DRAKE PLASTICS LTD CO v. INTERNAL REVENUE SERVICE (2026)

MEMORANDUM AND OPINION

This dispute is about a tax disclosure rule for captive insurers. The defendants (collectively referred to as the “IRS” or the “agency”) justify the rule by the potential for tax evasion that captive-insurance arrangements present. The plaintiffs, a company, a related captive insurer, and their tax-advising firm, do not want to make the disclosures and contend that the rule exceeds the IRS's authority, lacks support in the administrative record, and is arbitrary and capricious. The IRS has responded to each argument and submitted the administrative record. The court held a hearing to allow the parties to argue their positions.

Based on the cross-motions for summary judgment, the record, the arguments of counsel, and the applicable law, the court grants in part the plaintiffs' motion for summary judgment and a permanent injunction, (Docket Entry No. 58), and grants in part the IRS's cross-motion for summary judgment, (Docket Entry No. 63). The court finds that the IRS: (1) appropriately designated micro-captive transactions as transactions of interest through 26 C.F.R. § 1.6011-11; but (2) cannot justify on the current record the designation of micro-captive transactions as listed transactions through 26 C.F.R. § 1.6011-10. The court therefore vacates 26 C.F.R. § 1.6011-10.

The reasons for these rulings are set out below.

I. Background

Captive insurance transactions allow corporate entities to claim payments to their affiliated insurers as a business expense under 26 U.S.C. § 162(a). If the affiliated insurer receives premium payments below a statutory cap, that insurer can elect tax benefits under § 831(b) of the Internal Revenue Code. Captives that elect § 831(b) benefits are commonly called “micro-captives.”

Although some micro-captives are legitimate insurers, some are not. In a typical, if oversimplified, micro-captive transaction, Company A owns a separate but related Company B whose identified purpose is to insure the risks of Company A. When Company A pays insurance premiums to Company B, it can deduct those payments as a § 162(a) business expense. If Company B makes a § 831(b) election, it does not need to include those payments in its taxable income. As a result, neither Company A nor Company B pays an immediate tax on the money passed through both companies: Company A gets the business-expense treatment, and Company B gets the election. Company A benefits because Company B can return the non-taxed income to Company A, by making a loan, by issuing a dividend (taxed at a lower rate), or by using the income for Company A's benefit. These transactions may be legitimate, but the potential for tax avoidance or evasion is apparent.

To ensure that micro-captives like Company B are legitimate insurance companies and not tax-avoidance entities, the Secretary of the Treasury promulgated, after notice-and-comment, a rule governing certain micro-captive arrangements. See Micro-Captive Listed Transactions and Micro-Captive Transactions of Interest, 90 Fed. Reg. 3,534 (Jan. 14, 2025) (codified at 26 C.F.R. § 1.6011-10 and § 1.6011-11) [hereinafter the Final Rule]. The Final Rule designates two categories of micro-captive transactions, each with its own reporting criteria: (1) micro-captive “transaction[s] of interest,” 26 C.F.R. § 1.6011-11, which are “of a type” that has “a potential for tax avoidance or evasion,” 26 U.S.C. § 6707A(c)(1); see 26 C.F.R. § 1.6011-4(b)(6) (defining “transactions of interest” as “reportable transactions”); and (2) “listed” micro-captive transactions, 26 C.F.R. § 1.6011-10, which are “a tax avoidance transaction for purposes of section 6011,” 26 U.S.C. § 6707A(c)(2). Listed micro-captive transactions are a subset of reportable transactions. There is a higher penalty for a failure to report listed transactions, because they are more likely to be tax avoidant than are micro-captive transactions of interest.

Drake Plastics Ltd. Co. is a Houston limited liability company that specializes in extruding, injection molding, postprocessing, and machining ultra-high-performance polymers. (Docket Entry No. 1 ¶ 15). Drake Insurance Co. is a captive insurance entity that provides workers' compensation, product liability, errors and omissions, and other types of insurance for Drake Plastics and over thirty other related entities. (Id. ¶ 16). Strategic Risk Alternatives, LLC creates and manages micro-captive insurance plans for closely held businesses. (Id. ¶ 24). The plaintiffs sued the Internal Revenue Service and federal officials to challenge the Final Rule, arguing that compliance is costly and burdensome and that labelling micro-captive insurance as presumptively tax-avoidant scares clients away from using a micro-captive to achieve even legitimate tax benefits. (See id. ¶¶ 10, 11, 18). The plaintiffs moved for a preliminary injunction, (Docket Entry No. 17), which the court deferred on the limited record then presented. The court required the parties to present the issues in cross-motions for summary judgment, accompanied by the administrative record. (Docket Entry No. 33). These motions and attachments followed.

A. Captive Insurance and the Internal Revenue Code

“A micro-captive transaction is typically an insurance agreement between a parent company and a ‘captive’ insurer under its control.” CIC Servs., LLC v. IRS, 593 U.S. 209, 213 (2021). Fortune 500 companies, as well as many other businesses, rely on captive insurance. (Docket Entry No. 51 at AR22100). Captive-insurance arrangements allow insureds to tailor coverage to specific needs and enable insurance against low-frequency, high severity risks that are “extremely difficult or even impossible to place in traditional markets.” (Docket Entry No. 38-2 at AR5283). Because captive insurance provides a baseline level of insurance for companies, it may also enable “access to the reinsurance market,” ensuring a “sound program[ ]” for specialized risks. (Id.). Some commenters have noted the positive downstream effects of the captive-insurance market, including promoting risk-management and safe-workplace practices by companies that are responsible for their affiliates' risks, business development spurred by tax savings and lower premiums, and a bulwark against rare but potentially devastating “black-swan” risks. (Docket Entry No. 37-1 at AR4413, 4463, 4467, 4601; Docket Entry No. 38-2 at AR5282).

The Internal Revenue Code does not define insurance. F.W. Servs., Inc. & Subsidiaries v. Comm'r, 459 F. App'x 389, 392 (5th Cir. 2012) (per curiam); Swift v. Comm'r, 144 F.4th 756, 764 (5th Cir. 2025). But the Code and Treasury regulations account for insurance in myriad ways. For example, and as relevant here, insurance premiums qualify as an “ordinary and necessary expense[ ] paid or incurred ․ in carrying on any trade or business.” 26 U.S.C. § 162(a); see 26 C.F.R. § 1.162-1(a) (including “insurance premiums against fire, storm, theft, accident, or other similar losses in the case of a business”). In the absence of a statutory definition, case law definitions use the common features of insurance: “risk-shifting and risk-distributing.” Helvering v. Le Gierse, 312 U.S. 531, 539 (1941). Captive insurance schemes operate under the usual rules of insurance set out in the Code and its regulations. See CIC Serv., 593 U.S. at 212–13 (explaining how micro-captive transactions fit within the usual tax rules).

Although the Code does not define “captive” or “insurance,” it does address micro-insurance. In 1986, Congress enacted § 831(b) to help small insurers. The Code allows any non-life insurance company that has “net written premiums (or, if greater, direct written premiums) for the taxable year” that “do not exceed $1,200,000” to elect under § 831(b) to be taxed on only its taxable investment income. See Tax Reform Act of 1986, Pub. L. No. 99-514, § 1024(a), 100 Stat. 2085, 2405 (1986). Section 831(b) simplifies taxes for small insurance companies by eliminating the distinction between mutual companies and other companies, extending “the benefit of the small company provision to all eligible small companies, whether stock or mutual.” (Docket Entry No. 38-2 at AR5657; see also id. at AR5667). Section 831(b) applies to all non-life insurance companies, whether captive or not.

In 2015, Congress amended § 831(b) through the Protecting Americans from Tax Hikes (“PATH”) Act. Pub. L. No. 114-113, § 333(a), 129 Stat. 2242, 3106 (2015). To qualify for § 831(b) benefits under current law, a captive insurer must: (1) not exceed limits on “net written premiums (or, if greater, direct written premiums) for the taxable year” ($2,200,000 after inflation adjustments); (2) meet certain diversification requirements; and (3) elect to apply § 831(b) for that taxable year. 26 U.S.C. § 831(b)(2); see CIC Servs., LLC v. IRS, No. 3:25-CV-146, 2026 WL 622836, at *1 (E.D. Tenn. Mar. 5, 2026). A captive insurer is sufficiently diversified if:

(I) no more than 20 percent of the net written premiums (or, if greater, direct written premiums) of such company for the taxable year is attributable to any one policyholder, or

(II) such insurance company does not meet the requirement of subclause (I) and no person who holds (directly or indirectly) an interest in such insurance company is a specified holder who holds (directly or indirectly) aggregate interests in such insurance company which constitute a percentage of the entire interests in such insurance company which is more than a de minimis percentage higher than the percentage of interests in the relevant specified assets with respect to such insurance company held (directly or indirectly) by such specified holder.

26 U.S.C. § 831(b)(2)(B)(i)(I)–(II). Congress added the diversification requirements to prevent the “abuse of captive insurance companies for estate planning purposes.” S. REP. NO. 114-16, at 3 (2015).

The tax benefits of captive insurance arrangements also present a risk of abuse. The benefits provide taxpayers with an incentive and opportunity to characterize their business arrangements as insurance, even when they are not, to avoid taxes that they would otherwise have to pay. See Charlene D. Luke, Captivating Deductions, 46 HOFSTRA L. REV. 855, 857 (2018) (“As a result of the difference in treatment between savings and premiums, businesses have an incentive to create structures that take the form of insurance but that have the substantive features of an ordinary investment.”). If a claimed insurance transaction is “really for insurance,” then it does not get taxed as a matter of “congressional choice.” CIC Servs., 593 U.S. at 213. But if the transaction “is a sham, which the affiliated companies have entered into only to escape tax liability,” then “no tax benefit should accrue.” Id.

Captive-insurance arrangements pose a potential for abuse because they combine the separate benefits that Congress enacted in §§ 162(a) and 831(b). “The insured party can deduct its premium payments as business expenses,” and “the insurer can exclude up to $2.2 million of those premiums from its own taxable income.” Id. “The result is that the money does not get taxed at all,” id., and can be reused by the insurer for the insured's financial benefit.

Unsurprisingly, in the last decade, the IRS has found that some captive insurance schemes are not “really for insurance.” Id. The IRS has consistently convinced tax and appellate courts that claimed captive and micro-captive insurance benefits under §§ 162(a) and 831(b) were illegitimate. See, e.g., Rsrv. Mech. Corp. v. Comm'r, 34 F.4th 881, 916 (10th Cir. 2022); Swift, 144 F.4th at 769; Avrahami v. Comm'r, 149 T.C. 144, 197 (2017); Syzygy Ins. Co. v. Comm'r, 117 T.C.M. (CCH) 1165 (T.C. 2019); Caylor Land & Dev., Inc. v. Comm'r, 121 T.C.M. (CCH) 1205 (T.C. 2021); Keating v. Comm'r, T.C.M. (RIA) 2024-002 (T.C. 2024); Patel v. Comm'r, T.C.M. (RIA) 2024-034 (T.C. 2024); Royalty Mgmt. Ins. Co., Ltd. v. Comm'r, T.C.M. (RIA) 2024-087 (T.C. 2024), appeal dismissed, No. 24-9006, 2025 WL 512503 (10th Cir. Jan. 10, 2025); Jones v. Comm'r, T.C.M. (RIA) 2025-025 (T.C.), opinion clarified, T.C.M. (RIA) 2025-078 (T.C. 2025); Kadau v. Comm'r, T.C.M. (RIA) 2025-081 (T.C. 2025); CFM Ins., Inc. v. Comm'r, T.C.M. (RIA) 2025-083 (T.C. 2025); see also Stearns-Roger Corp. v. United States, 774 F.2d 414, 416–17 (10th Cir. 1985); Beech Aircraft Corp. v. United States, 797 F.2d 920, 923 (10th Cir. 1986); Clougherty Packing Co. v. Comm'r, 811 F.2d 1297, 1307 (9th Cir. 1987); Humana Inc. v. Comm'r, 881 F.2d 247, 251–52 (6th Cir. 1989); Gulf Oil Corp. v. Comm'r, 914 F.2d 396, 411–12 (3d Cir. 1990).¹

To gain more information about potentially illegitimate micro-insurance transactions, the Department of the Treasury promulgated the Final Rule's disclosure requirements. This challenge followed.

B. The Code's Disclosure Requirements

The federal tax system “is based on a system of self-reporting.” United States v. Bisceglia, 420 U.S. 141, 145 (1975). “Congress delegated to the Secretary of the Treasury—acting through the Internal Revenue Service—the authority to collect information and prescribe regulations as necessary to assess and collect federal taxes.” Green Rock LLC v. IRS, 104 F.4th 220, 222 (11th Cir. 2024) (Pryor W., C.J.). Taxpayers must file “return[s] or statement[s]” as required by “regulations prescribed by the Secretary.” 26 U.S.C. § 6011(a).

“In response to the proliferation of certain corporate tax sheltering strategies, the Secretary designed a comprehensive disclosure regime—a ‘reportable transaction’ regime—to target those shelters and ferret out improper tax-avoidance transactions.” Green Rock, 104 F.4th at 222; see 26 C.F.R. § 1.6011-4. Because taxpayers did not sufficiently comply with the reportable-transaction requirements, Congress passed—and the President signed—section 6707A of the Code, which provides civil penalties for violators. See American Jobs Creation Act of 2004, Pub. L. No. 108- 357, 118 Stat. 1418, 1575 (codified in part at 26 U.S.C. § 6707A); see also S. REP. NO. 108-192, at 90 (2003) (recognizing the need for penalties).

Taxpayers must disclose, under the penalty of § 6707A, “reportable transactions” and “listed transactions.” 26 U.S.C. § 6707A(c). A “reportable transaction” is a transaction that is “of a type” that the Secretary has “determine[d]” has “a potential for tax avoidance or evasion.” Id. § 6707A(c)(1). A “listed transaction” is a “reportable transaction” that the Secretary has “specifically identified” as “a tax avoidance transaction for purposes of section 6011.” Id. § 6707A(c)(2). “Listed transactions are potentially the most abusive transactions.” Green Rock, 104 F.4th at 223. “Listing is intended to provide taxpayers with a ‘bright line’ and ‘clear and objective’ definition of the specific conduct that the [IRS] considers presumptively suspicious and subject to heightened disclosure.” Id. (emphasis added) (quoting Elaine Church & Corina Trainer, Reportable Transactions: A Comprehensive Disclosure Regime to Combat Tax Shelters, 57 MAJOR TAX PLAN. 12-1, 12-8 (2005)). Although the IRS's designation of transactions as “reportable” or “listed” triggers disclosure requirements, those designations do “not affect the legal determination of whether the taxpayer's treatment of the transaction is proper.” 26 C.F.R. § 1.6011-4(a); accord Turnham v. Comm'r, 979 F.3d 1322, 1326 (11th Cir. 2020); Green Rock, 104 F.4th at 223.

The reportable-transaction category includes “transactions of interest.” 26 C.F.R. § 1.6011-4(b)(6). “A transaction of interest is a transaction that is the same as or substantially similar to one of the types of transactions that the IRS has identified by notice, regulation, or other form of published guidance as a transaction of interest.” Id. This definition is “circular,” but it is a “catch-all” that allows the IRS to designate a “reportable transaction,” consistent with its own regulations, “if it believes the transaction has the potential for tax avoidance or evasion.” CIC Servs., LLC v. IRS, No. 3:17-CV-110, 2021 WL 4481008, at *5 (E.D. Tenn. Sept. 21, 2021); see Ryan, LLC v. IRS, No. 3:25-CV-0078-B, 2025 WL 3089415, at *3 (N.D. Tex. Nov. 5, 2025) (equating a “transaction of interest” with transactions “deemed potentially abusive”).

C. The Final Rule

In 2016, through Notice 2016-166, the Treasury identified certain § 831(b) captive transactions as transactions of interest. CIC Servs., 593 U.S. at 213–14. A material advisor—a professional who helps carry out reportable transactions and benefits from such aid—challenged the Notice under the Administrative Procedure Act. That challenge went to the Supreme Court on the question whether the Anti-Injunction Act barred the challenge. See id. The Court held that the case could proceed, and, on remand, the district court vacated the Notice for failure to comply with the APA's notice-and-comment procedures. See CIC Servs., LLC v. IRS, 592 F. Supp. 3d 677, 688 (E.D. Tenn. 2022). The district court also held that the administrative record did not include sufficient “facts and data showing that micro-captive insurance arrangements have a potential for tax avoidance or evasion.” Id. at 687.

In response to this ruling, Treasury went through notice-and-comment rulemaking and promulgated a Final Rule similar to, but with differences from, Notice 2016-166. The Final Rule designates two sets of micro-transactions, each with their own reporting criteria: (1) micro-captive “transaction[s] of interest,” 26 C.F.R. § 1.6011-11; and (2) “listed” micro-captive transactions, 26 C.F.R. § 1.6011-10. Treasury “developed ․ objective factors to ensure administrability and clarity for taxpayers.” Final Rule at 3,535. Three are relevant to this case: the Relationship Test; the Financing Factor; and the Loss-Ratio Factor. (See Docket Entry No. 59 at 13).

The Relationship Test puts the “captive” in the “micro-captive” regulations. The regulation defines a “captive” in part as an entity for which “[a]t least 20 percent of the entity's assets or the voting power or value of its outstanding stock or equity interests is directly or indirectly owned, individually or collectively, by an Insured, an Owner, or persons Related to an Insured or an Owner.” 26 C.F.R. §§ 1.6011-10(b)(1)(iii); see also id. § 1.6011-11(b)(1) (“Captive has the same meaning as provided in § 1.6011-10(b)(1).” (emphasis omitted)). A captive must disclose its claimed insurance transaction only if at least 20 percent of its assets or equity is owned by an insured or related entity. The Final Rule includes this requirement to exclude arrangements less likely to be tax avoidant, such as small mutual insurers with diversified ownership that are traditionally held by their members in a given geographical location solely to protect their own property and are not for profit. See Final Rule at 3,555 (using “groups of farmers and fire associations” as examples).

The Financing Factor asks whether the captive insurer is funneling the non-taxed premium payments back to the affiliated insurer. The Financing Factor is triggered if the captive “made available as financing or otherwise conveyed ․ to” the insured, the insured's owners, or related entities “in a transaction that did not result in taxable income or gain” “any portion of the amounts” the captive earned from insurance contracts. 26 C.F.R. § 1.6011–10(c)(1)(i)–(ii). The Financing Factor requires disclosure when “amounts paid as premiums have not only avoided ordinary taxation but have continued to avoid tax while back in the hands of the related parties who caused the premiums to be paid and deducted.” Final Rule at 3,546. In these cases, the parties can “generate nearly $1.2 million in tax deductions while arguably only moving money from one pocket to another.” Avrahami, 149 T.C. at 179.

The Loss-Ratio Factor “measures whether the amount of liabilities incurred for insured losses and claims administration expenses is significantly less than the amount of premiums earned, adjusted for policyholder dividends.” Final Rule at 3,540. The loss ratio is calculated over a ten-year period. See id. §§ 1.6011-10(b)(2)(ii), 1.6011-11(b)(2)(ii)(A). A transaction with a captive “that does not have at least ten taxable years cannot” be a listed micro-captive transaction. Id. § 1.6011-10(b)(2)(ii). It is classified instead as a micro-captive transaction of interest, and the loss ratio is based on “all taxable year(s) of the [c]aptive.” Id. § 1.6011-11(b)(2)(ii)(B). The loss ratio is important because if an arrangement is tax avoidant (and therefore not really insurance), a low loss ratio is expected. A low loss ratio results when high premiums are paid into the captive (to get the tax benefits) and “low amounts of claims [are] paid out from captive to insured (because the captive is not insuring legitimate risks or not paying claims as an insurance company would).” (Docket Entry No. 63-1 at 20). A listed micro-captive transaction has a loss ratio of “less than 30 percent,” 26 C.F.R. § 1.6011-10(c)(2)(i)–(ii), and a micro-captive transaction of interest has a loss ratio of “less than 60 percent,” id. § 1.6011-11(c)(2), (e)(1).

Combining these factors results in two distinct tests. First, a captive transaction is a “listed transaction” if the captive elects § 831(b) taxation and the transaction satisfies the Relationship Test, the Financing Factor, and the Loss-Ratio Factor (with an adjusted loss ratio of 30 percent). See id. § 1.6011-10(b)–(c). Second, a captive transaction is a “reportable transaction” or a “transaction of interest” if the captive elects § 831 tax treatment, satisfies the Relationship Test, and satisfies either the Financing Factor or Loss-Ratio Factor (with an adjusted loss ratio of 60 percent). See 26 C.F.R. § 1.6011-11(b)–(c).

Challenges to the Final Rule followed in the Eastern District of Tennessee; the Northern District of Texas; and, here, in the Southern District of Texas. The district court in the Eastern District of Tennessee recently granted summary judgment in the government's favor, holding that the Treasury neither exceeded its statutory authority nor acted arbitrarily or capriciously in enacting the Final Rule. CIC Servs., LLC v. IRS, No. 3:25-CV-146, 2026 WL 622836, at *9 (E.D. Tenn. Mar. 5, 2026). The district court in the Northern District of Texas, on the government's motion to dismiss, held that the Final Rule did not exceed Treasury's statutory authority but ruled that the plaintiff adequately alleged that the Final Rule is arbitrary and capricious. Ryan, 2025 WL 3089415, at *14. That litigation is ongoing.

In this court, the plaintiffs have moved for summary judgment, arguing that the Final Rule exceeds the Treasury Department's statutory authority and that the Final Rule is arbitrary and capricious. (Docket Entry No. 59). The plaintiffs have also moved to vacate the Final Rule and for a permanent injunction. (See id. at 33–35). The defendants have cross-moved for summary judgment. (Docket Entry No. 63).

II. The Legal Standard

A. Summary Judgment

“Summary judgment is required when ‘the movant shows that there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.’ ” Trent v. Wade, 776 F.3d 368, 376 (5th Cir. 2015) (quoting FED. R. CIV. P. 56(a)). “A genuine dispute of material fact exists when the ‘evidence is such that a reasonable jury could return a verdict for the nonmoving party.’ ” Nola Spice Designs, LLC v. Haydel Enters., Inc., 783 F.3d 527, 536 (5th Cir. 2015) (quoting Anderson v. Liberty Lobby, 477 U.S. 242, 248 (1986)). “The moving party ‘bears the initial responsibility of informing the district court of the basis for its motion, and identifying those portions of [the record] which it believes demonstrate the absence of a genuine issue of material fact.’ ” Id. (alterations in original) (quoting EEOC v. LHC Grp., Inc., 773 F.3d 688, 694 (5th Cir. 2014)); see also Celotex Corp. v. Catrett, 477 U.S. 317, 323 (1986).

If the burden of proof at trial lies with the nonmoving party, the movant may satisfy its initial burden by showing an absence of evidence to support the nonmoving party's case. Fret v. Melton Truck Lines, Inc., 706 Fed. App'x. 824, 827 (5th Cir. 2017) (per curiam) (quoting Lindsey v. Sears Roebuck & Co., 16 F.3d 616, 618 (5th Cir. 1994)). Although the party moving for summary judgment must demonstrate the absence of a genuine issue of material fact, it does not need to negate the elements of the nonmovant's case. Coastal Agric. Supply, Inc. v. JP Morgan Chase Bank, N.A., 759 F.3d 498, 505 (5th Cir. 2014) (citing Boudreaux v. Swift Transp. Co., 402 F.3d 536, 540 (5th Cir. 2005)). A fact is material if “its resolution could affect the outcome of the action.” Aly v. City of Lake Jackson, 605 Fed. App'x 260, 262 (5th Cir. 2015) (per curiam) (citing Burrell v. Dr. Pepper/Seven UP Bottling Grp., Inc., 482 F.3d 408, 411 (5th Cir. 2007)). “If the moving party fails to meet [its] initial burden, the motion [for summary judgment] must be denied, regardless of the nonmovant's response.” Pioneer Exploration, LLC v. Steadfast Ins. Co., 767 F.3d 503, 511 (5th Cir. 2014) (quoting Kee v. City of Rowlett, 247 F.3d 206, 210 (5th Cir. 2001)).

“When the moving party has met its Rule 56(c) burden, the nonmoving party cannot survive a summary judgment motion by resting on the mere allegations of its pleadings.” Duffie v. United States, 600 F.3d 362, 371 (5th Cir. 2010). The nonmovant must identify specific evidence in the record and articulate how that evidence supports that party's claim. Willis v. Cleco Corp., 749 F.3d 314, 317 (5th Cir. 2014). “This burden will not be satisfied by ‘some metaphysical doubt as to the material facts, by conclusory allegations, by unsubstantiated assertions, or by only a scintilla of evidence.’ ” Jurach v. Safety Vision, LLC, 642 Fed. App'x 313, 317 (5th Cir. 2016) (quoting Boudreaux, 402 F.3d at 540). In deciding a summary judgment motion, the court draws all reasonable inferences in the light most favorable to the nonmoving party. Darden v. City of Fort Worth, 866 F.3d 698, 702 (5th Cir. 2017).

When the parties cross-move for summary judgment, the court must review “each motion independently, viewing the evidence and inferences in the light most favorable to the nonmoving party.” Mid–Continent Cas. Co. v. Bay Rock Operating Co., 614 F.3d 105, 110 (5th Cir. 2010) (cleaned up).

B. Review of Agency Decisions

“[W]hen a party seeks review of agency action under the APA, the district judge sits as an appellate tribunal.” Am. Bioscience, Inc. v. Thompson, 269 F.3d 1077, 1083 (D.C. Cir. 2001). “The entire case on review is a question of law.” Id. (cleaned up). “Under the APA, it is the role of the agency to resolve factual issues to arrive at a decision that is supported by the administrative record, whereas ‘the function of the district court is to determine whether or not as a matter of law the evidence in the administrative record permitted the agency to make the decision it did.’ ” Stuttering Found. of Am. v. Springer, 498 F. Supp. 2d 203, 207 (D.D.C. 2007) (quoting Occidental Eng'g Co. v. INS, 753 F.2d 766, 769–70 (9th Cir. 1985)). “Summary judgment thus serves as the mechanism for deciding, as a matter of law, whether the agency action is supported by the administrative record and otherwise consistent with the APA standard of review.” Id.; accord Redeemed Christian Church of God v. USCIS, 331 F. Supp. 3d 684, 694 (S.D. Tex. 2018).

The Administrative Procedure Act directs courts to hold unlawful agency actions that are “in excess of statutory jurisdiction, authority, or limitations, or short of statutory right,” 5 U.S.C. § 706(2)(C), or “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law,” id. § 706(2)(A).

When addressing challenges brought under the statutory-authority limitation, “[c]ourts must exercise their independent judgment in deciding whether an agency has acted within its statutory authority, as the APA requires.” Loper Bright Enters. v. Raimondo, 603 U.S. 369, 412 (2024). The question is whether agency action regulates conduct that the relevant statute does not. See Bondi v. VanDerStok, 604 U.S. 458, 467–68 (2025) (assessing whether the challenged rule regulated more firearms than the statute did, and assuming, without deciding, that the challenged rule could be vacated only if it was “inconsistent on [its] face” with the relevant statute). When the authorizing statute includes an express delegation of discretionary authority, “the question is whether the Rule is within the outer boundaries of that delegation.” Mayfield v. U.S. Dep't of Lab., 117 F.4th 611, 617 (5th Cir. 2024) (citing Loper Bright, 603 U.S. at 403–04).

When addressing challenges brought under the arbitrary-and-capricious standard, courts must ensure that agency action is “reasonable and reasonably explained.” FCC v. Prometheus Radio Project, 592 U.S. 414, 423 (2021). The “well-worn arbitrary-and-capricious standard ensures that an administrative agency ‘examined the relevant data and articulated a satisfactory explanation for its action including a rational connection between the facts found and the choice made.’ ” FDA v. Wages & White Lion Invs., L.L.C., 604 U.S. 542, 567 (2025) (alterations adopted) (quoting Motor Vehicle Mfrs. Ass'n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983)). “The scope of this review ‘is narrow,’ and reviewing courts must exercise appropriate deference to agency decisionmaking and not substitute their own judgment for that of the agency.” Id. (quoting State Farm, 604 U.S. at 43). “A court simply ensures that the agency has acted within a zone of reasonableness and, in particular, has reasonably considered the relevant issues and reasonably explained the decision.” Prometheus, 592 U.S. at 423. With respect to final rules, courts focus on whether “the agency has relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.” State Farm, 463 U.S. at 43.

“The reviewing court should not attempt itself to make up for [any] deficiencies,” as it “may not supply a reasoned basis for the agency's action that the agency itself has not given.” Id. (quoting SEC v. Chenery Corp. (Chenery II), 332 U.S. 194, 196 (1947)). But courts will “uphold a decision of less than ideal clarity if the agency's path may reasonably be discerned.” Bowman Transp., Inc. v. Arkansas-Best Freight Sys., Inc., 419 U.S. 281, 286 (1974) (citing Colo. Interstate Gas Co. v. FPC, 324 U.S. 581, 595 (1945)). It is “relevant that Congress required a record of the rulemaking proceedings to be compiled and submitted to a reviewing court and intended that agency findings under the Act would be supported by substantial evidence on the record considered as a whole.” State Farm, 463 U.S. at 43–44 (cleaned up); see 5 U.S.C. § 706 (“[T]he court shall review the whole record ․”). As a result, “the agency's rationale for” a “policy need not be reiterated fully in the Final Rule itself so long as the rationale may ‘reasonably be discerned’ from the administrative record as a whole.” Nat. Res. Def. Council v. Nat'l Marine Fisheries Serv., 71 F. Supp. 3d 35, 58 (D.D.C. 2014) (Jackson, J.); see, e.g., Alaska Dep't of Env't Conservation v. EPA, 540 U.S. 461, 497 (2004) (relying on “accompanying explanatory correspondence from EPA” to explain its “skeletal orders”).

III. Analysis

The plaintiffs argue that the court must set aside the Final Rule because the IRS exceeded its statutory authority and acted arbitrarily and capriciously in selecting the reporting criteria for micro-captive transactions. (Docket Entry No. 59 at 13). The court considers each argument with respect to both micro-captive “transaction[s] of interest,” 26 C.F.R. § 1.6011-11, and “listed” micro-captive transactions, id. § 1.6011-10.

A. The Exceeds-Statutory-Authority Challenge

The key question on an exceeds-authority challenge is whether agency action regulates conduct that the relevant statute does not. See VanDerStok, 604 U.S. at 467–68. If “the Rule is within the outer boundaries of that delegation,” then the agency has the statutory authority to promulgate it. Mayfield, 117 F.4th at 617 (citing Loper Bright, 603 U.S. at 403–04). The plaintiffs' exceeds-authority challenge presents a mixed question of statutory interpretation and factual findings. Whether the Final Rule is consistent with the Code depends on both the Code's scope and whether the Secretary adequately made the factual findings necessary to place the Final Rule within that scope. To determine the Code's meaning, courts employ “all relevant interpretive tools,” Loper Bright, 603 U.S. at 400, starting with the text, see King v. Burwell, 576 U.S. 473, 486 (2015). To determine whether the Secretary made sufficient factual findings for the Final Rule to fall within the Code's scope, this court must “ensure” that the agency “exercise[d] [its] discretion consistent with the APA.” Loper Bright, 603 U.S at 404.

1. Statutory Interpretation

Taxpayers must file “return[s] or statement[s]” as required by “regulations prescribed by the Secretary.” 26 U.S.C. § 6011(a). Under this section, the IRS “has broad power to require the submission of tax-related information that it believes helpful in assessing and collecting taxes.” CIC Servs., 593 U.S. at 212. To ensure that enforcement penalties back the Secretary's regulations under § 6011(a), the Secretary may designate transactions as “reportable transactions” and “listed transactions.” 26 U.S.C. § 6707A(a), (c). A “reportable transaction” is a transaction that “is of a type” that the Secretary has “determine[d]” has “a potential for tax avoidance or evasion.” Id. § 6707A(c)(1). A “listed transaction” is a “reportable transaction” that the Secretary has “specifically identified” as “a tax avoidance transaction for purposes of section 6011.” Id. § 6707A(c)(2).

A transaction is tax avoidant or evasive if it “attempt[s] to defeat or circumvent the tax law in order to illegally reduce one's tax liability.” Tax Evasion, BLACK'S LAW DICTIONARY (8th ed. 2004). The defining characteristic is that but for the transaction, tax liability would have and should have accrued. See CIC Servs., 593 U.S. at 213 (explaining that a transaction is tax avoidant or evasive because it is “a sham” designed “to escape tax liability”). As relevant to micro-captive transactions, the Code allows deductions for “ordinary and necessary expenses paid or incurred ․ in carrying on any trade or business,” 26 U.S.C. § 162(a), including “insurance,” 26 C.F.R. § 1.162-1(a). The Code provides benefits to qualifying “insurance companies,” 26 U.S.C. § 831(b)(1), which are companies that derive “more than half” their business “during the taxable year” from issuing “insurance or annuity contracts” or “reinsuring ․ risks underwritten by insurance companies,” id. § 816(a)(2); see id. § 831(c) (adopting § 816(a)'s definition of “insurance company”). A micro-captive transaction is avoidant or evasive if it is claimed to be “insurance” or “ordinary and necessary expenses” but is “not really for insurance” or not really an ordinary and necessary expense. See CIC Servs., 593 U.S. at 213.

The common-law definition of “insurance” is based on the traditional characteristics of an insurance transaction. Those are: (1) risk-shifting; (2) risk-distribution; (3) insurance risk; and (4) whether an arrangement looks like commonly accepted notions of insurance. Swift, 144 F.4th at 764. To determine the fourth factor, courts ask: “(1) whether the insurer is organized, operated, and regulated as an insurance company by the States in which it does business; (2) whether the insurer is adequately capitalized; (3) whether the insurance policies are valid and binding; (4) whether the premiums are reasonable in relation to the risk of loss; and (5) whether premiums are duly paid and loss claims are duly satisfied.” R.V.I. Guar. Co. & Subsidiaries v. Comm'r, 145 T.C. 209, 231 (2015); accord Avrahami, 149 T.C. at 191. This analysis calls for case-by-case factual findings on whether a specific transaction constitutes insurance. See, e.g., Swift, 144 F.4th at 769 (the tax court did not err in finding that the micro-captive arrangements were not insurance); Rsrv. Mech. Corp., 34 F.4th at 916 (the tax court “did not clearly err in” finding that the captive “did not operate as an insurance company”).

The Code distinguishes between “reportable transactions” and “listed transactions.” Reportable transactions are “of a type” that has “a potential for tax avoidance or evasion.” 26 U.S.C. § 6707A(c)(1). The term “potential” is capacious. It means “[c]apable of coming into being; possible,” Potential, BLACK'S LAW DICTIONARY; “[a]n event that may or may not happen,” Possibility, BLACK'S LAW DICTIONARY. See also Potential, MERRIAM–WEBSTER'S DELUXE DICTIONARY (10th ed. 1998) (“something that can develop or become actual”). That is a low bar. A potential occurrence need not probable, more likely than not, or even plausible. See Bell Atl. Corp. v. Twombly, 550 U.S. 544, 557 (2007) (distinguishing “between [the] possibility and plausibility” of entitlement to relief); see id. at 556 (“Asking for plausible grounds to infer an agreement does not impose a probability requirement at the pleading stage.”). If a transaction “may” allow an individual or entity to avoid paying taxes when they should have, then that transaction has “a potential for tax avoidance or evasion.” Because the IRS can require taxpayers to report a transaction if it is “of a type” that has “a potential for tax avoidance or evasion,” 26 U.S.C. § 6707A(c)(1) (emphasis added), the IRS may require taxpayers to report transactions within a category even if an individual transaction in that category does not have a potential for tax avoidance or evasion, see Type, MERRIAM–WEBSTER'S DELUXE DICTIONARY (“one having qualities of a higher category”).²

A “listed transaction” is “a tax avoidance transaction for purposes of section 6011.” 26 U.S.C. § 6707A(c)(2). The Code does not define a “tax avoidance transaction.” But the term and the Section's structure compel the conclusion that a “tax avoidance transaction” has more than “a potential for tax avoidance or evasion.” “Tax avoidance” is an adjective defining “transaction.” When Congress uses a term as an adjective, it does so to limit the scope of the word the adjective modifies. See Weyerhaeuser Co. v. U.S. Fish & Wildlife Serv., 586 U.S. 9, 19 (2018) (“Adjectives modify nouns—they pick out a subset of a category that possesses a certain quality.”); Rimini St., Inc. v. Oracle USA, Inc., 586 U.S. 334, 341 (2019) (same). The term “tax avoidance” supports the conclusion that a listed transaction must avoid taxes and cannot merely have the potential to do so to be subject to the reporting requirement.

The Section's structure confirms that a “listed transaction” is more than a potentially tax-avoidance transaction. Section 6707A(c)(2) does not include the word “potential,” even though its immediate neighbor does. See Hamdan v. Rumsfeld, 548 U.S. 557, 578 (2006) (“[A] negative inference may be drawn from the exclusion of language from one statutory provision that is included in other provisions of the same statute.”); Stanley v. City of Sanford, 606 U.S. 46, 53 (2025) (“That Congress used different language in these two provisions strongly suggests that it meant for them to work differently.”). The IRS has the discretion to rescind penalties for a failure to disclose “a reportable transaction” but not for a failure to disclose “a listed transaction.” 26 U.S.C. 6707A(d)(1)(A). And failures to report listed transactions may be penalized up to “$200,000 ($100,000 in the case of a natural person),” while failures to disclose “any other reportable transaction” may be penalized up to “$50,000 ($10,000 in the case of a natural person).” Id. § 6707A(b)(2)(A)–(B). Congress intended listed transactions to have more than a mere potential for tax avoidance. See Burwell, 576 U.S. at 496 (relying on the statute's structure); Eli Lilly & Co. v. Revive Rx, LLC, 812 F. Supp. 3d 708, 749–50 (S.D. Tex. 2025) (recognizing that a statute's heightened regulatory requirements for one category of entities reflected “careful policy choices”).

The remaining interpretive issue is “the requisite risk of tax abuse” for a listed transaction. CIC Servs., 593 U.S. at 213. Must a listed transaction constitute tax evasion, or must a listed transaction be only presumptively (that is, more often than not) tax evasive? The Section's text, structure, and purpose, as well as case law, support the second approach. Section 6707A(c)(2) uses the full term, “a tax avoidance transaction for purposes of section 6011.” 26 U.S.C. § 6707A(c)(2) (emphasis added). Section 6011 provides the IRS with the power to “prescribe[ ]” regulations that govern the “return[s] or statement[s]” taxpayers must file. 26 U.S.C. § 6011(a). Congress defined “a tax avoidance transaction” with respect to a reporting requirement, which, by its nature, must be overinclusive so the government can find the tax-evader needles in the taxpayer haystack. See Mann Constr., Inc. v. United States, 27 F.4th 1138, 1141 (6th Cir. 2022) (Sutton, C.J.) (“This information-gathering imperative allows the government to ensure compliance with tax provisions and ferret out improper tax avoidance.”).

Confirmation of this interpretation comes from the Code's structure. The Code does not specify the relationship between the IRS's determinations under § 6707A and the lawfulness of specific transactions designated as “reportable” or “listed.” See, e.g., Turnham, 979 F.3d at 1326; Green Rock, 104 F.4th at 223; see also 26 C.F.R. § 1.6011-4(a) (“The fact that a transaction is a [listed] transaction shall not affect the legal determination of whether the taxpayer's treatment of the transaction is proper.”). The statutory distinction between actual tax-evasive transactions and listed transactions supports concluding that the former is narrower than the latter. For these reasons, courts have considered listed transactions to be those that are “presumptively tax-avoidant.” Green Rock, 104 F.4th at 225.

After “[p]lugging in our definitions,” Morrissey v. United States, 871 F.3d 1260, 1265 (11th Cir. 2017), two tests emerge. First, the IRS did not exceed its statutory authority in designating micro-captive transactions as transactions of interest, 26 C.F.R. § 1.6011-11, if they are “of a type” that “may” not be “really for insurance.” Second, the IRS did not exceed its statutory authority in designating micro-captive transactions as listed transactions, id. § 1.6011-10, if they are “presumptively” (that is, more than half the time) “not really for insurance.”

The plaintiffs reject these tests. They argue that Congress precluded the IRS from considering whether micro-captive transactions are “really for insurance.” According to the plaintiffs, agencies may not promulgate a rule that “frustrates [ ] congressional policy,” (Document No. 59 at 13 (citation omitted)); Congress made a policy judgment that micro-captive transactions are legitimate, (id. at 13–15); and the Final Rule, by labeling some as potentially or presumptively illegitimate and imposing a reporting requirement, “second-guess[es]” Congress's judgment, (id. at 16). The plaintiffs rely on the diversification requirements in § 831(b)(2)(B)(I)–(II) to suggest that Congress affirmatively approved any micro-captive transaction that meets them. (See id. at 14–16; Docket Entry No. 65 at 2–8). The plaintiffs' position is divorced from the Code's text and history, and is unpersuasive.

The plaintiffs point to no Code language that authorizes deductions for benefits for micro-captive transactions or even captive transactions. Congress instead created two separate, and generally applicable benefits: the business deduction for “ordinary and necessary expenses” that every corporate entity may claim for its insurance, 26 U.S.C. § 162(a), and a benefit election for “every [small] insurance company other than life,” id. § 831(b)(2)(A). Each provision has its own requirements. “To be deductible as business expenses under [§] 162(a) of the Internal Revenue Code, the insurance premium payments made to the Captives must have been ‘really for insurance.’ ” Swift, 144 F.4th at 764 (quoting CIC Servs., 593 U.S. at 213). To elect benefits under § 831(b) of the Internal Revenue Code, the captive must be “a bona fide insurance company.” Avrahami, 149 T.C. at 197. The plaintiffs do not dispute that the terms “ordinary and necessary expenses” and “insurance company” impose statutory requirements that transactions and micro-captive insurers must satisfy. Nor do they contest the many tax and appellate cases that have concluded that captive and micro-captive transactions were not deductible. See, e.g., Swift, 144 F.4th at 769; Rsrv. Mech. Corp., 34 F.4th at 916; Royalty Mgmt. Ins., 2024 WL 4200129, at *1 & n.3 (collecting tax-court cases); see also Clougherty Packing Co., 811 F.2d at 1307 (holding that a parent-subsidiary captive transaction is not insurance); Stearns-Roger Corp., 774 F.2d at 416–17 (same); Beech Aircraft Corp., 797 F.2d at 923 (same); Humana Inc., 881 F.2d at 251–52 (same); Gulf Oil Corp., 914 F.2d at 411–12 (same). The plaintiffs offer no reason to depart from the well-settled rule that “the substance rather than the legal form of the transaction ․ governs.” United States v. Smith, 418 F.2d 589, 596 (5th Cir. 1969); see Comm'r v. Ct. Holding Co., 324 U.S. 331, 334 (1945) (“To permit the true nature of a transaction to be disguised by mere formalisms, which exist solely to alter tax liabilities, would seriously impair the effective administration of the tax policies of Congress.”).

The plaintiffs' reliance on the diversification requirements in § 831(b)(2)(B)(I)–(II) are also misplaced. The plaintiffs argue that because Congress barred tax benefits for micro-captive transactions that do not meet the diversification criteria, it implicitly approved of the micro-captive transactions that do meet the diversification criteria. (Docket Entry No. 65 at 6–8). But statutes “may,” and often do, “take one step at a time, addressing itself to the phase of the problem which seems most acute to the legislative mind.” Williamson v. Lee Optical of Ok. Inc., 348 U.S. 483, 489 (1955); see Williams-Yulee v. Fla. Bar, 575 U.S. 433, 449 (2015) (“A State need not address all aspects of a problem in one fell swoop; policymakers may focus on their most pressing concerns.”); TikTok Inc. v. Garland, 604 U.S. 56, 76 (2025) (same); see, e.g., SEC v. Drexel & Co., 348 U.S. 341, 349 (1955) (emphasizing that Congress “put an end to the worst of these practices”); Silver v. Silver, 280 U.S. 117, 124 (1929) (“It is enough that the present statute strikes at the evil where it is felt and reaches the class of cases where it most frequently occurs.”). As a result, courts do not “indulge in the loose assumption that when Congress adopts a new” statute, “it thereby deals with all situations falling within the general mischief which gave rise to the legislation.” Kirschbaum v. Walling, 316 U.S. 517, 521 (1942). The plaintiffs rely on this “loose assumption” rather than the more reasonable explanation that Congress adopted § 831(b)(2)(B)(I)–(II) to stamp out only the worst abuses and the most likely tax-evasive transactions.³

The legislative history confirms this application of traditional interpretive principles. See Delaware v. Pennsylvania, 598 U.S. 115, 138–40 (2023) (relying on legislative history to confirm the Court's interpretation of the statute's text); Munoz v. Intercontinental Terminals Co., L.L.C., 85 F.4th 343, 350 (5th Cir. 2023) (same). When Congress enacted § 831(b), its “intent” was not “that the election [ ] be used as a means of eliminating tax liability ․ but rather as a simplification for small companies.” (Docket Entry No. 37 at AR3183). Congress intended that only legitimate insurance companies could claim the benefits, confirming that “[a] company that does not meet the definition of an insurance company is not eligible” under § 813(b). STAFF OF J. COMM. ON TAXATION, 108TH CONG., GENERAL EXPLANATION OF TAX LEGISLATION ENACTED IN THE 108TH CONGRESS 99 (Comm. Print 2005). Congress “intended that IRS enforcement activities [would] address the misuse” of pre-existing tax exemptions and “that regulations or other Treasury guidance [would] provide for anti-abuse rules so as to prevent improper use of the provision, including, for example, by attempts to characterize as premiums any income that is other than premium income.” Id. Congress understood that the IRS would promulgate regulations like the Final Rule and did not chart a different course in enacting the diversification requirements. Congress added the diversification requirements to prevent the clearest case of tax evasion: “abuse of captive insurance companies for estate planning purposes.” S. REP. NO. 114-16, at 3 (2015); see Luke, Captivating Deductions, at 862 n.38 (“Legislation enacted in 2015 attempts to limit the extent to which captives used in such [estate-planning] strategies can obtain income tax benefits available to small captives.”). Congress's treatment of the worst abuses did not foreclose Treasury's authority to address the mine run of them.

Under the Internal Revenue Code, the IRS may designate micro-captive transactions as reportable transactions if they are of a type that may not be insurance, and the IRS may designate micro-captive transactions as listed transactions if they are presumptively not for insurance. Because this court has “police[d] the outer statutory boundaries of [Congress's] delegations,” it must now “ensure” the IRS “exercise[d]” its “discretion consistent with the APA.” Loper Bright, 603 U.S at 404.

2. Agency Findings

The next question is whether the agency reasonably found that the designated micro-captive transactions have the “the requisite risk of tax abuse,” CIC Servs., 593 U.S. at 213, to be designated as “transactions of interest” or “listed transactions,” 26 C.F.R. §§ 1.6011-10, 1.6011-11. “The issue is whether the agency reached a rational conclusion based on the facts before the agency and the agency's stated reasons for its decision.” Araujo Perez v. Blinken, 569 F. Supp. 3d 539, 550 (S.D. Tex. 2021) (citing State Farm, 463 U.S. at 43). This court cannot “substitute its judgment for that of the agency.” State Farm, 463 U.S. at 43. If the agency considered the “relevant data,” “articulate[d] a satisfactory explanation,” and did not rely on facts “which Congress has not intended it to consider,” this court must sustain the agency's action. Id. This court can overturn the agency's findings only if there was “a clear error of judgment,” id. (quoting Bowman Transp., Inc. v. Arkansas-Best Freight Sys., Inc., 419 U.S. 281, 285 (1974)), such that the agency's findings are “so implausible that [they] could not be ascribed to a difference in view or the product of agency expertise,” id.⁴

Applying this standard of review to the Final Rule, the court concludes that the agency (1) appropriately designated micro-captive transactions as transactions of interest under 26 C.F.R. § 1.6011-11 because they are of a type that has the potential for tax avoidance or evasion; but (2) exceeded its statutory authority in designating micro-captive transactions as listed transactions because there was no finding, supported by the administrative record, that the majority of the transactions covered by 26 C.F.R. § 1.6011-10 are to avoid or evade taxes.

i. Transactions of Interest (26 C.F.R. § 1.6011-11)

The IRS did not exceed its statutory authority in promulgating 26 C.F.R. § 1.6011-11. The agency designated transactions in which a captive: elects § 831 tax treatment; satisfies the Relationship Test; and satisfies either the Financing Factor or Loss-Ratio Factor (with an adjusted loss ratio of 60 percent). See 26 C.F.R. § 1.6011-11(b)–(c). The administrative record is clear that these transactions are of a type that may not really be for insurance, see CIC Servs., 593 U.S. at 213; the “potential for tax avoidance or evasion” is patent, 26 U.S.C. § 6707A(c)(1).

The IRS affirmatively found that micro-captive transactions designated in § 1.6011-11 are of a type that has the “potential for tax avoidance or evasion.” Final Rule at 3,538. The regulation “identif[ies] fact patterns” similar to those in which the IRS and courts have found that “the transaction at issue lacked the necessary characteristics, based on the specific facts in each case, to qualify as insurance for Federal tax purposes.” Id. at 3,539. The Final Rule relies on the case law's multi-factor definition of insurance. The Final Rule explains that an insurance contract must involve “risk shifting, risk distribution, [and] insurable risk” and look like “insurance in the commonly accepted sense.” Id. at 3,534–35. The Final Rule relies on the tax-court rulings and their reasoning to justify its finding that micro-captive transactions have a potential for tax avoidance or evasion. Id. at 3,538–39. As the Final Rule explains, “[t]he transactions described in § 1.6011-11 have many of the characteristics of self-insurance, and as such, taxpayers who deduct amounts paid to captives in such transactions may be engaged, as a matter of substance, in self-insurance.” Id. at 3,539. Because “more information is needed to determine if” transactions are not insurance or a disguised form of self-insurance, the IRS determined that a reporting requirement is appropriate. Id.

Two features of micro-captive transactions and insurance case law support the agency's findings. First, precedents show that payments for insurance premiums are deductible under § 162(a) but “amounts set aside in a loss reserve as a form of self-insurance are not.” Avrahami, 149 T.C. at 174. A captive can provide bona-fide insurance. For example, a captive transaction may increase the size of a captive's asset pool and reduce “the ratio between expected and actual losses”; put the captive's “reserves at risk”; and assign “claims administration to persons with a comparative advantage at that task.” Sears, Roebuck & Co., 972 F.2d at 864. But in other cases, a captive's reserve can be “treated ․ as if it were a tax-free savings account,” no different from self-insurance or a slush fund. Keating, 2024 WL 50234, at *24, *26. Because a captive transaction requires additional factual investigation to determine whether it is really for insurance, it has the requisite potential for tax avoidance. See Final Rule at 3,539.

Second, if the captive and the insured are parties to a financing arrangement, the insured may obtain the tax benefit of the premium payments without losing their economic benefit. Money will “move[ ] from one entity owned by” the taxpayers “to another,” allowing them to maintain “the same wealth despite the transfer.” Rsrv. Mech. Corp., 34 F.4th at 885. It is well settled that “a circular flow of funds among related entities does not indicate a substantive economic transaction for tax purposes.” Patel, 2025 WL 3158814, at *13 (quoting Merryman v. Comm'r, 873 F.2d 879, 882 (5th Cir. 1989)). If money is consistently flowing from the insured to the insurer and back to the insured, then the insurer may not have enough money in its reserves to satisfy legitimate claims. See, e.g., Malone & Hyde, Inc. v. Comm'r, 62 F.3d 835, 840 (6th Cir. 1995) (“[T]he establishment of a tax deduction is not, in and of itself, an ‘otherwise bona fide transaction’ if the deduction is accomplished through the use of an undercapitalized foreign insurance captive that is propped-up by guarantees of the parent corporation.”); see Patel, 2025 WL 3158814, at *13 (“The parallels between Malone & Hyde and the [micro-captive] cases before us are easy to draw.”); see also Final Rule at 3,551–52 (citing Malone & Hyde, 62 F.3d at 842–43). Even if the insurer is adequately capitalized—for example, when it is “licensed and regulated in a jurisdiction with extremely low reserve requirements,” Avrahami, 149 T.C. at 179—the circular flow of funds still raises suspicion that the insurer may not be acting as a traditional insurance company, such as by contracting through arm's-length negotiations or based on sound actuarial analyses. See, e.g., id. at 190–97; Rsrv. Mech. Corp., 34 F.4th at 844–85; Patel, 2025 WL 3158814, at *14–15; Caylor Land, 2021 WL 915613, at *13–16; see also Final Rule at 3,543. The potential for tax avoidance exists because when the insured pays its premiums to the micro-captive and the micro-captive returns to them to the insured, the insured may receive only the tax-free value of the premium payments, not insurance protection.

A similar concern arises even if the Financing Factor is not satisfied. If the Loss-Ratio Factor is satisfied, then the captive insurer may be excessively pricing premiums despite “artificially low or nonexistent claims activity.” Final Rule at 3,540. Because the Loss-Ratio Factor “measures whether the amount of liabilities incurred for insured losses and claims administration expenses is significantly less than the amount of premiums earned,” it is triggered when policies are not priced at a competitive level—that is, “price[d] in such a way that the premiums brought in cover losses and the insurer's business expenses with enough profit left over to keep investors happy.” Id. (quoting Avrahami, 149 T.C. at 152); see Avrahami, 149 T.C. at 194–95 (finding that the captive premiums far exceeded those of competitive commercial insurance carriers); see also Caylor Land, 2021 WL 915613, at *15 (finding that the captive premiums far exceeded the taxpayers' history of loss for the covered risk). The Loss-Ratio Factor suggests that the micro-captive transaction is creating “a loss reserve as a form of self-insurance,” Final Rule at 3,540, or insuring a risk for which coverage is not legitimately needed, see Helvering, 312 U.S. at 539 (“[T]he amounts must be received as the result of a transaction which involved an actual ‘insurance risk’ at the time the transaction was executed.”); cf. Harper Grp. & Includible Subsidiaries v. Comm'r, 96 T.C. 45, 58 (1991) (“[I]ts exposure, was real, not illusory.”), aff'd, 979 F.2d 1341 (9th Cir. 1992).

The plaintiffs do not contest the IRS's finding that some of the micro-captive transactions in § 1.6011-11 have the potential for tax avoidance or evasion. (See Docket Entry No. 61 at 14:21–15:1). The plaintiffs argue instead that the IRS should have used a narrower or different set of criteria focused on transactions that are actually or more likely to be abusive. (Docket Entry No. 65 at 9–10). The plaintiffs argue that the “mere fact of related-party financing,” without more, does not establish that a micro-captive transaction is tax evasive, (id. at 19–20), and that the Loss-Ratio Factor suggests only that the captive is insuring “low-frequency, high-severity risks,” (id. at 3–4). But requiring the disclosure of such transactions is far from establishing that they are in fact tax-evasive. The plaintiffs' over-inclusiveness arguments are by their nature a separate arbitrary-and-capricious challenge. The court addresses them below. The plaintiffs' argument that the Relationship Test, Financing Factor, and Loss-Ratio Factor are not narrowly tailored to transactions that are actually tax-evasive does not undermine the agency's finding that some transactions that satisfy § 1.6011-11's disjunctive test are tax-evasive.

This finding is sufficient to conclude that § 1.6011-11 is within the agency's statutory authority. The IRS may require taxpayers to report a transaction if it is “of a type” that has “a potential for tax avoidance or evasion.” 26 U.S.C. § 6707A(c)(1) (emphasis added). Making the “type” of transaction the object of the agency's determination allows it to designate transactions at a higher level of generality. If the IRS determines that a “type” or “category” of transactions has a potential for tax avoidance or evasion, see Paroline v. United States, 572 U.S. 434, 447 (2014), then it can require taxpayers to report transactions within that category. As long as the designated transactions have the characteristics of the type of transaction that has the potential for tax avoidance or evasion, it is immaterial that some of the individual transactions within that type do not have the potential for tax avoidance or evasion.

Section 6707A's structure confirms this definition of a “reportable transaction.” The IRS's ability to designate reportable transactions is broader than its ability to designate listed transactions. A listed transaction is “a transaction specifically identified by the Secretary as a tax avoidance transaction.” 26 U.S.C. § 6707A(c)(2) (emphasis added). Section 6707A(c)(2)'s focus on specific transactions, rather than types of transactions (and transactions that are of types of transactions), further supports the conclusion that the IRS did not exceed its statutory authority under § 6707A(c)(1) by using overinclusive reportable-transaction designations.

Congress's use of “of a type” allows the IRS to designate reportable transactions at a high level of generality. Reporting requirements must use “an administrable test that would adequately put all relevant taxpayers on notice of their obligations.” Final Rule at 3,541. Administrability sometimes comes at the cost of precision, and Congress did not intend to hamstring the agency by requiring precision in both subsections (c)(1) and (c)(2). Congress rationally favored administrability through breadth in subsection (c)(1) but precision through specificity in subsection (c)(2). The agency's finding that § 1.6011-11 micro-captive transactions are of a type that has the potential for tax avoidance or evasion, Final Rule at 3,538, is sufficient to satisfy the statutory requirements.

In sum, “the administrative record includes recent multiple tax court decisions in which courts determined that a taxpayer in a micro-captive transaction remitted amounts treated as insurance premiums for something other than insurance.” CIC Servs., 2026 WL 622836, at *5. These rulings “provide the ‘facts and data’ necessary for the Treasury Department and the IRS to conclude that captive insurance arrangements are potentially used for tax avoidance” and “a legitimate basis for the Treasury Department and the IRS to seek additional information from captives and material advisors to aid in determining whether a captive is, in fact, impermissibly avoiding tax liability.” Id.

ii. Listed Transactions (26 C.F.R. § 1.6011-10)

The IRS exceeded its statutory authority in promulgating 26 C.F.R. § 1.6011-10. The IRS designated transactions for which a captive that has at least ten years of insurance transactions elects § 831(b) taxation and satisfies the Relationship Test, the Financing Factor, and the Loss-Ratio Factor (with an adjusted loss ratio of 30 percent). See id. § 1.6011-10(b)–(c). The agency exceeded its statutory authority because the record does not include or support a finding that the transactions designated by § 1.6011-10 are presumptively not really for insurance. See CIC Servs., 593 U.S. at 213.

There is no finding in the Final Rule that transactions under § 1.6011-10 are presumptively tax avoidant. The Final Rule justifies § 1.6011-10 through an explanation of what typically occurs in an abusive micro-financing arrangement: “amounts paid as premiums have not only avoided ordinary taxation but have continued to avoid tax while back in the hands of the related parties who caused the premiums to be paid and deducted.” Final Rule at 3,546. The Final Rule made § 1.6011-10 a conjunctive test, requiring the presence of both the Financing Factor and the Loss-Ratio Factor, so that it covers transactions that are “consistently associated with transactions that are ․ tax avoidance transactions.” Id. But this analysis addresses only half the problem. Identifying typical features of an abusive transaction does not identify transactions that are typically abusive, so the finding that § 1.6011-10 covers the typical abusive micro-captive transactions is not a finding that § 1.6011-10 covers presumptively abusive transactions. If the Final Rule purported to make the latter finding, then there was a “clear error of judgment,” State Farm, 463 U.S. at 43 (quoting Bowman Transp. Inc., 419 U.S. at 285), because neither Final Rule nor the administrative record includes information about how many captives qualify under § 1.6011-10 or how many of those qualifying captives have been evading taxes.

The dearth of data is material because the Final Rule recognizes that § 1.6011-10 will cover some legitimate transactions as well as those that are abusive. “[T]he Treasury Department and the IRS agree that the presence of related-party financing in a micro-captive transaction by itself may not rise to the level of tax avoidance, as it may be that such financing was determined at arm's length or otherwise treated as a bona fide financing arrangement between the related parties.” Final Rule at 3,546 (citing Avrahami, 149 T.C. at 199–204). A loan “to a related party may serve as an investment that is not correlated to stock or bond market performance, providing a fixed income stream that can be matched with potential claims obligations.” (Docket Entry No. 37-1 at AR4276). The IRS agrees “that low loss ratios may be the result of coverage of low-frequency, high-severity risks.” Final Rule at 3,541. The administrative record reflects that lower loss ratios may be reasonable because micro-captives may have different structures, risk pools, and cost profiles than do commercial insurance carriers. (Docket Entry No. 37-1 at AR4649). Although the agency found that § 1.6011-10's conjunctive test “significantly narrows the scope” of the final regulation, Final Rule at 3,541, there is no finding that the Final Rule is sufficiently narrow to satisfy § 6707A(c)(2)'s requirement of presumptive tax avoidance, see Green Rock, 104 F.4th at 222; (see also Docket Entry No. 66 at 19 (“The Final Rule does not contend that every transaction with a loss ratio below a certain numerical threshold is tax avoidant.”)).

The agency defends § 1.6011-10 by arguing that the Final Rule's “objective factors are needed to reasonably capture such transactions in a fair and administrable way” and “that a modified loss ratio and related-party financing provide reasonable metrics to do” so. (Docket Entry No. 66 at 14). In its view, “[o]nce Treasury has cleared each reasonableness checkpoint along its path, all it has left to do is draw a non-arbitrary line.” (Id.). Because the “Final Rule aims for percentages ․ significantly-below-the-norm for listed transactions,” it is reasonable enough to satisfy APA review. (Id.). The defendants' arguments find some support in the Supreme Court's APA case law. Agencies may rely on “materially incomplete” data when they “acknowledge[ ] the gaps in the data” and “repeatedly ask[ ] for data on [an] issue.” Prometheus, 592 U.S. at 425. But their reliance on Prometheus is misplaced in this context, in which Congress required a specific finding in order to designate transactions as listed transactions. See Finberg v. U.S. Dep't of Agric., 6 F.4th 1332, 1336 (D.C. Cir. 2021) (“Agency action is arbitrary and capricious if the agency fails to ․ make statutorily-required findings.”). Congress required the agency to find that transactions designated under § 1.6011-10 are presumptively tax avoidant. See 26 U.S.C. § 6707A(c)(2). Rather than designate transactions based on the “the requisite risk of tax abuse,” CIC Servs., 593 U.S. at 213, the Final Rule focuses on transactions that bore the typical features of tax-avoidance, Final Rule at 3,546, and were abnormal, id. at 3,541. But without the statutorily-required finding of presumptive tax avoidance or an administrative record that supports this finding, see State Farm, 463 U.S. at 43–44, § 1.6011-10 exceeds the agency's authority.

For these reasons, 26 C.F.R. § 1.6011-10 exceeds the defendants' statutory authority.

B. The Arbitrary-and-Capricious Challenge

The plaintiffs argue that even if the defendants acted within statutory authority, the failure to narrowly tailor § 1.6011-11 was arbitrary and capricious. See Mayfield, 117 F.4th at 615 (“Holding that a rule survives a particular type of APA review does not determine how it fares under other types of review, even if both types of review raise similar issues.”). The plaintiffs argue that the Financing Factor and the Loss-Ratio Factor do not alone show a potential for tax-abuse, so the agency acted arbitrarily and capriciously by failing to make the test conjunctive or use other factors that exclude those transactions guaranteed to be legitimate. (See Docket Entry No. 65 at 16–22). For example, the plaintiffs argue that because some related-party financing arrangements are legitimate, the IRS should have used different tests more tailored to the facts that the Financing Factor is supposed to be a proxy for. (Id. at 19–20). Similarly, for the Loss-Ratio Factor, the plaintiffs argue that loss ratios of 30% or 60% are not indicative of even the potential for tax abuse in the micro-captive context. (See id. at 11–13). The defendants' response leads with a broader point that the IRS balanced the risks of overinclusive categories and proxies against the burdens of a reporting requirement. (Docket Entry No. 66 at 10–11).

The defendants' broader argument is persuasive. The defendants chose the Financing Factor and the Loss-Ratio Factor “to ensure administrability and clarity for taxpayers whose transactions are identified in the regulations, so taxpayers can clearly determine whether they are participants or material advisors, and thus be on clear notice of their obligations.” Final Rule at 3,535. Those factors identify trends that raise concerns about tax avoidance, even if they are overinclusive. See id. at 3,542 (reducing the Loss-Ratio Factor for transactions of interest from 65% to 60% in response to comments that 65% was “still too high” and explaining that the 60% threshold was “the lowest loss ratio supported by available data”); id. at 3,546 (rejecting objections to the Financing Factor “[b]ecause the potential for tax avoidance still exists when there is related-party financing”). Erring on the side of some over-inclusiveness for the reporting obligation is reasonable because “the Treasury Department and the IRS expect the reporting burden to be low,” at around 8.11 hours per form. Id. at 3,558–59. “[T]he information sought is necessary for regular annual return preparation and ordinary recordkeeping.” Id. at 3,558. The defendants did not act arbitrarily by favoring a broader reporting requirement to combat tax evasion, even if some taxpayers will suffer the minimal burden of reporting some transactions that are likely to be legitimate insurance arrangements. See U.S. Sportsmen's All. Found. v. CDC, 167 F.4th 813, 822 (6th Cir. 2026) (“That an agency regulation might be over-or under-inclusive does not necessarily render it arbitrary and capricious.”); see, e.g., TMJ Implants, Inc. v. U.S. Dep't of Health & Hum. Servs., 584 F.3d 1290, 1294–95 (10th Cir. 2009) (explaining that the FDA rejected comments that “broad reporting requirements may impose unduly onerous burdens” on regulated entities based on its need for “an expansive reporting system”); United States v. Kanchanalak, 192 F.3d 1037, 1046 (D.C. Cir. 1999) (“[W]e know of no bar to an agency's interpretation of a prophylactic disclosure rule, such as this one, that may overshoot the mark a bit, so long as it stays in reasonable range.”).

The defendants' specific rebuttals to the plaintiffs' objections to the Relationship Test, Financing Factor, and Loss-Ratio Factor are successful. The justifications for using each of them are “reasonable and reasonably explained for purposes of the APA's deferential arbitrary-and-capricious standard.” Prometheus, 592 U.S. at 426.

1. The Relationship Test

The plaintiffs object to the Relationship Test because “Congress itself has decided that closely held micro-captives are eligible for Section 831(b)'s tax benefits if they satisfy the ‘relatedness’ test.” (Docket Entry No. 59 at 30). They also argue that “a captive's ownership structure is irrelevant” to “whether a transaction is insurance for federal tax purposes.” (Docket Entry No. 65 at 21 (citing Humana Inc., 881 F.2d at 252)). Neither argument is persuasive.

For the reasons already discussed, Congress did not decide that micro-captives that satisfy § 831(b)'s diversification criteria are automatically eligible for tax benefits, even if they are not providing insurance. See Final Rule at 3,555; Ryan, 2025 WL 3089415, at *11 n.9; see also Kirschbaum, 316 U.S. at 521. The Final Rule's Relationship Test excludes “small insurers” that Congress intended to elect § 831(b) and that do not pose a high risk of tax evasion, such as “groups of farmers and fire associations.” Final Rule at 3,555. The Relationship Test also helps identify tax-avoidance transactions because the risk that the captive is acting as a mere self-insurance reserve or otherwise not acting as a legitimate insurance provider is heightened when the insurer and the insured are under common ownership. See id. at 3,550–51 (“Because these transactions involve closely held related entities, there is little to no barrier to the manufacture of claims in these arrangements.”).

The plaintiffs' reliance on Humana Inc. is misplaced. They are correct that Humana Inc. held, consistent with general principles of corporate and tax law, that courts must treat related subsidiaries as “separate corporate entities.” 881 F.2d at 252. But Humana Inc. reasoned from that premise only that the tax court did not clearly err in finding “there was both risk shifting and risk distribution between the subsidiaries and the captive insurance company.” Id. The Sixth Circuit recognized that courts do not need to respect separate corporate entities when a transaction between them is a “sham” or “lack[s] [a] business purpose under the relevant tax statute.” Id. at 255; accord CIC Servs., 593 U.S. at 213. The factors that determine whether a transaction looks like commonly accepted notions of insurance address that business-purpose concern. See R.V.I., 145 T.C. at 231; Avrahami, 149 T.C. at 191; see also Swift, 144 F.4th at 765, 768–69. The Final Rule respects that separate subsidiaries are legally distinct while recognizing that the risk of a sham transaction is higher where companies are related and controlled by common ownership. Final Rule at 3,550–51; see, e.g., Caylor Land, 2021 WL 915613, at *15–16 (finding the captives did not set premiums based on actuarial practices); Keating, 2024 WL 50234, at *26 (finding the captives did not negotiate at arm's-length); Patel, 2024 WL 1270772, at *28–29 (same).

The use of the Relationship Test is “reasonable and reasonably explained.” Prometheus, 592 U.S. at 423.

2. The Financing Factor

The plaintiffs argue that the Financing Factor does not target transactions that are potentially tax avoidant because “related-party financing is an expected and intended characteristic of micro-captive insurance transactions.” (Docket Entry No. 59 at 28). They highlight the Final Rule's concession that related-party financing does not itself “rise to the level of tax avoidance” and can instead be a “bona fide financing arrangement between the related parties.” (Id. at 28–29 (quoting Final Rule at 3,546)). They further argue that the Financing Factor is not an appropriate proxy for concerns that a transaction is not really for insurance. (Id. at 29–30). A related-party financing arrangement does not address whether a captive is undercapitalized or simply returning premiums to the insurer, for example, because the financing arrangement may involve a captive with $5 million in assets that provides a $250,000 secured loan to its affiliates on arm's-length terms after receiving regulatory approval. (Id. (citing Docket Entry No. 37-1 at AR4727)). The plaintiffs argue that that the Financing Factor should have focused on “how a captive invests its assets, including whether the captive has loaned out the majority of its assets” or is transferring them all back to the insurer. (Docket Entry No. 65 at 19–21). The argument is not persuasive.

The Financing Factor identifies transactions in which the insured entity receives part of its premium payments back through a guarantee, loan, or other transfer of the captive's capital. 26 C.F.R. § 1.6011-11(c)(1)(i). Through these arrangements, the “amounts paid as premiums have not only avoided ordinary taxation but have continued to avoid tax while back in the hands of the related parties who caused the premiums to be paid and deducted.” Final Rule at 3,546. The Financing Factor identifies the heart of the agency's tax-avoidance concern.

The agency considered and reasonably rejected the plaintiffs' argument that the Financing Factor is overinclusive. See State Farm, 463 U.S. at 43. The Final Rule acknowledges that related-party financing may be “determined at arm's length or otherwise treated as a bona fide financing arrangement between the related parties.” Final Rule at 3,546. But the Financing Factor captures many transactions in which the “key abuses” occur. Id. “[T]he IRS has seen multiple transactions for which approval was required but not sought, or for which approval may have been granted but, nevertheless, the parties' treatment of the financing arrangement did not comport with industry standards.” Id.⁵ “Because the potential for tax avoidance still exists when there is related-party financing, the final regulations include the Financing Factor in the identification of a Micro-captive Transaction of Interest.” Id.

The agency also considered and reasonably rejected the plaintiffs' arguments that it should have used reporting criteria that more directly targets the insurance factors recognized in the case law, such as undercapitalization or the return of premiums to the insured. In response to comments that favored measurements of a captive's liquidity, the Final Rule explained that those commenters “did not specify what potential measure of liquidity (such as the character of assets, amount of assets, or comparison of assets to Captive's liabilities) would better identify micro-captive transactions that are or may be tax avoidance transactions.” Id. The Final Rule also considered that “regardless of the specific measure of liquidity used, determinations thereof would be too fact-specific (and dependent upon individual policy terms and jurisdictional requirements) to be administrable” for purposes of a tax-reporting rule. Id. Focusing on “a complete loop” of insurance premiums, Avrahami, 149 T.C. at 186, poses similar administrability concerns. Insureds could still illegitimately collect “some portion of those tax-deferred amounts directly or indirectly,” Final Rule at 3,546, and could “avoid[ ]” the “characterization of their micro-captive arrangements as ‘transactions of interest,’ ” cf. id. at 3,545 (addressing reporting games with the Loss-Ratio Factor), simply by returning fewer than all the premiums paid to the insured. The Final Rule uses a relevant, clear, and administrable reporting requirement. Although it is overinclusive, it is not unreasonably so.

The use of the Financing Factor was “reasonable and reasonably explained.” Prometheus, 592 U.S. at 423.

3. The Loss-Ratio Factor

The plaintiffs argue that the Loss-Ratio Factor for transactions of interest creates too many false positives because a low loss ratio does not suggest the absence of legitimate insurance risk or that an insurance policy itself is illegitimate. (Docket Entry No. 59 at 20). The plaintiffs argue that low loss ratios, including loss ratios far lower than the 60% threshold for transactions of interest, are common for micro-captives because they “provide coverage for low-frequency, high-severity risks (such as global pandemics, large oil spills, or terrorist attacks) that require them to maintain large capital reserves in the event the catastrophic risk comes to pass and claims materialize.” (Id. at 21). The plaintiffs add that lower loss ratios are typical for micro-captives, which have “very different cost structures, risk pools, and business models than Section 831(b) captives (which do not compete for broad third-party business and lack the marketing and distribution costs that drive commercial carriers' loss ratios).” (Docket Entry No. 65 at 14). Based on these points, the plaintiffs argue that the defendants erroneously relied on flawed data from commercial insurance carriers. (See Docket Entry No. 59 at 22–28; Docket Entry No. 65 at 14–16). The plaintiffs' arguments do not show arbitrary and capricious reliance on the Loss-Ratio Factor.

The Loss-Ratio Factor helps identify transactions that may not be legitimate insurance. The Loss-Ratio Factor is based on the premiums that come in and the claims that are paid out. If premiums are excessively priced relative to “artificially low or nonexistent claims activity,” then the transaction may be tax avoidant. Final Rule at 3,540. “Stated another way, the loss ratio factor attempts to identify situations in which a captive arrangement may not be functioning like insurance in the commonly accepted sense.” CIC Servs., 2026 WL 622836, at *7. The administrative record lays out this logic and shows that the agency considered and reasonably rejected the commenters' objections to the Loss-Ratio Factor. See Final Rule at 3,540–45.

The agency considered and reasonably rejected the argument that loss ratios below 60% are common for micro-captives because they may cover low-frequency, high-severity risks. Id. at 3,541. The Final Rule recognized that “low loss ratios may be the result of coverage of low-frequency, high-severity risks” but reaffirmed that the Loss-Ratio Factor is not “intended to act as a proxy for the actuarial basis of premium pricing.” Id. Instead, the Loss-Ratio Factor requires reporting of transactions in which “premiums were consistently priced to meet the target threshold under section 831(b) without regard to reasonable estimates for loss experience.” Id. The Final Rule concedes, as the plaintiffs point out, that low loss-ratios may capture legitimate insurance arrangements. But over-inclusiveness is not fatal to an agency's final rule, U.S. Sportsmen's, 167 F.4th at 822, especially when, as here, it is targeted to capture a common feature of abusive micro-captive transactions, see Final Rule at 3,541 (collecting cases); (see Docket Entry No. 63-1 at 20 (“If an arrangement is tax avoidant, one would expect it to have a low loss ratio ․”)).⁶

In addition, the Final Rule addresses the argument that the loss ratio is too overinclusive. The Final Rule responds to the argument that years' worth of excessive premiums are justified for low-frequency, high-severity claims by calculating loss ratios over ten years, when possible. See 26 C.F.R. § 1.6011–11(b)(2)(ii). The Final Rule explains that the longer computation period allows “a Captive significant time to develop a reasonable loss history that supports the use of a micro-captive for legitimate insurance.” Final Rule at 3,541. The Final Rule cites case law that refutes the plaintiffs' premise that a micro-captive provides legitimate insurance when it is of a type that is expected to have low loss ratios. See id. at 3,543 (collecting cases). The plaintiffs rely, for example, on the “Treasury's own terrorism reinsurance program,” which “had a 0% loss ratio from 2003 to 2019.” (Docket Entry No. 65 at 11–12). But in Avrahami, the tax court found that the micro-captive was not providing legitimate terrorism insurance because, among other reasons, it charged “grossly excessive” premiums, 149 T.C. at 188, and negotiated contracts under which it was unlikely to pay out claims, see id. at 189–90. Using a micro-captive arrangement to insure low-frequency risks may present a high risk of abuse, because a taxpayer could be willing to spend the insurer's reserves on the unlikely chance that they will ever be needed. See, e.g., id. at 190 (recognizing the circular payment scheme); Caylor Land, 2021 WL 915613, at *15 (finding both premiums far in excess of the entities' damage from the covered risk and a circular payment scheme); see also Keating, 2024 WL 50234, at *31 (explaining that “loans to [the insured] to finance its operations” showed “confidence” that the insurer's “surplus would not be needed to pay substantial claims on the captive policies and that [the insurer] was overcharging for the coverage provided”).

The agency also considered and reasonably rejected the plaintiffs' argument that the Final Rule's 60% ratio is flawed. The agency found that a 60% loss ratio was appropriate based on national loss-ratio averages drawn from data from the National Association of Insurance Commissioners (NAIC). Final Rule at 3,542. National averages “ranged between 67.2 and 76.2 percent per year from 2012 to 2021” and “69.0 and 76.4 percent per year from 2014 to 2023.” Id. The Final Rule adjusted the loss ratio down to 60% in response to comments about potential inadequacies with the NAIC data set. See id. at 3,541–45.

Commenters argued, like the plaintiffs do, that the NIAC data is based on “auto and homeowners insurance that are rarely covered by captives and exhibit higher loss ratios due to market competition.” (Docket Entry No. 59 at 25). The Final Rule responded “[b]y removing the high frequency, low severity coverages that captives are unlikely to cover for each year from 2013 through 2022 from the annual data and computing the comparison of liabilities incurred for insured losses and claim administration expenses to premiums earned less policyholder dividends.” Final Rule at 3,542. This modified calculation resulted in an “average nine-year modified loss ratio” of about “66 percent.” Id. Then, in light of commenters' concerns that “the proposed 65 percent modified loss ratio [was] still too high,” the Final Rule adjusted downward the “the Loss Ratio Factor percentage for identification of a transaction of interest ․ to 60 percent.” Id. The Final Rule explained that its 60% line represented “the lowest loss ratio supported by available data.” Id.

The plaintiffs argue that the NAIC data is nonetheless “useless” and that better data was available. (Docket Entry No. 65 at 12–18). The plaintiffs overstate their argument that the NAIC data was “useless” for micro-captives. The plaintiffs cite to case law that addressed “commercial insurance companies' premiums to surplus ratio.” CFM Ins., Inc. v. Comm'r, T.C.M. (RIA) 2025-083, 2025 WL 2207492, at *30 (T.C. 2025). A premium-to-surplus ratio is not the same as a loss ratio, see Rent-A-Center, 142 T.C. at 12, and the Final Rule attempted to adjust the loss ratio it used to account for the differences between commercial and micro-captive insurance, see Final Rule at 3,542. Even if the NAIC data is flawed, agencies can rely on “materially incomplete” data so long as they “acknowledge[ ] the gaps in the data,” try to obtain better data, and cannot do so. Prometheus, 592 U.S. at 425. That is what the record shows. As the Final Rule acknowledges, “although commenters were critical of the aggregated data provided by the NAIC, commenters did not point to, and the IRS is not aware of, an alternative publicly-available data set that would be more appropriate.” Final Rule at 3,542. The agency used the data it had and made reasonable adjustments to the dataset to address commenters' concerns.

The plaintiffs argue that the defendants should have relied on the AM Best dataset “because it excludes certain large insurance companies.” (Docket Entry No. 65 at 16–17). But the agency considered and reasonably rejected the plaintiffs' argument. The agency explained that the AM Best dataset was “a table illustrating the distribution of AM Best Company's average loss and loss administration expenses ratios for small insurance companies, described as insurers grouped by capital and surplus up to $10 million.” Final Rule at 3,542. The agency found that the AM Best dataset included claims vastly different from those that micro-captives covered and included policies that micro-captives generally are not permitted to offer. Id. Although the plaintiffs argue that the NAIC dataset suffers the same flaws, (Docket Entry No. 65 at 16–17), the Final Rule explained that the AM Best dataset was only a “table,” not a “published data set,” Final Rule at 3,542. As a result, the agency could not adjust the AM Best data to better reflect the micro-captive industry. The agency concluded that the NAIC data was a better “starting point,” as it could be “modified.” Id. The Final Rule's reliance on the NAIC data was reasonable. Compare Prometheus, 592 U.S. at 425–26 (permitting use of “materially incomplete” data because the agency tried to use the best data available), with Chamber of Com. of U.S. v. SEC, 85 F.4th 760, 775 (5th Cir. 2023) (finding arbitrary and capricious the agency's failure to consider data that it received “in response to its solicitation”).

Lastly, the plaintiffs argue that the 60% threshold is facially unreasonable because it is around the industry norm and far from the 0–15% loss ratios in cases in which the tax court found tax fraud or the defendants previously asserted tax fraud. (Docket Entry No. 59 at 26; Docket Entry No. 65 at 14–15; see also Docket Entry No. 48-1 at AR16550); see Ryan, 2025 WL 3089415, at *13. The court disagrees. The 60% threshold is reasonable enough for APA review. The plaintiffs' argument proceeds from the faulty premise that if a micro-captive transaction barely qualifies as a transaction of interest, it is a proper insurance transaction. That is not necessarily the case.

The administrative record supports the conclusion that micro-captives that satisfy the Loss-Ratio Factor are potentially tax evasive. An unmodified loss ratio can suggest the profitability of an insurance company; its profit rate is 100% less the loss ratio and business-expense ratio. See Final Rule at 3,540 (citing Avrahami, 149 T.C. at 152); (Docket Entry No. 48-2 at AR17783 n.103, 17826 n.158). Underwriting profits “in the commercial insurance market have typically been in the 5% to 6% range.” (Docket Entry No. 48-2 at AR17254 (property and casualty coverage)). A typical commercial-insurance loss ratio of 50-70% will result in profit margins of about 10-20% after taking investment income into account. (Docket Entry No. 48-1 at AR16686; see also Docket Entry No. 59 at 26 n.6 (plaintiffs cite these statistics)). The Loss-Ratio Factor uses a modified loss ratio that accounts for “claim administration expenses” and for “policyholder dividends,” which correct overpriced premiums. Final Rule at 3,542, 3,545. The Final Rule's modifications make the Loss-Ratio Factor a stronger indicator of an insurer's profits. If an insurance provider has a lower-than-average modified loss ratio, that provider is likely earning unreasonable profits. See id. at 3540 (explaining that the loss ratio helps identify whether an insurer is earning reasonable profits). And because, as the plaintiffs boast, micro-captives typically “do not require the same kinds of overhead costs as commercial insurers,” (Docket Entry No. 59 at 3 (first citing Docket Entry No. 37-1 at AR4634–35; and then citing Docket Entry No. 38-2 at AR5282–83)), a below-average modified loss ratio suggests an even more unusual underwriting profit margin, (contra id. (suggesting that captives should “provide lower and more stable premium prices” because they have lower costs)). The agency reasonably sought information about insurance transactions resulting in unusual profit margins, which may suggest that taxable revenue is being diverted into unnecessary insurance premiums.⁷

The substantially lower loss ratios in the tax-court cases do not undermine the Final Rule's approach. The plaintiffs' focus on the tax-court cases suffers from a form of selection bias. See David H. Kaye & Hal S. Stern, Reference Guide on Statistics and Research Methods, in REFERENCE MANUAL ON SCIENTIFIC EVIDENCE 229, 480–83 (4th ed. 2026); see, e.g., In re Chevron U.S.A., Inc., 109 F.3d 1016, 1019–20 (5th Cir. 1997) (selecting the “best” and “worst” cases for bellwether trials would not result in a representative sample for settlement). As the defendants highlighted at oral argument, the government is not the plaintiff in tax-court litigation. Taxpayers choose to litigate disputes with the IRS, and the administrative record shows that they do so when the assessed tax deficiencies and penalties are the highest. See, e.g., Swift, 144 F.4th at 773 (affirming penalties); Caylor Land, 2021 WL 915613, at *17 (imposing penalties “across the board”); Keating, 2024 WL 50234, at *40 (imposing “accuracy-related penalties across the board”). The plaintiffs err by extrapolating from the tax-court cases that a loss ratio above 1% is unreasonable.⁸

The defendants disclaim reliance on the individual facts of the tax-court cases to establish the 60% threshold. (Docket Entry No. 63-1 at 32 n.12). They instead highlight the tax-court cases because they show a pattern of improper premium pricing “to meet the target threshold under section 831(b) without regard to reasonable estimates for loss experience.” Final Rule at 3,541. The Loss-Ratio Factor identifies cases in which that practice could occur. See id. at 3,540 (explaining that the Loss-Ratio Factor “generally identif[ies] ․ excessive pricing of premiums and artificially low or nonexistent claims activity” (emphasis added)). The defendants are reasonably skeptical of insurance transactions that generate unnecessary (even if modestly unnecessary) profits for captives, which may reflect non-deductable “distributions of [the insured's] profits,” rather than legitimate “expenses.” Caylor Land, 2021 WL 915613, at *9 (citing Elliotts, Inc. v. Comm'r, 716 F.2d 1241, 1243 (9th Cir. 1983)).

In the end, the agency is correct that the plaintiffs' arguments against the Loss-Ratio Factor reflect an irreconcilable policy disagreement. See State Farm, 463 U.S. at 43. The agency “reasonably concluded that when the other relevant criteria are satisfied, the arrangements with modified loss ratios below the industry average deserve a second look.” (Docket Entry No. 63-1 at 33). The Final Rule imposes a two-page reporting requirement. In using a capacious definition of “reportable transaction,” Congress made the policy judgment that the agency should have wide discretion to require taxpayers to report transactions that are of a type of transaction with a potential for tax abuse. See 26 U.S.C. § 6707A(c)(1) (“The term ‘reportable transaction’ means any transaction” that “is of a type which the Secretary determines as having a potential for tax avoidance or evasion.”). In light of the “minimal” and “low” reporting burdens, Final Rule at 3,558, the agency did not act arbitrary and capriciously in requiring taxpayers to report transactions that have slightly below average loss ratios, even if those transactions are likely legitimate.

The use of the Loss-Ratio Factor was “reasonable and reasonably explained.” Prometheus, 592 U.S. at 423.

* * *

26 C.F.R. § 1.6011-11 survives review under the Administrative Procedure Act.⁹

C. Remedy

The plaintiffs move for (1) vacatur of the Final Rule; (2) a declaratory judgment; (3) a permanent injunction barring the defendants from enforcing the Final Rule against the plaintiffs, their clients, and affiliates; (4) and a permanent injunction requiring the defendants to destroy or return all materials provided in response to the Final Rule. (Docket Entry No. 59 at 34–35). The defendants object to each of the proposed remedies. They argue primarily that this court should remand the case for further agency action that can correct any deficiencies, rather than vacate the Final Rule in its entirety. (Docket Entry No. 63-1 at 33–35; Docket Entry No. 66 at 18–20). They argue further that if the court remands for further agency action, the additional relief that the plaintiffs seek ultimately may be unnecessary. (See id.).

Under the APA, reviewing courts shall “hold unlawful and set aside agency action, findings, and conclusions.” 5 U.S.C. § 706(2). Because the APA directs courts to “set aside agency action,” id., the “default rule is that vacatur is the appropriate remedy,” Data Mktg. P'ship, LP v. DOL, 45 F.4th 846, 859 (5th Cir. 2022); accord Rest. L. Ctr. v. DOL, 120 F.4th 163, 177 (5th Cir. 2024). But vacatur is not required, and courts may remand a case for further agency action to correct the errors found. Texas v. United States, 50 F.4th 498, 529 (5th Cir. 2022); see Tex. Ass'n of Mfrs. v. U.S. Consumer Prod. Safety Comm'n, 989 F.3d 368, 389 (5th Cir. 2021) (“However, only in rare circumstances is remand for agency reconsideration not the appropriate solution.” (cleaned up)); Fertilizer Inst. v. EPA, 935 F.2d 1303, 1312 (D.C. Cir. 1991) (“[W]hen equity demands, an unlawfully promulgated regulation can be left in place while the agency provides the proper procedural remedy.”).

To determine whether vacatur is the appropriate remedy, courts consider: “(1) the seriousness of the deficiencies of the action, that is, how likely the agency will be able to justify its decision on remand; and (2) the disruptive consequences of the vacatur.” Texas, 50 F.4th at 529 (quoting United Steel v. Mine Safety & Health Admin., 925 F.3d 1279, 1287 (D.C. Cir. 2019)). “Remand without vacatur of the agency action is ‘generally appropriate when there is at least a serious possibility that the agency will be able to substantiate its decision given an opportunity to do so.’ ” Id. (quoting Tex. Ass'n of Mfrs., 989 F.3d at 389–90). If the agency is likely able to remedy the found errors on remand, the disruptive consequences are “only barely relevant.” Fox Television Stations, Inc. v. FCC, 280 F.3d 1027, 1049 (D.C. Cir. 2002).

Courts also must consider whether a final rule is severable. “Whether an administrative agency's order or regulation is severable, permitting a court to affirm it in part and reverse it in part, depends on the issuing agency's intent.” North Carolina v. FERC, 730 F.2d 790, 795–96 (D.C. Cir. 1984). “Severance and affirmance of a portion of an administrative regulation is improper if there is ‘substantial doubt’ that the agency would have adopted the severed portion on its own.” Davis Cnty. Solid Waste Mgmt. v. EPA, 108 F.3d 1454, 1459 (D.C. Cir. 1997) (quoting North Carolina, 730 F.2d at 795). Severance is inappropriate if invalidation of part of a rule will “impair the function of the statute as a whole.” K Mart Corp. v. Cartier, Inc., 486 U.S. 281, 294 (1988). Severance is appropriate if a final rule's unlawful parts are not “intertwined” with its lawful ones, Tel. & Data Sys. v. FCC, 19 F.3d 42, 50 (D.C. Cir. 1994), and they “operate entirely independently of one another,” Solid Waste Mgmt., 108 F.3d at 1459.

Based on these principles, the court severs and vacates 26 C.F.R. § 1.6011-10. First, the unlawful aspects of the Final Rule are severable. The Final Rule is codified in two separate sections of the Code of Federal Regulations. See 26 C.F.R. §§ 1.6011-10, 1.6011-11. The Final Rule relies on separate statutory authority for each regulation. See 26 U.S.C. § 6707A(c)(1), (c)(2). And the separate regulations “operate entirely independently of one another.” Solid Waste Mgmt., 108 F.3d at 1459. Eliminating § 1.6011-10 does not affect a taxpayer's obligation to report under § 1.6011-11. The court's conclusion that the record does not show the factual findings necessary to promulgate § 1.6011-10 has no bearing on the validity of § 1.6011-11. The regulation designating micro-captive listed transactions, 26 C.F.R. § 1.6011-10, is severable, and the appropriate remedy must be limited to that section.

Second, the equities favor a vacatur and corresponding declaratory relief. See FED. R. CIV. P. 57 (“The existence of another adequate remedy does not preclude a declaratory judgment that is otherwise appropriate.”). There is not a “serious possibility” that the agency can substantiate § 1.6011-10 as currently drafted. Tex. Ass'n of Mfrs., 989 F.3d at 389. Section 1.6011-10 is unlawful because it is not supported by the “statutorily-required findings.” Finberg, 6 F.4th at 1336. Section 1.6011-10 is based on information not directly relevant to whether a transaction is presumptively tax avoidant: the Final Rule relies on information about common features of a tax-avoidance transaction, not factors that show a transaction is more often than not tax-avoidant. To make this finding, different or altered criteria may be needed, such as a lowered modified loss ratio. Remand without vacatur is generally inappropriate if the agency is going to adopt substantively different standards. Cf. Corner Post, Inc. v. Bd. of Governors of Fed. Rsrv. Sys., 603 U.S. 799, 837 n.6 (2024) (Kavanaugh, J.) (identifying an exception for when an agency “must regulate some entity or activity more extensively”). Because there is little “prospect of the rule's being readopted upon the basis of a more adequate explanation of the agency's reasoning,” and vacatur is unlikely to cause disruptive consequences, it is the appropriate remedy. Illinois Pub. Telecomms. Ass'n v. FCC, 123 F.3d 693, 693 (D.C. Cir. 1997) (emphasis added).

Lastly, the plaintiffs' requests for injunctive relief are denied. Injunctive relief is inappropriate if vacatur affords the parties complete relief. Louisiana v. EEOC, 784 F. Supp. 3d 886, 910 (W.D. La. 2025). Because § 1.6011-10 is vacated, the defendants cannot enforce it against the plaintiffs; an injunction is not needed to protect the plaintiffs from unlawful agency action. See Franciscan All., Inc. v. Azar, 414 F. Supp. 3d 928, 946 (N.D. Tex. 2019). The plaintiffs baldly assert that an injunction is necessary because there is a “risk that Defendants will seek to implement the Final Rule's classification scheme through alternative mechanisms.” (Docket Entry No. 65 at 25). But there is nothing improper about the agency repromulgating § 1.6011-10 or a similar rule after compiling an adequate record to do so. And there “is currently no indication that, once the Rule is vacated, Defendants will defy the Court's order and attempt to apply the Rule against Plaintiffs.” Azar, 414 F. Supp. 3d at 946.

Nor is there a reason to order the defendants to “return all documents and information that Plaintiffs and their clients produced pursuant to the Final Rule.” (Docket Entry No. 59 at 35). The court upholds § 1.6011-11, which has materially identical disclosure requirements to § 1.6011-10. The plaintiffs do not argue that the defendants obtained unique information under § 1.6011-10.

For these reasons, the court declares 26 C.F.R. § 1.6011-10 unlawful, vacates it, and remands the case to the agency for further action consistent with this opinion.

IV. Conclusion

The court grants in part the plaintiffs' motion for summary judgment and a permanent injunction, (Docket Entry No. 58), and grants in part the defendants' cross-motion for summary judgment, (Docket Entry No. 63). The defendants (1) appropriately designated micro-captive transactions as transactions of interest through 26 C.F.R. § 1.6011-11; but (2) exceeded their statutory authority in designating micro-captive transactions as listed transactions through 26 C.F.R. § 1.6011-10. The court declares unlawful 26 C.F.R. § 1.6011-10 and vacates it. The case is remanded to the Department of the Treasury and the Internal Revenue Service for further agency action consistent with this opinion.

The vacatur is stayed until May 1, 2026, to avoid taxpayer confusion on Tax Day. Cf. Purcell v. Gonzalez, 549 U.S. 1, 4–5 (2006) (“Court orders ․ can themselves result in ․ confusion ․”). Final judgment is entered separately.

SIGNED on April 15, 2026, at Houston, Texas.

FOOTNOTES

1.   The IRS has not won all its captive-insurance cases. See, e.g., Rent-A-Ctr., Inc. v. Comm'r, 142 T.C. 1, 25 (2014); Securitas Holdings, Inc. v. Comm'r, 108 T.C.M. (CCH) 490 (T.C. 2014); see also AMERCO & Subsidiaries v. Comm'r, 96 T.C. 18, 42 (1991), aff'd, 979 F.2d 162 (9th Cir. 1992); Allstate Sears, Roebuck & Co. v. Comm'r, 972 F.2d 858, 864 (7th Cir. 1992); Ocean Drilling & Expl. Co. v. United States, 988 F.2d 1135, 1137 (Fed. Cir. 1993) (per curiam). But it has won all its micro-captive insurance cases before the Tax Court. See Royalty Mgmt. Ins., 2024 WL 4200129, at *1 & n.3. The court is aware of one case where a jury rejected the United States's argument that a taxpayer's micro-captive transactions did not qualify as insurance, but, in that case, the court found that the United States proved “that its positions were substantially justified,” including “its key position that the captive program did not qualify as ‘insurance’ for tax purposes.” Ankner v. United States, No. 2:21-CV-330-JES-NPM, 2024 WL 4825337, at *7 (M.D. Fla. Nov. 19, 2024).

2.   Consider the following analogy: Congress allows the Secretary of Agriculture to require persons to report if they “own birds of a type that may have blue feet.” The agency would act within its authority if it required persons to report if they owned a booby. A booby is a type of bird that could have blue feet; it could be, for example, a blue-footed booby. But a booby may also have red feet (if it is a red-footed booby) or yellow feet (if it is a brown booby). But because red-footed and brown boobies are still boobies—that is, a type of bird that may have blue feet—the agency could still require persons to report their red-footed and brown boobies.

3.   The Ryan court added an additional, persuasive point:It is hard to see any limiting principle to this argument. It is not unlike a claim that, because Congress allows for the formation of “S Corporations” and grants them pass-through taxation status if they meet certain criteria under 26 U.S.C. § 1361, any Treasury regulation that imposes listed transaction disclosure obligations on some qualifying S Corporations, see, e.g., I.R.S. Notice 2004-30, is automatically in excess of the Treasury's constitutional authority because it “undermines” Congress's approval of S Corporations generally. Under [the plaintiff's] approach, any regulation would be “in excess of authority” if imposed on an activity or entity receiving general Congressional approval, regardless of whether the regulation is authorized by Congress.Ryan, 2025 WL 3089415, at *11 n.9.

4.   The parties do not distinguish between the arbitrary-and-capricious and substantial-evidence tests. See In re Gartside, 203 F.3d 1305, 1311–15 (Fed. Cir. 2000) (discussing their differences). The parties rely on arbitrary-and-capricious principles, which this court adopts as a matter of party presentation.

5.   The plaintiffs repeatedly argue that the Final Rule inappropriately relies on Treasury and IRS “experience.” (Docket Entry No. 59 at 28; Docket Entry No. 65 at 20–21). The plaintiffs are correct that an agency's experience and expertise “do not substitute for” a reasonable “explanation.” Texas v. Biden, 10 F.4th 538, 556 (5th Cir. 2021) (per curiam) (quoting CS Wind Viet. Co. v. United States, 832 F.3d 1367, 1377 (Fed. Cir. 2016)). But “[a]n agency ․ can ‘rely on its experience, even without having quantified it in the form of a study.’ ” Safari Club Int'l v. Haaland, 31 F.4th 1157, 1173 (9th Cir. 2022) (quoting Sacora v. Thomas, 628 F.3d 1059, 1069 (9th Cir. 2010)). “Courts have not permitted agencies to rely on their ‘experience’ only when the agency fails to actually explain what that experience was and how that experience supports the promulgated regulation.” Nat'l Mining Ass'n v. United Steel Workers, 985 F.3d 1309, 1322 (11th Cir. 2021) (citing Int'l Union, United Mine Workers of Am. v. Mine Safety & Health Admin., 626 F.3d 84, 94 (D.C. Cir. 2010)). Jeffrey A. Van Hove, Associate Chief Counsel of the IRS, certified that the IRS considered materials from “taxpayer-specific files” that it could not disclose under the Code's privacy provisions. (Docket Entry No. 34-1 ¶¶ 5–6). The Final Rule's reliance on data that the IRS has reviewed and received, see Final Rule at 3,538, and expert reports that are representative of these additional files, (see Docket Entry No. 48 at AR14615–67, Docket Entry No. 48-2 at AR17715–829; Docket Entry No. 49 at AR18065–105; Docket Entry No. 51 at AR21977–79), is likely sufficient “experience” to satisfy the APA.The plaintiffs respond that the defendants should have anonymized the returns and placed them into the administrative record. (Docket Entry No. 65 at 17–18). But the “mere removal of identifying details from return information” is not enough to satisfy 26 U.S.C. § 6103(b)(2). See Church of Scientology of Cal. v. I.R.S., 484 U.S. 9, 15 (1987). Agencies do not need quantify their experience in the form of a study, see Safari Club, 31 F.4th at 1173 (citing Sacora, 628 F.3d at 1069); Nat'l Mining Ass'n, 985 F.3d at 1322, yet that is the only way the IRS could place the data on which it relies into the administrative record, see Church of Scientology, 484 U.S. at 17 (discussing permissible “statistical studies and compilations”). The IRS may have disclosed “as much information publicly” as he needed to under the APA. See Flyers Rts. Educ. Fund, Inc. v. FAA, 864 F.3d 738, 745 (D.C. Cir. 2017); (Docket Entry No. 63-1 at 28 (“The fact pattern included in the Final Rule is the anonymized, aggregated summary of the information seen in audit files.” (citing Final Rule at 3,539))). In any event, without resolving this issue, the discrete fact to which the court cites is not itself “critical” or “vital,” so the failure to disclose more fully the underlying tax data is not fatal under APA review. See Time Warner Ent. Co., L.P. v. FCC, 240 F.3d 1126, 1140 (D.C. Cir. 2001) (first quoting Cmty. Nutrition Inst. v. Block, 749 F.2d 50, 57 (D.C. Cir. 1984); and then quoting Int'l Union, UAW v. OSHA, 938 F.2d 1310, 1324–25 (D.C. Cir. 1991)).

6.   This argument displays the fundamental difference between the parties' positions. The plaintiffs are concerned about false positives, and the defendants are concerned about true positives. The plaintiffs prioritize excluding legitimate transactions, but the defendants prioritize including illegitimate transactions. This is a fundamental policy disagreement, and, under APA review, the agency's reasonable policy choice prevails. See State Farm, 463 U.S. at 43 (requiring courts to uphold rules if they can “be ascribed to a difference in view”).

7.   The plaintiffs' reply brief articulates much of this logic. The plaintiffs explain that captives “do not compete for broad third-party business and lack the marketing and distribution costs.” (Docket Entry No. 65 at 14). They add that these facts typically “drive commercial carriers' loss ratios.” (Id.). But if micro-captives do not have usual loss ratios because they are not pricing their premiums competitively—that is, pricing them at the level to cover costs and expected claims—then they are not pricing premiums as an insurance company commonly would. See Final Rule at 3,540 (citing Avrahami, 149 T.C. at 152); cf. Rent-A-Center, 142 T.C. at 12 (explaining that captives must price their premiums to earn an underwriting profit because they do not have the surplus to earn substantial investment income). And if micro-captives do not have the marketing and distribution costs that commercial carriers do, then they should have higher loss ratios that reflect lower premium rates enabled by lower costs. (See Docket Entry No. 59 at 3 (explaining that micro-captives offer “lower and more stable premium prices”)). The agency was reasonable to think that even a slightly below-average modified loss ratio warrants a second look because the insured may be paying above-market premiums (or market-rate premiums that could have been negotiated to a lower-but-profitable captive price), and doing so only because those premiums go to its captive and generate federal tax benefits. See Rsv. Mech. Corp., 34 F.4th at 914–15 (reviewing whether transactions were negotiated at arm's length and whether a reasonable party would pay the premiums charged); cf. Patel, 2025 WL 3158814, at *13–15 (concluding that unreasonable premium payments to a micro-captive lacked economic substance and were only for federal tax benefits). The agency is also entitled to investigate whether the delta between market-priced commercial premiums and at-cost-priced captive premiums should have been taxed, not deducted as an “ordinary and necessary expense[ ].” 26 U.S.C. § 162(a).

8.   The plaintiffs' reliance on Puglisi and R.V.I. fares no better. The plaintiffs, like many commenters, point to the IRS's concession in Puglisi v. Commissioner, No. 13487-20, 2021 WL 7162530, at *9–10 (T.C. Oct. 29, 2021), as proof that loss ratios of around 15% are legitimate. (Docket Entry No. 65 at 14 n.4). But the IRS made no concession on the merits in Puglisi, see 2021 WL 7162530, at *6–10 (opposing the taxpayer's motion for entry of an opinion on the merits), and, as the Final Rule explained, the “IRS's decision to concede or settle a given case in no way alters” its findings based on the mine run of them, Final Rule at 3,540; see also Puglisi, 2021 WL 7162530 at *9 (discussing the tactical considerations of conceding or litigating the specific case). The plaintiffs also argue that the tax court's ruling in R.V.I. undermines the Loss-Ratio Factor because it held that a 30% loss ratio represents a “significant risk of loss.” 145 T.C. at 228. But the policies at issue in R.V.I. concerned “catastrophic-type coverage” of “low-frequency, high-severity risks.” Id. The Final Rule does not dispute that a 30% loss ratio for those insurance risks may be legitimate. See Final Rule at 3,541 (discussing “low-frequency, high-severity risks”). R.V.I. does not speak to the reasonableness or suspiciousness of loss ratios for other insurance risks relevant to the loss-ratio calculation that the Final Rule used. See id. at 3,541–43 (discussing the modifications to the NAIC data). R.V.I. shows only that the Loss-Ratio Factor is overinclusive, which, as the court already discussed, does not warrant vacatur under the APA.

9.   The plaintiffs make a final argument that the agency failed to consider comments that the reporting requirement would have a detrimental effect on the micro-captive industry. (Docket Entry No. 59 at 32–33; Docket Entry No. 65 at 22). The plaintiffs argue that the Loss-Ratio Factor would incentivize captives to “limit their surplus reserves—a move that jeopardizes their ability to pay future claims (and may also violate state regulations).” (Docket Entry No. 59 at 33). The Final Rule considered and reasonably rejected this argument as “misplaced.” Final Rule at 3,545. The Final Rule explains that captives should and “would avoid insolvency in the same way they always have; that is, by insuring risks that are selected and duly reserved for in accordance with sound business judgement and the regulatory requirements of their domicile.” Id. The conclusion was that nothing in the Final Rule would “encourage[ ] ․ micro-captives to make contractual promises that exceed risk-bearing capabilities.” Id. The Final Rule stresses that it does not “define insurance for Federal tax purposes,” id. at 3,546, and that its reporting requirements are likely to impose “minimal” and “low” burdens on taxpayers, id. at 3,558. There is little incentive to avoid a minimal reporting requirement that does not affect the legitimacy of a claimed tax benefit—unless the taxpayer believes it may not be providing legitimate insurance as defined in the relevant case law. In such a case, the reporting requirement will have the intended and permissible effect of deterring tax evasion. See id. at 3,538 (explaining that the Final Rule does not “harass otherwise valid businesses” but supports the “IRS's long-standing positions with respect to abusive micro-captives”).

Lee H. Rosenthal Senior United States District Judge