tax court declined to consider the Swifts' substantial authority argument, reasoning that they did not raise the argument until their answering brief. We find that even if the Swifts preserved the argument, they failed to show substantial authority because the IRS letter rulings they rely on are materially distinguishable.

As substantial authority, the Swifts rely on a string of IRS private letter rulings that found adequate risk distribution where captive insurers derived at least 30% of their premiums from reinsurance pools. See, e.g., I.R.S. P.L.R. 201224018 (June 15, 2012); I.R.S. P.L.R. 201219009 (May 11, 2012); I.R.S. P.L.R. 201219010 (May 11, 2012); I.R.S. P.L.R. 201219011 (May 11, 2012). For example, one ruling found that where an insurer “contributes a substantial amount of its direct consideration (reinsured from its insured) and associated risks to the pool,” and “receives a quota share of the consideration and associated risks from the pool,” the “arrangement achieves adequate risk shifting and risk distribution” to constitute insurance for tax purposes. I.R.S. P.L.R. 201219011.

The rulings may provide substantial support for the use of reinsurance pools generally, and the 30% threshold. But the issue here was whether these particular reinsurance pools, consisting of Clark-affiliated captives, were bona fide insurers. We agree with the government that the rulings are materially distinguishable because “the private letter rulings presuppose that the insurance arrangement discussed were otherwise bona fide.” In fact, the particular ruling the Swifts cite assumes that “[a]ll insureds use recognized actuarial techniques, based, in part, on commercial rates for similar coverage, to determine the premiums charged to an individual insured.” I.R.S. P.L.R. 201219011. Accordingly, the Swifts failed to establish the defense of substantial authority, and the tax court did not err in sustaining the penalties.

III. Conclusion

For the foregoing reasons, the decision of the tax court is AFFIRMED.

FOOTNOTES

1.   $1.2 million was the cap in effect for the years at issue in this case. I.R.C. § 831 (2012). Since then, the statute has been amended to increase the cap to $2.2 million, and to include certain diversification requirements not applicable here. See I.R.C. § 831; Pub. L. No. 114-113.

2.   The notices of deficiency also included a 40% accuracy-related penalty which the Commissioner has since conceded.

3.   The Swifts also argue that the tax court erred in stating it would “sustain the Commissioner's determinations with respect to the Swift captives, so the Swift captives must recognize the premiums they received as income for the years at issue,” Swift, 2024 WL 378671, at *26, because the Commission did not make any determinations with respect to the Captives, and the Captives were not parties to this case. Because the tax court's decisions for each year are properly limited to determining the Swifts' liabilities, we do not address this argument further.

4.   The Swifts also argue that 199 physicians is comparable to the 200 stock brokers the tax court found sufficient in Estate of Moyer v. Comm'r, 32 T.C. 515 (1959). That case contains little discussion on risk distribution, and we do not read it to support the idea that 199 exposures is per se sufficient.

5.   The nine factors are (1) whether the company was created for legitimate nontax reasons; (2) whether there was a circular flow of funds; (3) whether the entity faced actual and insurable risk; (4) whether the policies were arm's-length contracts; (5) whether the entity charged actuarially determined premiums; (6) whether comparable coverage was more expensive or even available; (7) whether it was subject to regulatory control and met minimum statutory requirements; (8) whether it was adequately capitalized; and (9) whether it paid claims from a separately maintained account.

6.   For example, the pool's loss ratio, or the percentage of premiums spent on claims, was 8% as compared to a 66% industry standard in 2015.

7.   The tax court applies this interpretation of the statute in circuits, like ours, that have not held otherwise.

8.   The Ninth and Tenth Circuit have adopted this interpretation of the statute. Laidlaw's Harley Davidson Sales, Inc. v. Comm'r, 29 F.4th 1066, 1074 (9th Cir. 2022); Minemyer v. Comm'r, No. 21-9006, 2023 WL 314832, at *5 (10th Cir. Jan. 19, 2023) (holding the statute requires approval before the issuance of a notice of deficiency, when discretion is lost). The Eleventh Circuit has held the statute requires approval before assessment but has not addressed the nuance around supervisor discretion. Kroner v. Comm'r, 48 F.4th 1272, 1276, 1279 n.1 (11th Cir. 2022). The Second Circuit has also held that the statute requires approval before the issuance of a deficiency notice, when discretion is lost, but arguably left open the question of whether approval might be required sooner. Chai v. Comm'r, 851 F.3d 190, 220–21 (2d Cir. 2017).

9.   On one hand, as the tax court emphasized, the letter stated that given the statute of limitations, “it will be necessary to move the case forward for the possible issuance of a Statutory Notice of Deficiency” and that “Internal Revenue Service District Counsel will review the issue and will make a determination as to whether a Statutory Notice of Deficiency will be issued.” (emphasis in original). On the other hand, as the Swifts emphasize, the letter stated the agent had completed his review, his “report includes the 20% accuracy related penalty,” and a Form 870 “for an agreement of the tax and [20%] penalties at the Examination level” was enclosed. The letter also mentioned consideration of a 40% penalty, but stated “approval has not been given” so the increase “has not been included,” suggesting that approval had been given for the 20% penalty.

10.   Once a notice of deficiency is sent, “the Commissioner begins to lose discretion over whether the penalty is assessed,” because the Internal Revenue Code provides that the deficiency “shall be assessed” if the taxpayer does not file a petition with the tax court within the required time period. Laidlaw's, 29 F.4th at 1071 n.4 (citing I.R.C. § 6213(c)).

11.   To the extent the Swifts argue the government did not comply with the statute even under the interpretation we adopt, we are unconvinced. The Swifts assert “[t]he government provided no evidence regarding the timing or identities of the individuals who actually were authorized to make the initial determination and provide the subsequent supervisory approval.” This misunderstands the tax court's reasoning and our own interpretation of the statute. While the letter itself was not an initial determination requiring approval under the statute, it does not follow that the agent sending the letter was not authorized to make the initial determination.

King, Circuit Judge: