Bills / Cases / IRS

Reserve Mechanical Corp. v. Commissioner; No. 18-9011

RESERVE MECHANICAL CORP., f/k/a Reserve Casualty Corp., Petitioner-Appellant, v. COMMISSIONER OF INTERNAL REVENUE,
Respondent-Appellee.

ALABAMA CAPTIVE INSURANCE ASSOCIATION, INC.; ARIZONA CAPTIVE INSURANCE ASSOCIATON, INC.; DELAWARE CAPTIVE INSURANCE ASSOCIATION INC.; GEORGIA CAPTIVE INSURANCE ASSOCIATION, INC.; HAWAII CAPTIVES INSURANCE COUNCIL; KENTUCKY CAPTIVE ASSOCIATION, INC.; MISSOURI CAPTIVE INSURANCE ASSOCIATION; MONTANA CAPTIVE INSURANCE ASSOCIATION, INC.; NORTH CAROLINA CAPTIVE INSURANCE ASSOCIATION,; UTAH CAPTIVE INSURANCE ASSOCIATION; SELF INSURANCE INSTITUTE OF AMERICA,

Amici Curiae.

PUBLISH

UNITED STATES COURT OF APPEALS
FOR THE TENTH CIRCUIT

Appeal from the Commissioner of Internal Revenue
(CIR No. 014545-16)

Val J. Albright, Foley & Lardner, LLP, Dallas, Texas (Michelle Y. Ku, Foley & Lardner, LLP, Dallas, Texas, E. John Gorman, Logan R. Gremillion, and Coby M. Hyman, The Feldman Law Firm LLP, Houston, Texas, with him on the briefs) for the Petitioner-Appellee.

Geoffrey J. Klimas, Attorney, Tax Division (Richard E. Zuckerman, Principal Deputy Assistant Attorney General, Joshua Wu, Deputy Assistant Attorney General, Francesca Ugolini, Attorney, Arthur T. Catterall, Attorney, Tax Division, with him on the brief), Department of Justice, Washington, D.C., for Respondent-Appellee.

Elizabeth J. Bondurant (Jonathan Reid Reich, with her on the brief), Womble Bond Dickinson (US) LLP, Atlanta, Georgia, filed a brief for Amici Curiae The Alabama Captive Insurance Association, Inc., Arizona Captive Insurance Association, Inc., Delaware Captive Insurance Association Inc., Georgia Captive Insurance Association, Inc., Hawaii Captives Insurance Council, Kentucky Captive Association, Inc., Missouri Captive Insurance Association, Montana Captive Insurance Association, Inc., North Carolina Captive Insurance Association, Utah Captive Insurance Association, and Self Insurance Institute of America.

Before HARTZ, HOLMES, and PHILLIPS, Circuit Judges.

HARTZ, Circuit Judge.

Reserve Mechanical Corp. appeals the decision of the Tax Court affirming the decision of the Commissioner of Internal Revenue that it did not qualify for an exemption from income tax as a small insurance company and that the purported insurance premiums it received must therefore be taxed at a 30% rate under I.R.C. § 881(a). We hold that the record supports the Tax Court's decision that the company was not engaged in the business of insurance. The court had two grounds for deciding that Reserve was not an insurance company. First, it determined that Reserve had not adequately distributed risk among a large number of independent insureds — a hallmark of any true insurance company. Virtually all the insured risk was that of one insured, a company that had the same ownership as Reserve itself. To appear to distribute risk, Reserve entered into an insurance pool with other purported insurance companies, each owned by an affiliate of its insured, but the arrangement lacked substance and the pool itself did not distribute risk. Second, the Tax Court determined that the policies issued by Reserve were not insurance in the commonly accepted sense. For example, the premiums were not the result of arm's-length transactions and were not reasonable, and Reserve was not operated in the way legitimate insurance companies operate. In addition, Reserve argues that if it was not an insurance company, the premiums it received must be treated as nontaxable capital contributions. We also reject that argument.

I. OVERVIEW

From 2008 to 2010, when Reserve Mechanical Corp. was known as Reserve Casualty Corp., it issued a number of insurance policies to Peak Mechanical Corp. Two men, Norman Zumbaum and Cory Weikel, owned both Reserve (through Reserve's parent corporation, Peak Casualty) and Peak. Before these policies were issued, Peak had limited its insurance coverage to commercial policies that cost about $100,000 a year. Peak maintained those policies but also paid Reserve more than $400,000 a year for the supplemental insurance obtained through the new policies. The relationship between Reserve and Peak is often termed "captive" insurance. See 3 Steven Plitt et al., Couch on Insurance § 39:2 (3d ed. 2021) ("A captive insurer is a corporation organized for the purpose of insuring the liabilities of its shareholders or their affiliates." (internal quotation marks omitted)).

Peak did not appear to get much in return for its $400,000 annual payment to Reserve. The appellate record indicates that Peak recovered on only one loss, receiving a payment of slightly less than $340,000; and even then, as we shall see, the bona fides of the claim were questionable and the handling of the claim was highly irregular. The high premiums on the policies could, however, be a significant financial benefit to Zumbaum and Weikel even if — indeed, especially if — Peak never suffered a loss covered by the policies issued by Reserve. The benefit arises from the tax treatment of small insurance companies, which has special consequences when the small insurer is a captive insurer, sometimes referred to as a "micro-captive." As the Supreme Court recently explained:

A micro-captive transaction is typically an insurance agreement between a parent company and a "captive" insurer under its control. The [Internal Revenue] Code provides the parties to such an agreement with tax advantages. The insured party can deduct its premium payments as business expenses. And the insurer can exclude . . . those premiums from its own taxable income, under a tax break for small insurance companies. The result is that the money does not get taxed at all.

CIC Servs., LLC v. IRS, 141 S. Ct. 1582, 1587 (2021) (citations omitted). Thus, Peak could treat the $400,000 in annual premiums it paid to Reserve as a deductible business expense on its federal income-tax returns, while Reserve would be exempt from income taxation so long as it qualified as an insurance company under the tax laws. (Reserve relied on I.R.C. § 501(c)(15), which exempts insurance companies from income taxation under § 501(a) if they receive no more than $600,000 a year in premiums.) The $400,000 moved from one entity owned by Zumbaum and Weikel to another entity they owned; so, pre-tax, they had the same wealth despite the transfer. But their businesses paid significantly less tax. In particular, the more paid in premiums on the insurance policies, the greater the tax deduction, so there would be a strong financial incentive for those who owned both the business and its captive to set the premiums as high as possible, unlike the usual incentive for a business to reduce its expenses. Such tax benefits and incentives have led micro-captive transactions to come under scrutiny because of "their potential for tax evasion." CIC Servs., 141 S. Ct. at 1587.

Capstone Associated Services, Ltd., which consulted for and managed a number of captive insurance companies besides Reserve, advised Zumbaum and Weikel in creating Reserve and handled the technical and management issues, such as preparing policies and recommending premiums. It believed that for Reserve to be a qualified insurance company it would have to receive at least 30% of its premiums from companies not affiliated with it, a threshold we can assume to be correct for purposes of this appeal.

In the Background section of this opinion we will describe in some detail how Reserve purported to obtain this diversification of risks. But it may be useful to orient the reader by sketching the key aspects of the arrangement now. Capstone ostensibly created diversification of risks in two ways, which together accounted for about 30% of the "premiums" received by Reserve. First, it arranged for 50-some captives under its management to, in essence, be liable on reinsurance policies issued to each other. In a reinsurance relationship one insurance company, the reinsurer, acts as an insurer of another insurance company; typically, the reinsured insurance company pays a premium to the reinsurer and the reinsurer assumes a portion of the liabilities of the reinsured company on the insurance policies issued by the reinsured company — that is, the reinsured company "cedes" some of its liability to the reinsurer. See Black's Law Dictionary 1539 (11th ed. 2019) (defining reinsurance as "[i]nsurance of all or part of one insurer's risk by a second insurer, who accepts the risk in exchange for a percentage of the original premium"); 13A John Alan Appleman & Jean Appleman, Insurance Law and Practice § 7681, at 480 (1976) ("Reinsurance, to an insurance lawyer, means one thing only — the ceding by one insurance company to another of all or a portion of its risks for a stipulated portion of the premium, in which the liability of the reinsurer is solely to the reinsured, which is the ceding company, and in which contract the ceding company retains all contact with the original insured, and handles all matters prior to and subsequent to loss."); but cf. Colonial Am. Life Ins. Co. v. Comm'r, 491 U.S. 244, 247 (1989) (adopting a more expansive notion of reinsurance). For example, a company that issues homeowners insurance may pay a premium to a reinsurer to protect the homeowner-insurance company from unsustainable losses if a major fire destroys too many homes insured by the company.

The reinsurance arranged by Capstone was accomplished through PoolRe Insurance Corp., another company managed by Capstone. Through PoolRe each of the captive insurers in effect reinsured, and was reinsured by, each of the other captives, with PoolRe acting as the intermediary. See Figure 1.

The process involved two steps. To begin with, on each policy issued by Reserve and the other captives, PoolRe provided what was termed stop-loss coverage (purportedly to protect the captive insurers from excess losses) through an endorsement on the policy that required PoolRe to assume a portion of the liability incurred by the captive insurer on the policy covering the insured (such as Peak). The restrictions on payment under the stop-loss coverage, which will be explored later, were sufficiently intricate and demanding that it appears they were never satisfied during the years in question (either on the stop-loss coverage for Reserve policies or the stop-loss coverage provided for the other captives), so there were no payouts on the coverage. For this stop-loss coverage, PoolRe received a fixed percentage (18.5% the first year) of the premiums paid on the policies issued by the captives.1 As the second step, the captives in turn reinsured all of PoolRe's stop-loss coverage, with each captive receiving a premium from PoolRe equal to the premium its insured paid to PoolRe. See Figure 2.

The net result of this arrangement was that each captive insurer received the full premium paid by its insureds — the 81.5% paid directly to the captive by the insured plus the 18.5% paid to PoolRe, which in turn later paid that amount to the captive. Through this arrangement, all the captive insurers shared the entire risk of the stop-loss coverage provided by PoolRe. If one of the captives incurred liability that triggered the stop-loss coverage, PoolRe would pay its share of the loss but would be fully reimbursed through the reinsurance it obtained from the captives as a whole, with each captive paying its proportionate share. As previously mentioned, however, this risk apparently never materialized. At least during the period relevant to this appeal, PoolRe did not need to pay on any stop-loss coverage, so the payments to and from PoolRe were a wash. But Reserve could argue that 18.5% of the premiums paid by Peak came to Reserve from PoolRe, rather than from its affiliate (Peak), and that it thereby distributed risk beyond its affiliates.

The second way in which Capstone arranged for the captives to ostensibly diversify risks was by purportedly arranging for each captive to reinsure a small percentage of risk that PoolRe assumed from coinsuring thousands of vehicle-service contracts with another insurance company. See Black's Law Dictionary 954 (11th ed. 2019) (providing one definition of coinsurance as "[i]nsurance provided jointly by two or more insurers"). Reserve claimed to have received about 15% of its premiums through this arrangement. In each year for which we have a record, Reserve reported that it incurred liability from this reinsurance approximately equal to the amount it was owed in premiums. Taken together, the premiums from both these plans constituted more than 30% of the premiums Reserve received from Peak and PoolRe.

Questioning the bona fides of Reserve's various arrangements, the IRS rejected Reserve's claim to be a qualified insurance company and assessed it for back taxes. Reserve challenged the assessment in the United States Tax Court but lost. Reserve has appealed to this court. Exercising jurisdiction under 26 U.S.C. § 7482(a), we affirm.2 In the Tax Court proceedings Reserve had the burden of proving that the IRS's assessment was incorrect. On the record before it, the court could properly find that Reserve had not satisfied its burden — in particular, Reserve had not proved that its purported insurance transactions were truly arrangements for insurance. We also reject Reserve's challenge to the Tax Court's refusal to accept Reserve's claim that its receipts from Peak were, if not insurance, nontaxable capital contributions.

Our discussion will proceed as follows: First, we describe at length the facts relevant to the issues before us. Second, we review the applicable law and the decision of the Tax Court holding that Reserve was not an insurance company. Third, we explain why we affirm that holding. Fourth, we reject Reserve's argument that it owes no taxes even if it was not an insurance company because the "premium" payments from Peak must then be deemed nontaxable contributions of capital.

II. BACKGROUND

In this section we will discuss the formation of Reserve, the policies it issued to Peak, the reinsurance arrangements it made with PoolRe that ostensibly allowed it to distribute risk to entities not affiliated with Peak, and the tax claim instituted against Reserve.

We recite the evidence of record in some detail. The amicus brief submitted to this court, as well as the briefs of Reserve itself, suggest that the decision of the Tax Court undermines perfectly proper practices in the creation and operation of captive insurance companies. But the specific evidence presented can make all the difference. Two transactions that appear similar in form may be treated quite differently under the law because of differences in the underlying substance. We do not hold that the forms of the transactions involving Reserve are improper (for example, insurance pools such as PoolRe may be perfectly legitimate in other circumstances); we hold only that the Tax Court could properly conclude that they were not insurance transactions in substance.

A. Formation of Reserve

Zumbaum and Weikel formed Peak in the mid-nineties to operate near deep underground mines in Idaho's Silver Valley. Peak continues to do business in Idaho, primarily on the Bunker Hill Superfund Site, but also in other locations in Idaho and Nevada. It manufactures and sells equipment that supports underground mining operations. This equipment includes ventilation fans that control the temperature of the mines, submersible pumps that remove the groundwater from the mines, and vehicles that transport workers, explosives, and fuel to, from, and within the mines. Peak also repairs and cleans mining equipment that is often contaminated with hazardous waste like lead or zinc. Because the cleaning operations risk creating hazardous-waste runoff, Peak employs various measures, such as the use of cleaning bays, sumps, and containment areas, to prevent spreading the contaminants.

In 2008 and 2009 Peak had 17 employees, including two shop managers, ten shop staff, and two outside salespersons. By 2010 it was down to 13 employees. Zumbaum and Weikel also owned two other business entities: RocQuest holds the real estate that Peak leases for its operations, and ZW was created by Zumbaum and Weikel to lend money to Zumbaum's secretary when she wanted to buy a bar in Silver Valley. At trial Zumbaum described the operations of RocQuest and ZW as insignificant.

Before obtaining insurance from Reserve, Peak relied on several commercial policies for its insurance needs: Most recently it had paid premiums of $95,828 for 2007, and $57,300 for the first half of 2008. The coverage limits for the policies ranged from $5,000 to $2 million. Peak filed few claims under these policies — some claims under its auto-insurance policies and a claim under its commercial-property policy for snowstorm damage to the roof of one of Peak's buildings (Peak spent $25,000 to replace the roof but recovered only $2,000 from the insurer). Although Peak claimed that it was unhappy with its insurers' handling of these claims, Peak continued its policies with them, even after procuring the additional coverage from Reserve.

Zumbaum and Weikel reached out to Capstone after a mentor recommended taking a look at forming a captive insurance company. Stuart Feldman, chief executive officer of Capstone's general partner, Capstone Holdings Corp., and Lance McNeel, Capstone's director of insurance operation, conducted an on-site visit of Peak in August 2008. Before the visit Peak provided Capstone with background documents on Peak's finances, taxes, and current insurance. The visit lasted six to eight hours. McNeel and Feldman toured Peak's facilities and discussed Peak's business operations and insurance risks.

After the visit Capstone began preparing a "Captive Insurance Company Feasibility Study" to "evaluate[ ] Peak's desire to explore the option of forming a captive insurer for the purpose of writing coverages that are generally unavailable or impractical to obtain in the conventional insurance marketplace." Aplt. App., Vol. 7 at 2027, 2030 (emphasis added). For reasons not explained in the record, Capstone did not produce the final version of the study until August 2009. The study outlined advantages of captives, such as "lower risk costs," investment income, tailored policies, coverage prices that "track closely with the risks inherent in an insured's own exposures," access to reinsurance, and "complete control over the operation of [the] captive[ ]." Id. at 2040-42. It said that "[c]overage lines that address reasonably predictable, non-catastrophic exposures are good candidates for coverage by a captive," and that other "unpredictable exposures may also be good candidates," but that those exposures would likely require "pooling or reinsurance." Id. at 2040.

As for Peak specifically, the study noted that Peak's "current conventional policies . . . offer broad and comprehensive coverage that is appropriately designed and priced," id. at 2047, and it acknowledged that Peak "had no losses of any significance" on its current policies, id. at 2061. But it stated that a captive could insure against additional risks, and it mentioned 13 categories of such risks, describing each in one to four sentences. There was no discussion of the likelihood of any risk. The study did not mention either ZW or RocQuest.

Zumbaum and Weikel did not wait for the final feasibility study before beginning their insurance project. On December 3, 2008, about four months after the site visit, they incorporated Reserve (as a subsidiary of their holding company named Peak Casualty) in Anguilla, British West Indies, with an initial capital investment of $100,000. Reserve had no employees and was managed by Capstone. Also in December, Reserve issued its first set of policies.

B. Reserve's Direct Policies

Reserve issued 13 policies to Peak (RocQuest and ZW were also named insureds, but we will generally refer only to Peak) on December 4, 2008:

1) Excess Directors & Officers Liability (for liability for wrongful acts committed by directors and officers acting in such capacity)

2) Special Risk — Loss of Major Customer (for reduction in net income caused by loss of major customer)

3) Special Risk — Expense Reimbursement (for expenses to mitigate adverse publicity arising from incidents such as a liability incident, product recall, labor dispute, or bankruptcy; and civil-liability defense costs if there was no underlying insurer or all defense expenses have been exhausted)

4) Special Risk — Loss of Services (for loss of services of employees to be specifically named)

5) Special Risk — Weather Related Business Interruption (for losses from interruption of business caused by weather)

6) Excess Pollution Liability (for cost of cleaning up on-site pollution and liability for creating pollution)

7) Special Risk — Tax Liability (if tax liability exceeds 115% of filed tax liability)

8) Excess Intellectual Property Package (for liability for wrongful acts by Peak and for damage to Peak's intellectual property caused by wrongful acts of others)

9) Special Risk — Regulatory Changes (for damages to business from changes in the law)

10) Special Risk — Punitive Wrap Liability (for punitive damages that would be paid by one of the other Reserve policies to Peak except that a law or judicial ruling precludes insuring punitive damages)

11) Excess Employment Practices Liability (for liability for wrongful discharge, workplace harassment, retaliation for exercising employment-related legal rights, breach of employment contract, etc.)

12) Excess Cyber Risk (for liability and for business and property loss caused by others)

13) Special Risk — Product Recall (for expenses of recall of products manufactured or sold by Peak)

For each policy the policy period was less than a month, extending from December 4, 2008, to January 1, 2009, and the liability limit was $1 million. The total premium was $412,089.02. The policies were claims-made policies: The claim must have been based on acts, errors, or omissions after the policy inception date or, if applicable, after the earlier retroactive date set forth in the policy. And the claim must have been made and reported to Reserve after the policy inception date and during the policy period or, if applicable, during the extended reporting period set forth in the policy. Policy # 7 — the tax-liability policy — had special provisions on retroactivity and reporting, apparently intending to cover tax returns due before 2009 if covered losses were reported within four years of the due date. Six policies (1, 6, 8, 10, 11, 12) had a retroactive date of January 1, 2005 (so the act or omission giving rise to the claim could have predated the policy by as much as four years) and an extended reporting period of three or four years (so the claim could have been made and reported to Peak as late as December 2012). The remaining policies had no retroactive date but extended reporting periods of one year or, in one case (policy #13), four years. Each policy contained an other-insurance clause, which provided, "The limits and deductibles stated herein only apply after coverage is exhausted from any and all other valid insurance policies issued by any other insurer," Reserve Mech. Corp. v. Comm'r, T.C. Memo. 2018-86 at 14, 115 T.C.M. (CCH) 1475 (T.C. 2018) (Reserve) (capitalization omitted);3 so if Peak suffered a loss that could be covered by both a Reserve policy and one of its commercial policies, Reserve would pay nothing unless the claim exhausted the benefits under the commercial policy.

Several of these policies were executed with singular carelessness. For example, two of the policies erroneously listed Pacific Arts Entertainment, LLC and Pacific Arts Presents, LLC as the insureds, rather than Peak, RocQuest, and ZW. And although the directors-and-officers policy stated that it covered the specific officers and directors listed in Schedule 1-A, an attachment to the policy, Schedule 1-A did not list a single insured person, so the policy — for which Peak paid $17,122 — would provide no coverage. Also, in the apparent rush to issue the policies (and pay premiums that would be deductible in 2008), Peak paid for three policies — employment-practices liability, weather-related business disruption, and cyber risk — that apparently were deemed unnecessary after a little further consideration, as they were dropped in 2009, after being in place for less than one month.4 (Notably, the later-issued feasibility study described one of the discontinued coverages — employment-practices liability — as a "major liability concern[ ]" for Peak. Aplt. App., Vol. 7 at 2062.)

1. Policy Premiums

There are also remarkable errors in pricing the premiums on two policies. McNeel, Capstone's director of the insurance operation, prepared a rating worksheet for each of the policies. On the 2008 worksheet, for example, one column set the annual premium for each policy, and an adjacent column contained a pro rata percentage to account for how premiums calculated on an annual basis should be adjusted for retroactive and partial-year coverage. For those policies with retroactive coverage the pro rata percentage was usually 95%. For four of the policies (loss of major customer, loss of services, product-recall reimbursement, and weather-related business interruption) with no retroactive coverage (so the occurrence had to be during December 2008) the pro rata percentage was 10% (presumably reflecting that the coverage was for occurrences during less than 10% of a full year). But for Special Risk — Regulatory Changes, which also had no retroactive coverage, the pro rata percentage was 95%. When asked about this at oral argument, counsel for Reserve responded, "It certainly strikes me as an error." Oral Arg. at 37:30.5 A similar error was made with respect to the policy for Special Risk — Expense Reimbursement.

Perhaps more important, the manner of arriving at the premium prices on McNeel's rating worksheet was questionable. The core task in setting premiums for an insurance policy is predicting risk: the size and frequency of losses covered by the policy. See Owens v. Aetna Life & Cas. Co., 654 F.2d 218, 240 (3d Cir. 1981) ("The [insurance] company must set its premiums based on its prediction of two cost variables: the probability of a particular risk of loss occurring and the magnitude of the loss if it occurs."). But the record is devoid of evidence of the necessary risk assessment by Reserve. Peak had no history of any losses that would be covered by the Reserve policies, so the premiums could not be based on Peak's actual experience. The feasibility study briefly described the risks that would be covered by the policies, but it contained no discussion of the probability or size of the risks. For example, when the feasibility study discussed Peak's need for employment-practices liability coverage, it merely stated that this liability "has become a hot topic over the past several years as complaints and legal action nationwide have skyrocketed for wrongful termination, discrimination, harassment, and other employment-related practices." Aplt. App., Vol. 7 at 2056. (Again, Peak dropped its excess-employment-practices-liability coverage after one month, before the feasibility study was issued.)

One source of information would have been insurance-industry data. Yet even though one of Reserve's expert witnesses testified that commercial insurance policies were available as an alternative to several of the Reserve policies, there is no evidence that anyone compared the rates on such policies or otherwise considered industry standards. Instead, the record suggests that McNeel based the rates on the premiums charged by other captive insurers managed by Capstone. For the most part those premiums were apparently based solely on the annual projected sales revenue of the insured — for example, regardless of the particulars of the business (nature of the business, location, etc.) the premium for the coverage depended only on the projected sales revenue. The only exception was employment-practices-liability coverage, the premiums for which were based on the number of employees. McNeel prepared a document that summarized the premium rates per $250,000 of coverage on all the different policies issued by Capstone entities. The document, in the words of the Tax Court, "provided both an average and a range of rates from which [McNeel] could choose for each type of policy." Reserve, at 15. But McNeel's testimony contained no suggestion that he looked at the businesses insured under the other policies to see if they would be likely to have risks comparable to those of Peak. On the contrary, he acknowledged that he was unaware whether the other captives were in a business similar to Peak, or were similar in other respects to Peak. Nor was there any evidence showing that the premiums charged by the other Capstone entities (who likely had the same tax incentive as Reserve to charge as high a premium as possible) were themselves reasonable.

Reserve claims that McNeel's proposed premiums were supported by the independent work of Mid-Continent General Agency, Inc., "a managing general underwriter," which provided what are called "pricing indication[s]," Aplt. App., Vol. 5 at 1209, 1242, that closely correlated with McNeel's suggested premiums. But how these indications were generated was left unexplained. McNeel testified that someone from Mid-Continent visited Capstone's office for about 10 days and looked at its files, but what files in particular are not specified. No representative from Mid-Continent testified at trial. And the only documentation offered to show how Mid-Continent arrived at these prices was a letter from Mid-Continent to McNeel that purported to "comment on the methodology [Mid-Continent] use[s] to assist in developing premium quotations." Aplt. App., Vol. 12 at 3560. The letter, however, stated merely that Mid-Continent's methods "involve[ ] the evaluation of exposures for a given line of insurance, examination of historic loss date [sic], if any, the consideration of increased limits factors6 . . ., and acknowledgement of market rate adjustments." Id. at 3560-61. The letter did not provide further details, and it did not mention Peak or Reserve's policies.

We recognize that Reserve called two experts as witnesses to defend the pricing of the Peak policies, Esperanza Mead and Michael Solomon. Although Mead testified that the Reserve premiums were reasonable, her conclusions rested on questionable assumptions. Reserve had no loss data for her to use nor did she use any risk information from general industry sources. And to the extent that she relied on the premise that the premiums for the policies issued by the other Capstone captive insurers were reasonable, that premise was not supported by loss data from those insurers. Mead's report does tabulate the claims payments by Capstone captives for a five-year period; but she assessed this information as having "low credibility" for actuarial purposes because so few claims were paid, and the information had no material effect on her conclusions. Aplt. App., Vol. 5 at 1256. She instead reached her conclusion that the premiums charged by Reserve and the other Capstone captives were reasonable only by making the assumption, unsupported by any data, that the captives would ultimately pay claims equal to about 75% of the premiums. She employed that figure solely because it is standard in the insurance industry (which means merely that 25% of insurance premiums ordinarily go to overhead, administrative expenses, and profit). Under this assumption one would expect Peak, a business with almost no history of insurance losses and only general concerns about the future, to average over $300,000 in submitted claims per year going forward. Neither Reserve nor its experts offered data, or even an explanation, to justify this assumption.

As for Solomon, at least he looked at real-world data — for example, he assessed the risk of loss to Peak from recall of products it manufactured or sold by looking at the "annual economic impact of food safety outbreaks." Aplt. App., Vol. 16 at 4724. But he opined on only six of the 13 policy coverages; and rather than assessing the reasonableness of the premiums for each coverage, he said only that the total premium for the reviewed coverages was "reasonable in the aggregate." Aplt. App., Vol. 13 at 3889. In any event, his report (and Mead's as well) was prepared many years after issuance of the policies of concern on this appeal, and there is no evidence in the record that the premiums for those policies were based in any way on the data he used (much of which in fact postdated the policies). Neither expert examined what the real risks to Peak were.

As summarized in the Commissioner's brief:

[N]either Solomon nor Mead compared taxpayer's premiums to premiums charged by unrelated third-party insurers; therefore, neither had a basis to determine whether taxpayer's premiums were commercially reasonable. Solomon merely opined that the aggregate premiums charged by taxpayer were reasonable when compared to Capstone's internal pricing guidelines. Similarly, Mead merely opined that the aggregate premiums charged by taxpayer were reasonable when compared to rates charged by other Capstone-managed captives.

Aplee. Br. at 60-61 (citations omitted). Reserve's reply brief does not respond to, much less challenge, the quoted statement. Although the testimony of the two experts is quite technical, our understanding of the testimony supports what the Commissioner says.

2. Comparison to Commercially Available Policies

Nor is there evidence of any effort to determine what other insurers might have charged Peak as premiums for similar policies. True, Capstone's feasibility study stated that Peak wanted to create a captive insurer "for the purpose of writing coverages that are generally unavailable or impractical to obtain in the conventional insurance marketplace." Aplt. App., Vol. 7 at 2030 (emphasis added). But the report of another of Reserve's experts, Robert Snyder, who in fact co-authored that feasibility study, said that at least six coverages provided to Peak by Reserve — pollution liability, intellectual property, "commercial property gap," "punitive damages wrap," cyber risk, and tax liability coverages — were available in the commercial market. Aplt. App., Vol. 12 at 3582-85 (capitalization omitted).

Zumbaum and Weikel were apparently not particularly interested in Peak's saving money on insurance premiums. Zumbaum admitted that he did not explore coverage on the commercial insurance market before creating Reserve. And although McNeel and Feldman both testified that Zumbaum and Weikel were the most significant people in determining the policies, pricing, and coverage, McNeel could not recall whether Zumbaum or Weikel questioned the premium numbers. Zumbaum testified that he did not review the Reserve policies in detail; instead, he testified that he "[p]robably scanned through them." Aplt. App., Vol. 4 at 1134. Zumbaum also admitted that he relied on "Mr. McNeel, Mr. Feldman or other folks from Capstone" to tell him about the policies and that he would "believe[ ]" whatever they told him. Id.

Aside from the premium calculations, there is also no evidence in the record that the choice of policies, or their specific contents, was based on an assessment of Peak's particular needs. The feasibility study stressed that captive insurers are ideal for providing so-called "manuscript[ ]" policies, which it defined as policies with "coverage[ ] to address specific concerns." Aplt. App., Vol. 7 at 2031. And Feldman testified that one advantage of captive insurers was that captives "design[ ] coverages not with a blunt instrument but with a scalpel" — effectively "tailoring coverages directly to the needs of the business." Aplt. App., Vol. 5 at 1372. But there is no evidence of such tailoring for the policies issued to Peak. McNeel acknowledged that Capstone "had policies that were common to a lot of different clients." Id. at 1250. And the Tax Court described these as "'cookie-cutter'" policies, noted testimony from Capstone employees that "they administered many of the same policies for all of their clients," and observed that "[i]n many instances the [Peak] policies were not reasonably suited to the needs of the insureds, particularly Rocquest and ZW." Reserve, at 54. (Of course, form policies have a very large, and quite proper, role in the insurance industry. In themselves, they do not suggest impropriety.)7

3. Claim-Handling

In addition, Reserve's claims-handling operations were, to say the least, far from businesslike. Reserve handled only one claim during the years at issue — a claim from Peak under the special-risk policy for partial loss of a major customer. That policy, as the name suggests, reimbursed Peak for business losses "as a result of the loss of or reduction of services of a [m]ajor [c]ustomer." Aplt. App., Vol. 12 at 3302. But the policy imposed significant restrictions on when coverage would be provided. For example, it would not cover the loss of a major customer if Peak "initiate[d] the termination," or if Peak did "not attempt or intend to replace" the customer, or if Peak caused the customer to leave by breaching the terms of its contract with the customer. Id. And the claim would not be paid under the policy if the insured had knowledge of the loss before the effective date of the policy.

Peak filed its claim under the major-customer policy on April 6, 2009, about four months into the coverage period. Reserve's initial notice of claim dated the loss as occurring during the first week of the policy period, on January 5, 2009. It described the claim as arising from a "[s]ignificant reduction of orders from Stillwater Mining Company, a customer representing 35% of sales." Aplt. App., Vol. 19 at 5407. It said that "[t]he reduction in orders from Stillwater represents 16% of the insured's sales for the period," and said that its net income had dropped from $422,440 in the first quarter of 2008 to $232,620 in the first quarter of 2009, a drop of about 45%. Id. After subtracting the $25,000 deductible, the claim value for lost net income was $164,820. The underlying claim document that Peak allegedly submitted to Reserve to substantiate and explain these numbers is not in the record. In fact, the record contains no supporting documentation from Peak whatsoever. And Zumbaum gave conflicting testimony about whether Peak submitted the claim to Reserve or to Capstone. Moreover, there is no evidence that Reserve investigated the claim to determine if the loss was covered under the policy — did Peak know of the coming reduction in sales by January 1, 2009, the effective date of the policy;8 was the reduction caused by a breach of contract by Peak; etc.?

Reserve paid $150,000 on the claim on April 21, 2009 — about two weeks after Peak filed the claim. One month later, on May 27, Reserve and Peak executed a settlement agreement. The agreement (which misstates the coverage limit for the relevant policy as $1 million, rather than the actual $500,000) stated, among other things, that Reserve would pay Peak the full value of the claim — $164,820 — and in return, Peak would "completely release[ ] and forever discharge[ ] [Reserve] from any and all past or present claims, demands, obligations, actions, causes of action, judgments, expenses and compensation . . . [that] may in any way grow out of the specified loss." Aplt. App., Vol. 12 at 3553. Reserve then issued a second payment of $14,820 the same day as the agreement. Four months later, however, on September 10, 2009, Reserve issued a third check to Peak for $175,000. The sole possible explanation in the record is a line dated August 25, 2009, on the Notice of Claim form stating: "Re-open claim due to extended losses from lost customer." Aplt. App., Vol. 19 at 5407. There is no supporting documentation for the $175,000 sum. All three checks from Reserve to Peak were signed by a Peak employee. More than two years later, on January 30, 2012, the parties executed an addendum to the settlement agreement, which stated in full: "In consideration of the release set forth above, the Insurer agrees to pay the Insured the sum of Three Hundred Thirty Nine Thousand, Eight Hundred Twenty Dollars ($339,820.00). The foregoing sum [has] been paid by Insurer through two checks made payable to 'Peak Mechanical & Components, Inc.'" Aplt. App., Vol. 12 at 3556.

C. Reserve's Reinsurance Policies

Capstone believed that Reserve would need to obtain at least 30% of its premiums from insureds not related to it — that is, from insureds other than Peak — in order to receive favorable tax treatment as a qualified insurance company. To meet the 30% threshold Reserve therefore engaged in two arrangements with PoolRe, each of which contributed about 15% of the total premiums it received. PoolRe managed similar arrangements for some 50 other captive insurance companies that issued about 400 policies altogether. As previously noted, these captive insurance companies as well as PoolRe were all managed by Capstone.

1. Stop-Loss and Quota-Share Agreements

The first arrangement involved stop-loss coverage provided by PoolRe to Peak. Through endorsements on all of Reserve's direct policies with Peak, PoolRe received 18.5% of the premiums paid by Peak on each policy in 2008 and 2009, and 19.9% of the premiums in 2010. See Figure 3 (showing the 2008 and 2009 arrangements). In return, PoolRe assumed (through an endorsement on the policy) some of Reserve's exposure on each policy. A financial backstop for Reserve was ostensibly important because it could face liabilities exceeding $8 million per year on the policies issued to Peak; at least for the first few years of its operation, such liabilities could not be covered by the combination of Reserve's $100,000 initial capitalization and the additional capital acquired from annual premium payments. But PoolRe did not provide much of a backstop. The amount that PoolRe could be required to pay each year on the stop-loss coverage was strictly limited, and the requirements for it to make any payment at all were intricate, restrictive, and highly unlikely to materialize.

We describe in some detail the limits with respect to the 2009 policies to illustrate the general issues. First, PoolRe incurred no liability under the stop-loss coverage until Reserve received claims from Peak in excess of the total premiums paid by Peak (18.5% of which went to PoolRe). Thus Peak needed to submit claims to Reserve in excess of $448,127.03 to trigger the stop-loss coverage. If that condition was satisfied, PoolRe would then pay amounts exceeding the total premiums, but only up to a maximum of 150% of those premiums. PoolRe could therefore not be required to pay more than about $672,000 in stop-loss coverage on the 2009 policies, even though the liability of Reserve to Peak could have been as high as $8 million on those policies.

The amount that PoolRe would have to pay on Reserve's policies was also restricted by an additional requirement that had little to do with Reserve. PoolRe's total liability on all its stop-loss coverage for the captive insurers was limited to 125% of the total yearly premiums paid for stop-loss coverage by all the captive insurers. In 2009 PoolRe received just under $83,000 in premiums on the Reserve policies; from all the captive insurers together, it received about $6 million. If a number of the other captive insurers suffered large losses in the same year, the amount that Reserve could recoup from PoolRe for that year could be further reduced.

Another feature of the PoolRe coverage, however, made the above limitations more theoretical than real. The limitations were unlikely to ever come into play because PoolRe would incur liability only if what the policy called an "Attachment Point" were satisfied. Aplt. App., Vol. 12 at 3411. There were four possible attachment points, each requiring at least two significant losses. Thus, a single catastrophic loss on one of the Reserve policies would not trigger any payments by PoolRe. The first attachment point was that Peak suffer at least two losses where (1) each loss was caused by a separate event and (2) each loss cost Peak at least $100,000. A second attachment point would be triggered only when Peak experienced three losses, each caused by a separate event, that caused the loss of at least $60,000. The third required four separate losses of $36,000 or more; and the fourth required five separate losses of $20,000 or more.9 Even then, PoolRe would pay no more than the amount that exceeded the attachment point on the claim where the attachment point was first satisfied (and for all later claims). For example, if the first claim was for $1 million, and the second was for $125,000, PoolRe would pay only $25,000, the amount by which the second claim exceeded the $100,000 threshold. The other captive insurers in the program had similar arrangements with PoolRe, which made no payments under the stop-loss coverages to any of the 50-odd captive insurers (on their 400-plus policies) for at least the period at issue in this case.

Because it was so unlikely that any money would ever be paid under the stop-loss coverage provided by PoolRe, one might infer that it was of very little financial utility to Peak or Reserve. But PoolRe was an essential part of the mechanism that created the appearance that Reserve had substantially diversified its risks beyond those faced by its affiliate Peak. What created that appearance was Reserve's reinsurance of liabilities of PoolRe. Not only did PoolRe provide stop-loss coverage for the insureds of Reserve and the 50 other captive insurers, but Reserve and the other captives in turn reinsured PoolRe's stop-loss coverage for all the captives' insureds (such as Peak), thereby assuming all the liability of PoolRe for the stop-loss coverage. In other words, PoolRe was essentially only an intermediary. The liability for paying for stop-loss claims incurred by a captive insurer ultimately rested on all the captive insurers as a group. This risk-pooling reinsurance arrangement, which was called a quota-share agreement, gave the appearance that each captive insurer (such as Reserve) was spreading its risk beyond its affiliated companies (such as Peak) by incurring liability on the stop-loss coverage provided for the benefit of the other captive insurers.

For participating in the reinsurance pool, each captive received from PoolRe a premium equal to the amount that PoolRe itself had received from that captive's insureds for the stop-loss coverage, so that, in effect, the captive insurer (such as Reserve) ultimately received all the premiums paid by its insured (such as Peak). For example, in 2009 Peak paid 18.5% of its premiums (about $83,000) directly to PoolRe for the stop-loss coverage, but PoolRe then paid Reserve that same amount ($83,000) for its reinsurance of stop-loss coverage, so Reserve ended up receiving 100% of what Peak paid for insurance. See Figure 3 above. In return for receiving this premium from PoolRe, the captive insurer was liable for a corresponding percentage of the liability incurred on the reinsurance provided to PoolRe by the captive insurers as a group. If the captive insurer received 3% of the total premiums paid by PoolRe for reinsurance, it was liable for 3% of what PoolRe had to pay on its stop-loss coverage.

The nature of the arrangement can be illustrated by a simplified example. Say, the PoolRe risk pool had three captives, A, B, and C, which respectively reinsure 50%, 30%, and 20% of the total stop-loss risk, because A's insureds pay 50% of the premiums to PoolRe for the stop-loss coverage, B's pay 30%, and C's pay 20%. If Captive B's insured has a covered claim that requires PoolRe to pay it $100, then Captive A must pay PoolRe $50, Captive B must pay $30, and Captive C must pay $20. B ends up reducing its potential $100 loss to $30, while the other pool members are out $50 and $20. No matter which captive insurer incurs the loss, the loss is ultimately borne in the same 5:3:2 proportions by the three captives.

In the actual PoolRe arrangement for 2009, the percentage of the total reinsurance paid by a captive ranged from 0.0984% (for a captive whose insureds paid $5,550 in stop-loss premiums) to 3.3235% (for a captive whose insureds paid $187,434 in premiums). Reserve's percentage was about 1.47%, based on Peak's paying PoolRe $82,904 in premiums for the stop-loss coverage on Peak's policies with Reserve. PoolRe then paid Reserve the identical sum as the premium for reinsurance.

Although Reserve ultimately received premiums precisely matching the combined total that Peak paid to Reserve and PoolRe, the reinsurance arrangement gave the appearance that Reserve was receiving only 81.5% of the premiums paid by Peak and a substantial premium paid by PoolRe (equal to 18.5% of the premiums paid by Peak) for reinsurance of stop-loss coverage for about 400 businesses (the insureds of the other captives) who were not affiliated with Reserve. (Recall that Capstone believed that for Reserve to be qualified as an insurer under the tax laws, Reserve needed to receive at least 30% of its premiums from its insurance of unaffiliated insureds, and the quota-share arrangement purportedly provided about half of that amount.) We recognize that a legitimate pool with a similar arrangement might properly distribute risk among insurers, even captive insurers. Here, however, the question was whether the stop-loss coverage was a legitimate risk. As previously noted, no claims under that coverage were made during the years in question by any of the 400 or so companies insured by the captive insurers participating in the PoolRe pool.

In addition, even assuming that the stop-loss coverage provided by PoolRe was not illusory, the only support in the record for the pricing of this coverage for the 2008 and 2009 policies is a 2005 letter signed by Robert Snyder, the director of risk consulting for the firm Myron Steves, who later served as one of the authors of Reserve's 2009 feasibility study, and eventually as a director of PoolRe itself. The letter was in response to a request from PoolRe to comment on the reasonableness of its Stop-Loss/Reinsurance premium structure, under which PoolRe would receive 18.5% of the premium paid by the insured. The letter concluded that "both the quota share premium retained by PoolRe and the quota share reinsurance premium(s) ceded to the captives via the pooling mechanism are reasonable." Aplt. App., Vol. 12 at 3568. The letter explains, however, that its "analysis and observations are limited to general commentary regarding stop loss insurance and the reinsurance pooling concept, and specifically, our assessment of the proposed premium structure." Id. at 3567. It does not indicate that Snyder had undertaken any specific assessment of the risks arising from the underlying insurance policies of the PoolRe captives or examined how any particular provisions of the stop-loss coverage, such as the highly restrictive attachment points, affected the insurer's risks. It noted that a review of the premiums was "necessarily subjective," id., and stated that the premiums employed by PoolRe are reasonable, "recognizing that in our judgment stop loss coverage in general might fairly be priced at anywhere from 2.5% to 30% of written premium." Id. at 3568. Talk about going out on a limb.

There is no evidence Capstone performed any further risk analysis before settling on 18.5% as the price paid by all captive insureds in 2008 and 2009.10 The premium for stop-loss coverage should depend on both the likelihood of losses covered by the Reserve policy and the likelihood that the losses during the year would satisfy the attachment-point and other conditions of the stop-loss coverage, which depends both on the amount of the loss and the frequency of losses. Not only does nothing in the record provide reason to believe that 18.5% of the total premium is the appropriate premium for the stop-loss coverage for any particular captive insured, but Reserve has also failed to explain why it should be the same percentage for every captive insured. Reserve has provided no reason to believe, for instance, that for every (properly priced) insurance policy with a $1 million coverage limit, the probability of incurring two losses exceeding $100,000 in the same year is identical. And there is no indication in the record that the owners of Reserve or anyone else on its behalf exercised any diligence or investigated the risks it was assuming from participation in the PoolRe pool. One would think that pool members would want to make sure that the premium prices, and therefore their quota-share percentages, accurately reflected the risks involved because if one of the captives' stop-loss endorsements triggered a disproportionate number of claims, then that captive would receive a greater benefit than the rest of the pool members.11

In the absence of due diligence in assessing the risks of the other captives in the PoolRe pool, the willingness of the captives, including Reserve, to join the pool can perhaps be explained by the unusual, if not unique, fact that both PoolRe and all the captives were managed by Capstone. (One of Reserve's experts testified that he had never seen a risk pool where all the participants and the pool were managed by the same entity.) Zumbaum, for his part, testified that he completely relied on Capstone. Also, the captives may have determined, or been advised by consultants, that they would almost surely incur no liability under the stop-loss reinsurance policies, so the premiums for that coverage (which would come back to the captives via the reinsurance of PoolRe) was of no import.

2. Credit-Coinsurance Arrangement

The other way in which Reserve purportedly distributed risk was through a coinsurance arrangement between PoolRe and an insurer called CreditRe, which ostensibly accounted for about 15% of Reserve's premiums receipts. CreditRe was not a Capstone-managed entity; rather, it was owned and operated by a man named Gary Fagg. Under the agreement between PoolRe and CreditRe, CreditRe ceded to PoolRe a portion of the risk that CreditRe assumed from vehicle-service contracts between consumers and another insurer, Lyndon Property Insurance Co. PoolRe then executed a separate agreement with Reserve — called the "Credit Insurance Coinsurance Contract" — under which Reserve reinsured about 1% of PoolRe's liability from the coinsurance contract and, in return, PoolRe would pay a premium to Reserve. Aplt. App., Vol. 11 at 3261.

The bona fides of the arrangement, however, could be questioned. Very little money changed hands between Reserve and PoolRe. The financial records do not show any premium payment actually transmitted to Reserve during the years in question. Nor is there any record of particular claims paid by PoolRe or Reserve. The sole payments in the relationship appear to be annual deposits of less than $200 each from PoolRe to Reserve purportedly equal to the amount by which the annual premiums exceeded the annual reinsurance obligations incurred by Reserve. Reserve's ledgers reflect actually receiving only three payments totaling $530 across three years under this agreement.

Further, there is nothing in the record showing what liability Reserve was allegedly insuring. Reserve did not submit any underlying vehicle-service contracts. And when government counsel asked Fagg to "elaborate on what those claims [the vehicle-service claims] would generally entail," he replied: "I'm familiar with them in the sense of I know them in terms of dollars. But I was not directly involved in the adjudication of those claims or any of that." Aplt. App., Vol. 5 at 1285.

Nor was there any evidence of how the premiums for the credit-coinsurance agreements were calculated. The agreement stated merely that the premium amounts would correspond to the share of risk that PoolRe assumed. Fagg testified that "in discussion with PoolRe," he "basically asked them how much [risk] they would be willing to accept," and then ceded that amount. Id.

D. Tax Dispute

On its income-tax returns Peak deducted the premiums it paid Reserve. The Tax Code generally treats a business's payment of insurance premiums as a tax deductible "ordinary and necessary expense[ ]" of doing business. I.R.C. § 162(a); see 4 Couch on Insurance, supra, § 63.5 (internal quotation marks omitted). But there was no corresponding taxable income reported by Reserve. Reserve paid no income tax on the premiums it received, claiming that it qualified under § 501(c)(15) as a tax-exempt insurance company receiving premiums under $600,000 per year.

The IRS sent Reserve a Notice of Deficiency challenging this claim, disputing that Reserve was tax exempt under § 501(c)(15). The notice said that Reserve's "purported insurance and/or reinsurance transactions lack[ed] economic substance," and that the money it had received was "not paid to an insurance company and . . . [was] not paid for insurance." Aplt. App., Vol. 6 at 1592. As a result, the IRS assessed Reserve for taxes for 2008, 2009, and 2010. It classified Reserve's income as FDAP income, which is "[F]ixed or [D]eterminable [A]nnual or [P]eriodical gains, profits, and income" that is "received from sources within the United States," but "not effectively connected with the conduct of a trade or business within the United States." I.R.C. § 881(a). FDAP income is subject to a 30% tax rate. See id. Reserve contested the assessment, arguing that it was in fact an insurance company and that if it was not, the purported premiums it received were a nontaxable capital contribution to the company. The case was tried to the Tax Court, which affirmed the IRS. Reserve appeals.

III. APPLICABLE LAW AND THE TAX COURT'S DECISION THAT RESERVE WAS NOT AN INSURANCE COMPANY

Under § 501(c)(15) of the Tax Code, "[i]nsurance companies (as defined in section 816(a)) other than life" are tax exempt if "(I) the gross receipts for the taxable year do not exceed $600,000, and (II) more than 50 percent of such gross receipts consist of premiums." I.R.C. § 501(c)(15)(A)(i). The core question before the Tax Court was whether Reserve was an insurance company. I.R.C. § 816(a) defines insurance company as "any company more than half of the business of which during the taxable year is the issuing of insurance or annuity contracts or the reinsuring of risks underwritten by insurance companies."

But no part of the Code defines the term insurance. Unfortunately, this is not because the definition is obvious. Indeed, the meaning can depend on the context. See Robert E. Keeton & Alan I. Widiss, Insurance Law: A Guide to Fundamental Principles, Legal Doctrines and Commercial Practices § 1.1, at 5 (1988) ("There is no single conception of insurance that is universally applicable for use in disputes involving questions of law."). At the core of the notion of insurance, however, are risk transfer and distribution. The insured transfers its risk of loss to the insurance company, and the insurer distributes that risk by charging premiums to many insureds. See id. at 4 ("[R]isk transference and risk distribution are among the basic characteristics of almost all insurance transactions."); Grp. Life & Health Ins. Co. v. Royal Drug Co., 440 U.S. 205, 211 (1979) ("The primary elements of an insurance contract are the spreading and underwriting of a policyholder's risk."); Beech Aircraft Corp. v. United States, 797 F.2d 920, 922 (10th Cir. 1986) ("'Risk-shifting' means one party shifts [its] risk of loss to another, and 'risk-distributing' means that the party assuming the risk distributes [its] potential liability, in part, among others. An arrangement without the elements of risk-shifting and risk-distributing lacks the fundamentals inherent in a true contract of insurance."); Stearns-Roger Corp. v. United States, 774 F.2d 414, 415 (10th Cir. 1985) ("[F]or there to be 'insurance' there must be a shifting of the risk of loss or a spreading of the risk.").

The Supreme Court decision in Helvering v. Le Gierse, 312 U.S. 531 (1941), illustrates that a transaction must display the basic attributes of insurance and that substance trumps form. A woman of 80 purchased a life-insurance policy and an annuity at the same time from the same life insurance company. The two contracts were formally treated as distinct transactions, but neither would have been purchased without the other. She paid $4,179 for an annuity that provided an annual payment of $589.80 throughout her life, and she paid $22,946 for a life-insurance contract that would pay $25,000 upon her demise. To obtain the policy she did not have to submit to a physical examination or answer the typical questions. She died a month later. The government claimed that the $25,000 payment under the contract could not be excluded from her taxable estate as life-insurance proceeds. The Court agreed. It observed that "Congress used the word 'insurance' in its commonly accepted sense," and that "[h]istorically and commonly insurance involves risk-shifting and risk-distributing." Id. at 539-40. In those respects the decedent had not acquired insurance. The Court said that the desirability of life insurance stems from its being "a device to shift and distribute risk of loss from premature death," id. at 539, but the combination of the decedent's two contracts "fail[ed] to spell out any element of insurance risk," id. at 541. It noted that "annuity and insurance are opposites" — "[f]rom the company's viewpoint, insurance looks to longevity, annuity to transiency." Id. Through the combination of the two contracts, said the Court, "the one neutralizes the risk customarily inherent in the other." Id. It concluded that the risk assumed by the insurance company "was an investment risk similar to the risk assumed by a bank; it was not an insurance risk." Id. at 542.

We cannot anticipate every nuance that may arise and provide a comprehensive explication of what is required for a purported insurer to satisfy § 816(a), but this court has previously followed the approach of Le Gierse. See, e.g., Stearns-Roger Corp., 774 F.2d at 415. Taking our lead then from that Supreme Court decision, we think that to be a "company more than half of the business of which . . . is the issuance of insurance . . . contracts," I.R.C. § 816(a), the company must deal in the transference and distribution of risk and it must do so as a "business," conducting itself through practices that one would expect from a reasonably ordered insurance enterprise.

In our view, the Tax Court has captured the essence of the statutory requirement in the tests that it has employed to determine whether a taxpayer is an insurance company under § 816(a). In this case it stated that "[c]ourts have looked to four criteria in deciding whether an arrangement constitutes insurance: (1) the arrangement involves insurable risks; (2) the arrangement shifts the risk of loss to the insurer; (3) the insurer distributes the risk among its policy holders; and (4) the arrangement is insurance in the commonly accepted sense." Reserve, at 33 (citing Harper Grp. v. Comm'r, 96 T.C. 45, 57-58 (1991), aff'd, 979 F.2d 1341 (9th Cir. 1992), and AMERCO & Subs v. Comm'r, 96 TC 18, 38 (1991), aff'd 979 F.2d 162 (9th Cir. 1992)); see also Caylor Land & Dev., Inc. v. Comm'r, 121 T.C.M. (CCH) 1205 (2021) at *31-32; Avrahami v. Comm'r, 149 T.C. 144, 177 (2017). Neither party objects to this framework. The Tax Court held that Reserve's policies failed to satisfy requirements (3) and (4). We begin with risk distribution.

A. Risk Distribution

Risk distribution is "essential" to insurance. Le Gierse, 312 U.S. at 539; see Comm'r v. Treganowan, 183 F.2d 288, 291 (2d Cir. 1950) (describing risk distribution as "the very essence of insurance" (internal quotation marks omitted)); Keeton & Widiss, supra, § 1.3, at 12 ("When insurance is considered from the viewpoint of an insurer or of society, it is appropriately viewed as a system for distribution[.]"). "Insuring many independent risks," that is, independent in the sense that the likelihood of a loss under one policy is independent of the likelihood of a loss under a separate policy, "in return for numerous premiums serves to distribute risk." Clougherty Packing Co. v. Comm'r, 811 F.2d 1297, 1300 (9th Cir. 1987); see Beech Aircraft Corp., 797 F.2d at 922. The success of insurance derives from the "law of large numbers." Clougherty Packing Co., 811 F.2d at 1300. When, as is assumed when writing insurance policies, the risk of an adverse event to each insured is random, with a chance of, say, X% per year, there is always the possibility that fate will strike a cruel blow and a number of adverse events will occur at the same time or in quick succession. The insurer may have to pay out much more in claims than it anticipated when it set the premiums on the policies, perhaps jeopardizing its financial solvency. The law of large numbers is no more than an expression of a simple computation using the mathematics of probability — when there are a sufficiently large number of independent risks each having an annual probability of occurrence of X%, there is an extraordinarily small likelihood that the percentage of insureds that suffer a loss during a year will deviate significantly from X%. (If a coin is tossed a million times, it is highly unlikely that the percentage of heads will differ appreciably from 50%.) Thus, insuring a large number of independent risks protects the insurer against financial calamity. See Keeton & Widiss, supra, § 1.3 at 12-13 ("As the number of ventures included in a group or 'pool' is increased, there is a greater likelihood that the favorable and the harmful experiences will tend to be balanced — that is, grouping a large number of ventures in a pool increases the probability that the losses suffered by all the ventures will be spread over time."). By writing policies for enough independent risks, an insurer can predict "the frequency and amount of loss within this larger group . . . [and] set the premium at a level that will cover the losses, cover the insurer's overhead and expenses, and earn a profit." 1 Daniel W. Gerber et al., New Appleman on Insurance Law Library Edition § 1.01 (2021).

The Tax Court began by finding that Reserve's direct written policies for Peak, RocQuest, and ZW did not themselves distribute risk. It said that because "most or all of the risk of loss was associated with the business operations of just one insured [Peak]," "the number of insureds and the total number of independent exposures were too few to distribute the risk that Reserve assumed under [those] policies." Reserve, at 36. Reserve does not challenge this determination.

The court therefore turned to Reserve's contention that it distributed risk through the arrangements with PoolRe. It noted that Reserve relied on Harper, a Tax Court case holding that a particular captive insurer had "'a sufficient pool of insureds to provide risk distribution'" when "approximately 30% of the captive's business came from insuring unrelated parties." Id. at 38 (quoting Harper Grp., 96 T.C. at 60). It pointed out that when it had previously held that risk distribution had been achieved through insuring unrelated parties, it had "determined that the transactions with the unrelated parties were insurance transactions for Federal income tax purposes." Id. Therefore, the court said, "[b]efore we can determine whether Reserve effectively distributed risk through these agreements, we must determine whether PoolRe was a bona fide insurance company." Id. It set forth nine factors that the Tax Court had previously considered in making such a determination:

(1) whether [the purported insurance company] was created for legitimate nontax reasons;

(2) whether there was a circular flow of funds;

(3) whether the entity faced actual and insurable risk;

(4) whether the policies were arm's-length contracts;

(5) whether the entity charged actuarially determined premiums;

(6) whether comparable coverage was more expensive or even available;

(7) whether it was subject to regulatory control and met minimum statutory requirements;

(8) whether it was adequately capitalized; and

(9) whether it paid claims from a separately maintained account.

Id. at 38-39 (quoting Avrahami, 149 T.C. at 185, with numbering added).

1. Quota-Share Arrangement

Beginning with the quota-share arrangement, the Tax Court addressed only the six factors that it found to be "the most relevant." Id. at 39.12 First, it determined that the "arrangement looks suspiciously like a circular flow of funds." Id. at 41 (internal quotation marks omitted). It said that because Reserve never "had any losses or expenses in connection with its purported quota share liabilities[,] . . . the end result for each tax year . . . was that Reserve would receive payments from PoolRe in exactly the same amount as the payments that PoolRe was entitled to receive from Peak . . . for the stop loss coverage." Id.

Second, the Tax Court concluded "that the amounts that PoolRe was to pay Reserve under the quota share arrangement were not determined at arm's length or using objective criteria." Id. at 42. It explained:

The perfect matching of payments under the corresponding stop-loss endorsements and quota share policies (from insureds to PoolRe, and from PoolRe to captives) indicates that the quota share arrangement was not the product of arm's-length considerations. Peak's risks that were insured through PoolRe were different from the risks that PoolRe ceded to Reserve under the quota share policies. The risks that PoolRe purported to assume under the stop loss endorsements related to various unrelated business activities and the policies covering various unrelated lines of insurance. Reserve has not shown that the risks were comparable in scale.

Id. at 41. The court noted that Reserve had not produced "evidence which shows the risks of other Capstone entities, . . . [such as] evidence regarding their industries, locations, operations, types of risk, and exposure to risk." Id. at 42.

Third, for similar reasons, the Tax Court concluded that the quota-share premiums were not actuarially determined. Not only was there "no evidence to support the calculation of the premiums," but it was concerned that "all participants in the quota share arrangement agreed to direct their affiliated insureds to pay the same percentage of direct written premiums to PoolRe." Id. at 42-43. Such a "one-size-fits-all" approach appeared inconsistent with actual actuarial determination. Id. at 43.

As for whether the quota-share policies insured an actual and insurable risk, the Tax Court stated that "PoolRe was removed far from any actual risk associated with the business or operations of Reserve's insureds." Id. at 44. Before obtaining insurance from Reserve, Peak had "never suffered any losses that would even come close to triggering the stop loss coverage provided for in the stop loss endorsements." Id.

The fifth factor was whether PoolRe was licensed and regulated as an insurance company. The Tax Court observed that "Reserve executed both its 2008 and 2009 reinsurance agreements with PoolRe before it obtained an insurance license in Anguilla," and there was no evidence that PoolRe was a licensed insurer before then. Id. at 45.

Sixth, the Tax Court found that "Reserve has not established that PoolRe was created or operated for legitimate nontax reasons." Id. It said that "[a]ll the facts and circumstances in this case indicate that Reserve did not enter into the quota share arrangement with the intention of distributing its risk." Id. Instead, "[t]he only purpose PoolRe served through the quota share arrangement was to shift income from Peak to Reserve." Id.

The Tax Court concluded that "the facts surrounding Reserve's quota share policies with PoolRe establish that those agreements were not bona fide insurance agreements." Id. "PoolRe's activities as they relate to [the Reserve] policies," it said, "were not those of a bona fide insurance company." Id. at 46.

One aspect of the Tax Court's analysis must be emphasized. Contrary to what is suggested in some briefs filed in this court, the Tax Court did not criticize risk pools as a general matter. Instead, its findings were specific to the PoolRe risk pool.

2. Credit-Coinsurance Arrangement

Turning to the credit-coinsurance contract, the Tax Court summarily rejected Reserve's contention that it provided risk distribution. It said that Reserve had "failed to provide evidence that the vehicle service contracts, which formed the basis for the reinsurance that PoolRe re-ceded in the coinsurance contracts, actually existed." Id. at 46-47. And it found that even if there were valid coinsurance contracts, PoolRe assumed a "de minimis" amount of risk from CreditRe and the amount ceded to Reserve was also de minimis. Id. at 47. Thus, "the coinsurance contracts were not bona fide reinsurance agreements." Id.

B. Insurance in the Commonly Accepted Sense

As an alternative basis for rejecting Reserve's appeal, the Tax Court held that Reserve's policies "did not constitute insurance in the commonly accepted sense." Id. at 48. The Tax Court considered the following factors: "[1] whether the company was organized, operated, and regulated as an insurance company; [2] whether it was adequately capitalized; [3] whether the policies were valid and binding; [4] whether the premiums were reasonable and the result of an arm's-length transaction; and [5] whether claims were paid." Id. at 48. These are all factors that the Tax Court had previously considered over the years. See, e.g., Avrahami, 149 T.C. at 191; R.V.I. Guar. Co., Ltd. v. Comm'r, 145 T.C. 209, 231 (2015); Harper, 96 T.C. at 60; Securitas Holdings, Inc. v. Comm'r, 108 T.C.M. (CCH) 490 (T.C. 2014) at *10; see also Caylor, 121 T.C.M. (CCH) 1205 at *39-40. Reserve does not challenge that approach on appeal.

First, the Tax Court said that although Reserve had observed the necessary formalities in incorporating under the law of Anguilla and complying with the insurance regulations of that country, it had not been operated as an insurance company. The court observed that "[o]ther than the feasibility study that Capstone produced, there is no evidence that any due diligence was performed for the policies that Reserve issued." Reserve, at 50. And it noted that the feasibility study was not complete until after Reserve issued its policies for 2008 and 2009, that many of the background documents attached to the study covered periods after Reserve had been incorporated and issued policies, and that the study said nothing about RocQuest and ZW even though those entities were covered under the policies. Also, Reserve's operations (like its planning and incorporation) "were managed entirely by Capstone." Id. Reserve had no employees of its own and Zumbaum — who was Reserve's 50% owner, president, and chief executive officer — knew "virtually nothing about its operations," and at trial showed "very little knowledge of provisions in the policies that Peak and his other entities held with Reserve." Id. It further found that there was "no evidence that Reserve performed any due diligence with respect to the reinsurance agreements that it executed with PoolRe," that "[n]othing in the record indicates that Reserve or anyone performing activities on Reserve's behalf evaluated [the risks assumed by Reserve under the quota-share agreement] before executing the quota share policies," and that no one with a financial interest in Reserve considered whether the quota-share or credit-coinsurance agreements would actually distribute risk. Id. at 51-52.

The Tax Court also found that Reserve handled "in an irregular manner" its only claim — the one for loss of a major customer. Id. at 53. It said that "no supporting documentation accompanied the claim notice"; that Reserve never insisted on documentation of "the occurrence or the amount of the claimed loss"; that Reserve issued its first claim payment before Peak and Reserve executed their settlement agreement, and then issued another payment months after the settlement agreement without executing an addendum to the settlement agreement "reflecting this payment until years after the tax years in issue"; and that all the payments from Reserve to Peak were signed by a Peak employee. Id. at 52. The Tax Court concluded that "Reserve was not operated as an insurance company in the commonly accepted sense." Id. at 53. It saw no virtue in Capstone's management of Reserve, saying that "Capstone directed Reserve's activities and directed a series of transactions between its managed entities so that Reserve appeared to be engaged in the business of issuing insurance contracts." Id. (emphasis added).

As for the second factor — adequate capitalization — the Tax Court observed that "[g]enerally [its] caselaw holds that meeting the statutory requirements of the captive's domicile jurisdiction is sufficient to show that the captive was adequately capitalized." Id. And it concluded that because "Reserve met the minimum capitalization requirements of Anguilla," this factor favored Reserve. Id.13

The Tax Court next found that "Reserve's direct written policies contained the necessary terms to make them valid and binding insurance." Id. at 54. But it noted that the policies were "cookie cutter" — "[t]he policies on their face indicate[d] that they were the copyrighted material of Capstone, and Capstone employees testified at trial that they administered many of the same policies for all of their clients" — and "were not reasonably suited to the needs of the insureds, particularly Roc[Q]uest and ZW, both of which had extremely limited operations." Id. at 54 (internal quotation marks omitted). The Tax Court also reiterated that Reserve's only claim — the loss-of-a-major-customer claim — was paid in full without any "due diligence to determine whether the claim was actually covered by the relevant policy." Id. at 55. The court concluded that the third factor — whether the policies were valid and binding — was a "neutral factor." Id.

The fourth factor considered by the Tax Court was whether Reserve's premiums were reasonable and the result of arm's-length transactions. The court acknowledged that "Capstone calculated Reserve's premium using objective criteria and what appear to be actuarial methods," id. at 61, and that to make his recommendations for premium pricing, McNeel used a spreadsheet showing the premiums charged by other Capstone entities for various lines of coverage and "pricing indications" from Mid-Continent, id. at 55.

But the court said that "[d]espite [this] methodology . . . a number of factors indicate that the premiums that the insureds were required to pay under the direct written policies were not reasonable in relation to the risk of loss." Id. at 56. It observed that Peak's policies with Reserve caused a dramatic increase in Peak's insurance expenses: Peak's 2007 insurance expenses had been $95,828, but its Reserve premiums were more than "$412,089, and this was in addition to the premiums that Peak continued to pay for third-party commercial insurance." Id. And it noted that despite this immense increase in cost, Reserve's policies would be excess to any other policy that covered the same loss because of the other-insurance clauses in every Reserve policy.

The Tax Court also noted apparent irregularities in the pricing of the policies. For example, the 2008 regulatory-changes policy (discussed earlier in this opinion) provided coverage for occurrences during only one month and charged a $64,899 premium for $1 million in coverage, while the 2009 and 2010 versions of the policy charged a lower premium, $47,588, and offered $500,000 in coverage for a one-year coverage period. There was "no explanation" for this inconsistency. Id. at 57.

In addition, the Tax Court questioned Peak's assertion that it required unique pollution coverage from a captive insurer because one of its operations is located on a Superfund site. The court noted that "Peak itself did not engage in mining practices that spread pollutants, and it already had systems in place to control the fluid runoff when it cleaned equipment used in polluted mines." Id. Moreover, Peak had operated at the same location "continuously for over 10 years," and there was "no evidence that Peak had ever incurred costs during that time for excess pollution liability." Id.

Continuing this analysis, the Tax Court expressed doubts about Zumbaum's testimony on the need to create Reserve. Zumbaum testified that one reason Peak wanted to create a captive was because it was unhappy with how its current commercial-liability insurer handled a claim for roof damage. But "no documentation was produced to substantiate" this point, id., and Peak continued to keep its policies with the commercial-liability insurers. Also, Zumbaum "did not know which of Peak's policies with Reserve would have covered a loss like the one that was not covered by [its commercial insurer]." Id. at 58. And although Reserve contended that Peak risked substantial liability if one of its mining parts malfunctioned, "[t]here [was] no convincing evidence that legitimate concerns about this kind of liability should have been greater in 2008 than in previous years." Id. For example, the feasibility study "provided no information on the probability of a loss event that the direct written policies covered." Id. at 59. The court added that Reserve "failed to explain why Peak would maintain its full set of third-party commercial insurance coverage, which it contends was insufficient, even after it paid roughly 400% more for additional coverage from Reserve." Id. at 58.

The Tax Court also pointed out that Zumbaum's testimony about Peak's business projections was inconsistent with the documentary evidence. Zumbaum testified that Peak was interested in obtaining additional insurance coverage because it expected that its business would "continue growing during the next few years." Id. But "the rating worksheets that Capstone produced for calculating Reserve's premiums reflect that Peak's projected sales stayed the same for all of the tax years in issue," and as it turned out, "Peak actually had fewer employees in 2010 than it did in 2008." Id.

After "conclu[ding] that any purported concerns about increased risks for the insureds were unfounded," id. at 59, the court noted that in determining whether premiums charged by captive insurers were reasonable, it looked to whether "the amounts agreed upon by the parties were the result of arm's-length negotiations," id. at 60. It said: "In determining whether an arrangement constitutes insurance in the commonly accepted sense we consider more than whether the premiums chosen can be arrived at by actuarial means. We consider whether the facts demonstrate that the terms of the arrangement were driven by arm's-length considerations." Id. By that measure, Reserve failed to make its case. The Tax Court found that "the facts and circumstances of this case demonstrate that the direct written policies were not the result of arm's-length negotiations." Id. at 61. In its view, "no unrelated party would reasonably agree to pay Reserve the premiums that Peak and the other insureds did for the coverage provided by the direct written policies." Id. It summarized: "Although Capstone calculated Reserve's premiums using objective criteria and what appear to be actuarial methods, the absence of a real business purpose for Reserve's policies leads us to conclude that the premiums paid for the policies were not reasonable and not negotiated at arm's length." Id.

Turning to the fifth factor — the payment of claims — the Tax Court observed that Reserve did pay the single claim for loss of a major customer. And although it noted that "the circumstances surrounding the payment of [the loss-of-a-major-customer] claim were unusual," it ultimately determined that this factor "weigh[ed] slightly in Reserve's favor." Id.

The Tax Court said that on balance the evidence favored a conclusion "that Reserve's transactions were not insurance transactions in the commonly accepted sense." Id. at 62. It placed particular emphasis on (1) Reserve's not being operated as a bona fide insurance company, (2) the lack of a legitimate business purpose for the policies, and (3) the large increase in payments for new insurance while still maintaining prior coverages. See id.

IV. THE TAX COURT DID NOT ERR IN RULING THAT RESERVE WAS NOT AN INSURANCE COMPANY

"We review decisions of the Tax Court in the same manner as civil actions tried without a jury." Hamilton v. Comm'r, 955 F.3d 1169, 1171 (10th Cir. 2020). That is, we "review legal conclusions de novo and factual determinations only for clear error." Id. at 1172 (emphasis omitted). And we review the evidence in the light most favorable to the Tax Court's ruling. See Exxon Corp. v. Gann, 21 F.3d 1002, 1005 (10th Cir. 1994) ("On appeal, we view the evidence in the light most favorable to the district court's ruling and must uphold any district court finding that is permissible in light of the evidence." (citation omitted)).

The essence of the Tax Court's ruling was that Reserve's insurance policies issued to Peak did not provide true insurance even though Reserve complied with some (but by no means all) of the customary formalities for insurance companies and went through some of the motions associated with pricing insurance premiums.

Perhaps we would have emphasized different evidence than did the Tax Court. But that court's findings and conclusions were supported by overwhelming evidence in the record. No experience, expertise, or studies supported the need for Peak to obtain the policies. And the premiums for that additional insurance were not supported by any study of similar commercially available policies or careful analysis of Peak's risks of loss. Nor has Reserve raised any persuasive challenges to the Tax Court's conclusions. We now address those challenges in turn.

A. Risk Distribution — the Quota-Share Arrangement

To qualify as an insurance company, Reserve needed to distribute risk. Reserve acknowledges that issuing a dozen-odd policies to essentially one insured (Peak) did not fill the bill. Rather, it relies on the reinsurance arrangements it had with PoolRe to establish risk distribution. The Tax Court rejected that reliance. Reserve challenges that rejection, arguing that the court improperly held that risk could not be distributed via PoolRe simply because PoolRe was not a bona fide insurer.

Reserve does not challenge the Tax Court's conclusion that PoolRe was not a bona fide insurer with respect to the quota-share agreement. Instead, it argues that "the tax court misapplied the legal test for analyzing risk distribution" by considering whether PoolRe — rather than Reserve — was a bona fide insurance company. Aplt. Br. at 37. It contends that "the existence of a bona fide insurance company is not necessary for risk distribution to exist." Id. at 38.

As an abstract legal proposition, Reserve's point has some merit. The courts have recognized, and the Commissioner concedes, that certain arrangements distributing risk provide insurance even though there is no insurance company involved. See Ross v. Odom, 401 F.2d 464, 466-67 (5th Cir. 1968) (Although "[t]he amounts payable to beneficiaries under the system were not funded or reinsured by any independent insurance company [,] . . . we fully approve the Trial Court's holding that the [payment] constituted amounts received under a life insurance contract."); Treganowan, 183 F.2d at 290-91 (plan under which all members of an organization pledged to pay a set sum to a deceased member's beneficiary constituted insurance). Reserve argues that an insurance company could reinsure such risks without jeopardizing its status as an insurance company. Perhaps.14 But that proposition does not help Reserve here.

To begin with, the Tax Court's conclusion that PoolRe was not a bona fide insurance company was not based on the absence of formalities. True, one of the six factors discussed by the court was whether PoolRe was licensed and regulated as an insurance company; and the court noted that there was no evidence that it was so licensed when the first two sets of policies were issued to Peak (in December 2008 and January 2009). But that factor apparently had little or no effect on the court's conclusion, since it reached the same conclusion about PoolRe for the policies issued in 2010 as well, and the "Conclusion" section of this part of the court's opinion makes no mention of the licensing issue when it summarizes why it finds that "PoolRe's activities as they relate to [the quota-share] policies were not those of a bona fide insurance company." Reserve, at 46.

Elsewhere in that Conclusion the Tax Court used language that fairly encompasses a determination that the quota-share arrangement did not distribute risk for Reserve. The heart of the problem was not that PoolRe was not an insurance company but that its product was not actual insurance. The court concluded that "the facts surrounding Reserve's quota share policies with PoolRe establish that those agreements were not bona fide insurance agreements." Id. at 45. We agree with the Commissioner's characterization of the Tax Court's analysis: "[T]he Tax Court did not invalidate the quota share arrangement on the ground that PoolRe failed to meet the formal definition of an insurance company. Rather, it invalidated the quota share arrangement on the ground that, as a matter of substance, PoolRe did not perform the functions of an insurance company — regardless of label — vis-à-vis the quota share arrangement." Aplee. Br. at 40. In short, what the Tax Court determined is what Reserve contends should have been determined — whether the quota-share arrangement was a true insurance arrangement for the distribution of risk.

The Tax Court did not err when it assessed PoolRe's bona fides. Reserve's opening brief does not point to, and we do not see, any significant flaw in the analysis by the Tax Court in reaching its conclusions that "Reserve's quota share policies with PoolRe . . . were not bona fide insurance agreements" and that "Reserve did not enter into the quota share arrangement with the intention of distributing its risk." Reserve, at 45. On the contrary, the Tax Court opinion, if anything, understates the compelling evidence that the quota-share arrangement was a sham.

Because Reserve's reinsurance of PoolRe's quota-share arrangement did not create any meaningful risk for Reserve, Reserve did not satisfy even the distribution threshold that Capstone set for it — obtaining 30% of its insurance premiums by insuring unaffiliated risks. Reserve has not argued that it could reach the 30% figure based only on the reinsurance of the credit-coinsurance agreements. The Tax Court's ruling must therefore stand. But we briefly discuss the credit-coinsurance arrangement because the Tax Court's ruling on that arrangement is an alternative basis for affirmance.

B. Risk Distribution — The Credit-Coinsurance Arrangement

The Tax Court determined that the credit-coinsurance agreements were not bona fide reinsurance agreements for two reasons: (1) there was no evidence that the underlying risk existed because Reserve never produced the vehicle-service contracts and (2) even if the agreements insured an actual risk, both the amount of risk PoolRe assumed and the amount ceded to Reserve were too "de minimis" to constitute "actual risk." Id. at 47.

Reserve challenges the first of the Tax Court's reasons, contending that the court should not have faulted it for failing to produce the vehicle-service contracts. It argues that because the reinsurance arrangements were what are called "treaty" arrangements (under which the reinsurer is "bound to automatically accept all the policies and losses covered thereby"), Reserve "would not have examined risks, investigated claims, or even received notices of losses from the original insureds." Aplt. Br. at 44. Therefore, it says, the Tax Court's criticism for not providing documentary evidence of the vehicle-service contracts is "misplaced and ignores not only the evidence in the record but also the rationale behind treaty reinsurance. Indeed, if a reinsurer were required to duplicate the costly but necessary efforts of a primary insurer in evaluating risks and handling claims, reinsurance simply would not work because it would not be economical to place and administer." Id. at 44-45 (citations omitted). This argument ignores, however, (1) the need for the information sought by the court in checking the bona fides of the "reinsurance" arrangement, (2) the documents that apparently gave PoolRe the right of access to the information sought, and (3) the need for a prospective treaty insurer to examine such information in some fashion when deciding whether or not to enter into a treaty-insurance relationship in the first place. But we need not resolve the merits of Reserve's arguments on this point, because we must affirm on the other ground (de minimis risk) for the Tax Court's decision.

There is certainly substantial evidence suggesting that Reserve incurred very little risk from the agreements. Over the course of 2008-10, Reserve received no payment of annual premiums and made no payment on claims. It instead received only one deposit each year, purportedly equal to the net of premiums less claims for that year, for a total of $530 over three years. Perhaps there are factual or legal flaws in the Tax Court's analysis of the extent of risk assumed by Reserve through the credit-coinsurance arrangement, but Reserve's opening brief raised no factual or legal challenge to the analysis. It completely failed to address the court's determination that the risk from the vehicle-service contracts did not constitute actual risk. The issue is therefore waived, and we must affirm the Tax Court's determination that the credit-coinsurance arrangements did not distribute risk. See Lebahn v. Nat'l Farmers Union Unif. Pension Plan, 828 F.3d 1180, 1188 (10th Cir. 2016) ("When a district court dismisses a claim on two or more independent grounds, the appellant must challenge each of those grounds.").

C. Insurance in the Commonly Accepted Sense

As a third alternative ground for rejecting Reserve's appeal, the Tax Court ruled that the Reserve policies did not satisfy the requirement that they be insurance in the commonly accepted sense. We also affirm this ruling.

We have already discussed at some length the evidence supporting the Tax Court's determination that Reserve did not act like a company in the business of writing legitimate insurance policies. Its only true insured was Peak. The arrangements with PoolRe lacked substance. There was no apparent business reason for the credit-coinsurance contracts. And the Tax Court was fully justified in determining that the quota-share reinsurance arrangement was a sham. The attachment points that had to be met before PoolRe had any liability to the Capstone captive insureds defied explanation. Reserve provided no evidence to overcome the logical inference that the reason to set such intricate limitations on liability was to ensure that PoolRe would never have to make a payment (particularly when the risk of loss under any of the underlying policies issued by the captive insurers was apparently low to start with), so there was really nothing for Reserve or the other captive insurers to reinsure.

Moreover, there was no evidence of any reasonable risk assessments to determine whether Peak needed any of the additional policies. Nor was there any evidence of due diligence to determine whether the premiums charged on the policies were worth it to Peak. And Reserve was hardly run like a business. The preparation of the original one-month policies in an apparent rush to obtain a large business deduction for Peak in 2008 was laughable. It suffices to note that on two of the policies the listed insureds, rather than Peak, RocQuest, and ZW, were "Pacific Arts Entertainment, LLC" and "Pacific Arts Presents, LLC." Nor did Reserve operate in a more businesslike fashion thereafter. No self-respecting insurance company would have paid $340,000 on the loss-of-customer claim without any investigation or supporting documentation. And it was clear from his testimony that Zumbaum, the Chief Executive Officer of Reserve, knew zero about its business. The Tax Court was fully warranted in its finding that "Reserve's transactions were not insurance transactions in the commonly accepted sense." Reserve, at 62.

Nevertheless, Reserve raises three challenges to the Tax Court's analysis of this issue: (1) the court mischaracterized Reserve's policies as providing only excess coverage; (2) the court erroneously determined that the premiums were unreasonable and not negotiated at arm's length; and (3) the court erroneously determined that Reserve was not operated like an insurance company because it was managed by hired professionals. We address these challenges in turn.

1. Alleged Mischaracterization of Reserve's Policies

Reserve contends that the Tax Court misinterpreted the other-insurance clauses in Reserve's policies to mean that they provided only excess coverage. But those clauses stated: "The coverages afforded by this policy are excess over any other valid and collectible insurance policy issued by any other insurer. The limits and deductibles stated herein only apply after coverage is exhausted from any and all other valid insurance policies issued by any other insurer." Reserve, at 14 (capitalization and ellipsis omitted). In other words, if Reserve's policies covered the same loss as another policy, then Reserve's liability would not arise until that other insurance was exhausted. See 15 Couch on Insurance, supra, § 219:1-2 (discussing purpose of other-insurance clauses). The Tax Court quoted this language in its recitation of the facts, and later observed that "the polic[ies] would be valid only after insurance coverage from other insurers was exhausted." Reserve, at 60. This observation is not clearly erroneous. To be sure, it appears that at least some of the policies issued by Reserve covered risks that were not covered by Peak's commercial policies; consequently, there would often be no coverage from other policies that Peak would have to exhaust. But Reserve has not pointed to any language by the Tax Court to the contrary. There was no mischaracterization of the policies.

2. Unreasonableness of Premiums

Reserve's second challenge is that the basis for the Tax Court's determination that the premiums were unreasonable and not negotiated at arm's length "flowed directly from its misreading of Reserve's policies" as providing only excess insurance. Aplt. Br. at 53. We beg to differ. First, as just stated, the Tax Court did not misread the policies. And there were multiple reasons supporting the Tax Court's determination. The observation that the policies provided only excess insurance was only a small component of the total analysis. We have repeatedly set forth a number of the questionable features of the policies issued by Reserve. Most important, however, was Reserve's utter failure to provide a reasonable explanation of how it calculated the risks covered by those policies. The best it could do at the hearing before the Tax Court was to suggest, through the testimony of experts who were not familiar with Peak, that the premiums were in line with the premiums charged for similar insurance by other captive insurers managed by Capstone. But there was no evidence that those other captive insurers faced risks similar to those of Peak, nor was there any evidence that the premiums charged by those other captive insurers were themselves reasonable. The Tax Court was fully justified when it said that "[t]aking into consideration all the surrounding facts and circumstances, . . . no unrelated party would reasonably agree to pay Reserve the premiums that Peak and the other insureds did for the coverage provided by the direct written policies," and then found that "the polic[i]es were not reasonable and not negotiated at arm's length." Reserve, at 61.

Reserve argues that the Tax Court should not have faulted Reserve for issuing policies in "the absence of a significant history of losses for Peak." Aplt. Br. at 54. It likens the Tax Court's analysis to "saying that automobile insurance is unnecessary for drivers who have not yet had an accident." Id. But the Tax Court did not conclude that Reserve's policies were per se unreasonable just because Peak did not have a history of losses that would be covered by the policies. Instead, it observed that the lack of loss history was suspicious given that (1) the Reserve policies increased Peak's insurance expenses by roughly 400%, (2) the policies' other-insurance clauses meant that Reserve's policies would provide only excess coverage if Peak's other policies covered the same loss, (3) "Peak had never come close to exhausting the policy limits of its third-party commercial insurance coverage," Reserve, at 60, and (4) Reserve provided no risk assessment to support the high premiums.

The Tax Court did not clearly err when it concluded that Reserve's premiums were unreasonable and not negotiated at arm's-length.

3. Irregularities of Reserve's Operation

Reserve also argues that the Tax Court should not have considered the fact that Reserve was managed by the hired professionals at Capstone and should not have faulted Zumbaum for being unable to discuss Reserve's policies in any detail. It says that "existing caselaw recogniz[es] that most captive insurance companies operate without any employees and routinely delegate operational functions, financial reporting, regulatory compliance and day-to-day tasks to captive managers." Aplt. Br. at 58. It says that it is "meaningless that [Zumbaum] could not recall details about the very subjects that he had hired professionals to handle," and likewise of no importance that Reserve was managed by Capstone. Id. at 60.

But the Tax Court did not criticize Reserve simply because it was managed by Capstone. Instead, the Tax Court observed that Reserve's operations were irregular because it had no employees, it maintained an address in Anguilla without ever doing business there, and Zumbaum — the co-owner, president, and CEO — "knew virtually nothing about its operations." Reserve, at 50. Even if we credit Reserve's position that captives are frequently managed by third parties, Reserve does not explain why Zumbaum — the responsible officer of both Reserve and Peak — could not testify about the specifics of any of the policies or their pricing even though McNeel testified that Zumbaum and Weikel were "the most significant" people in determining premium prices, Aplt. App., Vol. 5 at 1247, and Feldman testified that Zumbaum had the final say on "[p]olicies, pricing, which coverages, everything." Aplt. App., Vol. 5 at 1430. Moreover, Zumbaum's lack of knowledge about Reserve and Capstone's management of Reserve contradicts Capstone's feasibility study, which emphasized that a prerequisite to forming a captive was "wholehearted top management support" and "strong internal oversight." Aplt. App., Vol. 7 at 2039-40 (capitalization omitted). It is one thing to seek technical and administrative assistance from professionals; it is quite another to abandon all responsibility for how your company is handling $400,000 of insurance and simply believe everything Capstone says.

Further, the Tax Court also observed that Reserve's highly unprofessional handling of its only claim suggested that it was not operating as an insurer. The Tax Court observed that "no supporting documentation accompanied the claim notice," that "Peak did not submit and Reserve did not insist on obtaining any documents to substantiate the occurrence or the amount of the claimed loss," that the first payment was made to Peak "more than a month before" the settlement agreement, that the final payment — which increased the total payments from $165,820 to $339,820 — was made several months after the settlement agreement, that Peak and Reserve did not prepare an addendum to the settlement agreement reflecting the increase in payments until years later, and finally, that all Reserve's payments were issued via checks signed by a Peak employee. Reserve, at 52. (As shown by our description earlier in the opinion of how the claim was handled, these irregularities do not exhaust the improprieties in the processing.) Reserve does not dispute on appeal any of these findings regarding its claim handling, or even the Tax Court's conclusions from these findings.

The Tax Court did not clearly err in concluding that Reserve did not operate as an insurance company in the commonly accepted sense.

V. THE TAX COURT DID NOT ERR IN REFUSING TO RECHARACTERIZE PEAK'S "PREMIUMS" AS CAPITAL CONTRIBUTIONS

Having concluded that the premiums Reserve received from Peak and PoolRe were not payments for insurance, the IRS determined that the receipts instead constituted FDAP income, which is "fixed or determinable annual or periodical gains, profits, and income" that is "received from sources within the United States" but "not effectively connected with the conduct of a trade or business within the United States." I.R.C. § 881(a). Such income is taxed at a rate of 30%.

Reserve challenges this characterization. It argues that if it does not qualify for treatment as an insurance company under the Internal Revenue Code, the IRS should have treated its receipts as contributions to capital, which are not taxable income for the recipient. The Tax Court rejected this argument below, correctly noting that whether a payment constitutes a capital contribution turns on the motive of the payor. See United States v. Chi., Burlington & Quincy R.R. Co., 412 U.S. 401, 411 (1973) (identifying "the intent or motive of the transferor" as "determin[ing] the tax character of the transaction" when deciding whether payment by nonshareholder was a contribution to capital); Wash. Athletic Club v. United States, 614 F.2d 670, 674 (9th Cir. 1980) ("[T]he motive or purpose and intent [of a club member] in making the contributions is a dominant factor in determining whether [the payments were] a capital contribution or payment for goods and services." (internal quotation marks omitted)); Bd. of Trade v. Comm'r, 106 T.C. 369, 381 (1996) (for both shareholders and nonshareholders, payor's intent controls whether a payment is a contribution to capital); cf. Sammons v. Comm'r, 472 F.2d 449, 451-53 (5th Cir. 1972) (applying both objective test and subjective-purpose test to determine whether to characterize the transfer of property from one corporation to another corporation as a dividend to an individual who has an ownership interest in both corporations followed by a capital contribution to the transferee corporation). The court pointed out that Reserve had the burden to show that the payments it received were not FDAP income but "did not produce evidence" supporting that position. Reserve, at 64. It said that "[t]he record does not reflect that the parties to the purported insurance transactions treated or intended the amounts paid to Reserve as additional capital contributions." Id. at 65. The absence of evidence is particularly striking since Zumbaum was a witness at the hearing and could have testified to a capital-contribution intent (although such testimony would, of course, have undermined the argument that the premiums were for insurance coverage).

We review for clear error a Tax Court finding that a taxpayer has not satisfied its burden of persuasion on a factual element of its claim. See Hamilton, 955 F.3d at 1172; Wash. Mut., Inc. v. United States, 856 F.3d 711, 721 (9th Cir. 2017) (reviewing failure of proof in tax-refund case for clear error).

On appeal Reserve does not dispute that it had the burden of proving its receipts were not FDAP income. Nor does it dispute that if its receipts were not capital contributions, they meet the criteria for FDAP income. It also does not dispute that it listed these payments as income on its tax returns and Peak reported them as deductible expenses; and it does not dispute that the payor's intent is a crucial element when determining whether a payment constitutes a capital contribution. It does not even dispute that it provided no direct or indirect evidence that its receipts were intended as capital contributions rather than business income. On the contrary, Reserve asserts that Peak did not intend to contribute to capital, stating, "Both the payors (the Direct Insureds) and the payee (Reserve) undeniably intended the payments to be insurance premiums." Reply Br. at 31. On this record alone, we can conclude that the Tax Court did not clearly err by finding Reserve failed to satisfy its burden.

Reserve nevertheless argues that because the Tax Court found no legitimate business purpose or non-tax reason for the issuance of the insurance policies, the only logical conclusion that the Tax Court could draw is that the payments from Peak to Reserve — the purported policy premiums — were capital contributions. See, e.g., Aplt. Br. at 63 ("The tax court's determination that there was no legitimate non-tax reason for Reserve's receipt of the payments is dispositive of the issue of the characterization of the amounts received from the perspective of Reserve."). It points to IRS Revenue Ruling 2005-40, 2005-2 C.B. 4, 2005 WL 1415557 (2005), which states that "an arrangement that purports to be an insurance contract but lacks the requisite risk distribution may instead be characterized as a deposit arrangement, a loan, a contribution to capital . . ., an indemnity arrangement that is not an insurance contract, or otherwise." Id. at *1. Reserve claims that among these options, the only possible characterization for a transaction without a business purpose or nontax reason is as a contribution to capital. Reserve relies on Carnation Co. v. Commissioner, 71 T.C. 400 (1978), aff'd, 640 F.2d 1010 (9th Cir. 1981), where the Tax Court determined that a corporation's payment to a subsidiary insurance company was not an insurance premium and recharacterized that payment as a capital contribution, see 71 T.C. at 415.

We disagree. To begin with, Carnation is not persuasive authority that whenever an affiliate of a captive insurer deducts a purported "premium" paid to the insurer and the insurer is found not to be an insurance company, then the premium should be recharacterized as a capital contribution to the insurer. In Carnation there was no dispute between the parties regarding how the payments should be recharacterized, see id.; and there was compelling evidence of intent in any event, since the parent had entered into an agreement with a third-party insurance company to provide an additional $3 million of capitalization to the subsidiary insurance company, see id. at 403-04. Nor is the situation here like that where a portion of a purported deductible business expense paid by one subsidiary to another is properly recharacterized as a dividend from the paying subsidiary to the parent company followed by a capital contribution from the parent to the other subsidiary. This may occur when the paying subsidiary pays an excessive amount (well above what would have been paid in an arm's-length transaction) for goods or services from a second subsidiary. In that circumstance, the Commissioner can consider the larger picture and infer that the purpose of the transaction was to increase the capital of the second subsidiary by, in essence, making a distribution (a taxable dividend) of the excess price from the first subsidiary to the parent, which then transfers that excess to the second subsidiary as an infusion of capital. See IRS Revenue Ruling 78-83, 1978-1 C.B. 79, 1978 WL 42300 (1978) (engaging in such reallocation under authority of IRC § 482, which permits Commissioner to reallocate gross income, deductions, etc., between two organizations controlled by the same interests if Commissioner "determines that such distribution, apportionment, or allocation is necessary in order to prevent an evasion of taxes or clearly to reflect the income of any of such organizations"); Sammons, 472 F.2d at 452-53 ("The concept that a transfer of property between corporations with common ownership may constitute a dividend to the common owner, even when the stockholder has not received funds or property from either corporation, has been well established. . . . In [this] situation . . . the theory is that the funds pass from the transferor to the common stockholder as a dividend and then to the transferee as a capital contribution." (emphasis added)). But Reserve can point to no evidence, and has conceded there is none, that would support such an inference regarding Peak's purpose here. In particular, it has proffered no reason why Peak would want to make a capital contribution to Reserve.

Revenue Ruling 2005-40 does not declare that when an ostensible insurance arrangement fails to adequately distribute risk it must be recharacterized as a contribution to capital. The Revenue Ruling mentions some alternatives but those are not exclusive. Whether an alternative characterization is appropriate depends upon both the objective facts of the transaction and the subjective intent of the parties. Here, the objective reality is that Peak and Reserve entered into contracts that required Reserve to pay Peak if Peak suffered losses covered by the contracts. Although the Tax Court held that those contracts were not insurance contracts and that the policies, with their unreasonable premiums, had no legitimate business purpose, that does not mean that the transfer of funds from Peak to Reserve could not serve a legitimate business purpose. The court's findings are consistent with the characterization of the transaction proffered by the Commissioner — namely, that the transaction was just a movement of funds offshore "to self-insure against business losses." Aplee. Br. at 66. Reserve does not challenge the Commissioner's assertion that the receipt of money by Reserve in such a transaction would be taxable under IRC § 881(a), by analogy to the taxation of the receipt of payments for underwriting or for providing a guarantee. Cf. Gulf Oil Corp. v. Comm'r, 914 F.2d 396, 411-13 (3d Cir. 1990) (although premiums paid to an insurance subsidiary were not deductible as payments for insurance because the risk was not truly distributed, court refused to characterize the premium payments by subsidiaries as dividends to the parent corporation followed by a capital contribution from the parent to the purported insurance company because the insurance subsidiary did provide a benefit — risk coverage — to its affiliates).

In any event, what is totally absent here is proof of the essential element of motive, purpose, or intent — evidence from which one can infer that Peak intended to make a capital contribution. The Tax Court's finding that the policies issued by Reserve were not true insurance policies said nothing about Peak's intent. The finding was based on objective facts regarding risk distribution and conducting affairs as an insurance business. Reserve has not suggested any reason why Peak or its owners would want to make a capital contribution to Reserve. Nor has it presented any argument why a factfinder could not infer that Peak's intent was simply the intent to create a plausible insurance company through which Peak could obtain a substantial tax deduction without reducing the funds available to its two owners. The intent behind the act does not change just because the act failed to achieve its purpose. (In fact, the arrangement did accomplish Peak's purpose. Counsel for the Commissioner reported at oral argument that it was too late to challenge Peak's business deductions for insurance premiums paid to Reserve for 2008-10.) By the same token, Reserve earned income by creating the appearance of issuing insurance policies to Peak, even if it ultimately failed to convince the Commissioner that the policies were bona fide. Reserve could have reported the transactions with Peak on its tax returns as the receipt of capital contributions, but it did not do so. "[W]hile a taxpayer is free to organize his affairs as he chooses, nevertheless, once having done so, he must accept the tax consequences of his choice, whether contemplated or not, and may not enjoy the benefit of some other route he might have chosen to follow but did not." Comm'r v. Nat'l Alfalfa Dehydrating & Milling Co., 417 U.S. 134, 149 (1974) (citations omitted). We have been pointed to nothing in the Tax Code that required the Commissioner to recharacterize the obvious intent of the parties to enable Reserve to escape taxation.

We must emphasize, however, that the Commissioner did not need to prove that the transfer of funds from Peak to Reserve was not a capital contribution. Reserve bears the burden of persuasion. In particular, Reserve had to prove Peak's intent to make a capital contribution. To rule against Reserve, it sufficed for the Tax Court to not be persuaded that this was Peak's intent. That court did not commit clear error in failing to be persuaded.

VI. CONCLUSION

We AFFIRM the judgment of the Tax Court.

FOOTNOTES

1. Strictly speaking this was not a reinsurance agreement since PoolRe was directly liable to the insured (such as Peak) under an endorsement on the policy from the captive insurer (such as Reserve) to the insured.

2. Reserve appealed to this circuit, perhaps because it filed its tax return in Utah. See 26 U.S.C. § 7482(b)(1)(B). The Commissioner has not challenged venue.

3. For simplicity and uniformity we refer to the pagination of the Tax Court's memorandum opinion throughout this opinion.

4. Besides dropping the three policies in 2009, Peak replaced the expense-reimbursement policy by two policies covering the same risks — one to mitigate adverse publicity and one to provide for civil-liability defense costs. Also, on six policies the liability limit was reduced from $1 million to $500,000, so the total limit of liability dropped from $13 to 8 million. Peak paid premiums of $448,127.03 on the 11 policies in 2009. Peak kept the same 11 policies and paid $445,314.01 in 2010.

5. In a letter to the court sent a few days after oral argument, counsel for Reserve retracted this statement, stating that there was no proration error. The letter asserts that the proration factor on the worksheet reflected McNeel's "judgment of the percentage of risk of the policy remaining," and notes that the premium was comparable to that on a document provided by Mid-Continent General Agency, Inc., which purportedly was produced independently and provided premiums that "were pro-rated for short term policies." Appellant Reserve Mechanical Corp.'s Suppl. Letter Br., at 2. What is absent from the retraction, however, is any plausible explanation of why the regulatory-changes premium for one month of coverage was essentially the same as the premium for a year's coverage, particularly when the premiums for the other non-retroactive policies were treated so differently. The loss-of-major-customer premium jumped from $7,268 (for a policy limit of $1 million) for the one-month coverage in 2008 to $50,625 (for a $500,000 limit) for the year-long coverage in 2009; the loss-of-services premium jumped from $4,874 ($1 million policy limit) in 2008 to $62,791 ($1 million limit) in 2009; and the product-recall-reimbursement premium jumped from $5,087 to $35,438 even though the policy limit dropped from $1 million to $500,000. (The weather-related-business-interruption policy was not continued in 2009.) In contrast, for the regulatory-changes coverage, the one-month premium in 2008 of $64,899 (for a $1 million policy limit) dropped to $47,588 for annual coverage in 2009 (with a $500,000 policy limit). In examining whether the Tax Court properly determined that Reserve did not satisfy its burden of establishing that it was providing insurance, we are not inclined to give any weight to an ipse dixit by counsel.

6. Limit factors are used to compute what the premium should be for a particular liability limit based on the premium already determined for a different liability limit. Because the likelihood of a large loss is generally significantly less than the likelihood of a smaller loss, the premium is not proportional to the liability limit. For example, if the proper premium for $250,000 of coverage is $1,000, one could calculate the proper premium for $1 million coverage by multiplying $1,000 by the appropriate limit factor (3.4 in McNeel's calculations) to arrive at a premium of $3,400.

7. Altogether, Reserve did not provide the advantages touted by the feasibility study. Rather than Peak having "complete control" over Reserve, Aplt. App., Vol. 7 at 2042, Reserve was managed entirely by Capstone. Rather than Reserve providing "manuscript" coverage, id. at 2041, it issued boilerplate policies. And rather than providing insurance that is "generally unavailable or impractical to obtain in the conventional insurance marketplace," id. at 2030, at least six of Peak's policies were commercially available.

8. The December 2008 policy indemnified Peak "for any Business Interruption loss of up to twelve (12 months) suffered as a result of losing the services of a Major Customer." Aplt. App., Vol. 11 at 3126 (emphasis added). The 2009 policy, which took effect five days before the "date of occurrence" stated on the Notice of Loss, changed the language to encompass "any Business Interruption loss . . . suffered as a result of the loss of or reduction of services of a Major Customer." Aplt. App., Vol. 12 at 3302 (emphasis added). The loss on which Reserve paid Peak — which was not for the total loss of services but just a reduction of services — may not have been covered under the language of the December 2008 policy.

9. In 2010 the conditions for payment under the stop-loss coverage were revised. For that year the stop-loss insurance from PoolRe would have been triggered if the total claims submitted from Peak to Reserve reached 35% of the total, aggregate premiums for all of Reserve's direct policies. Above that threshold, PoolRe would pay 50% of the claims, though it would not pay more than the total premiums themselves. There were no longer any separate Attachment Points that needed to be satisfied.

10. In 2010, in addition to the previously discussed changes in stop-loss coverage, Capstone changed the premium percentage received by PoolRe to 19.9%. Reserve submitted a letter from Glicksman Consulting, LLC in support of this number. But that letter did not assess any risk. Rather, it presented a range of options based on a number of (unsubstantiated) risk assumptions provided by PoolRe and said that it was beyond "the scope of th[e] review to recommend a specific change" to the stop-loss premium rate. Aplt. App., Vol. 12 at 3564.

11. Say the amount of premiums paid by Peak for its insurance was identical to the amount paid by the insureds of captive insurer A in the pool, but Peak was only 1% as likely as the insureds of A to incur a loss covered by the stop-loss coverage provided by PoolRe. Reserve and A each would forgo the same amount in premiums (18.5% of the total premiums in 2009) to PoolRe and each would be paid that same amount from PoolRe for reinsurance premiums. So Reserve and A would each effectively receive all premiums paid by its insureds. The problem would arise from the discrepancy between the liability each insurer would incur for reinsuring PoolRe's stop-loss liability and the benefit each would receive from that coverage. Since each had to forgo (at least in form) the same amount in premiums to PoolRe, each would be liable for the same percentage of losses incurred by PoolRe when it had to pay on claims under the stop-loss coverage to any of the captive insurers in the pool. This would be highly unfair to Reserve under our assumption that Reserve would expect to receive only 1% as much benefit as A would receive from the stop-loss coverage. Reserve would be shelling out a lot more to reinsure A's losses than it would be getting in return from A. If the stop-loss risks to the other captive insurers in the pool were more like the risks to A than the risks to Reserve, it would be foolish for Reserve to join the PoolRe pool.

12. The Tax Court did not address three factors — comparable commercial coverage, adequate capitalization, and a separate account for claim payments. Neither party complains about the failure to address those factors.

13. Reserve had plenty of capital if, as may well have been the case, the probability of a covered loss was extremely low. But, given the low probability that an attachment point would ever be satisfied, Reserve could not count on any reinsurance backup in case of a major loss that it could not pay out of its own funds, and it would take several years of accumulating premiums before Reserve could have sufficient capital to pay more than one policy-limits loss out of its own funds.

14. We note that the definition of insurance company cited by Reserve is "any company more than half of the business of which during the taxable year is the issuing of insurance or annuity contracts or the reinsuring of risks underwritten by insurance companies." I.R.C. § 816(a) (emphasis added). It is not clear whether reinsurance of risks that have not been underwritten by an insurance company would qualify in the way Reserve hopes.