The Tax Court's decision regarding captive insurance arrangements in Rent-A-Center, 142 T.C. No. 1 (2014), was an overall taxpayer-friendly result—albeit a lengthy opinion with a concurrence and two dissenting opinions. The Tax Court reached a similar decision in Securitas Holdings, Inc., T.C. Memo. 2014-225. Although both cases had complex facts (that ultimately favored the taxpayers) and both rulings delved into many aspects of captive insurance law, this item focuses specifically on the implications for satisfying the "risk distribution" requirement for having a valid captive insurance arrangement.
Captive insurance arrangements are often pursued as an alternative to third-party insurance and risk protection. In a captive arrangement, a controlled subsidiary of a group of entities (often corporations) typically operates as an insurance company, providing insurance against loss risks affecting those entities. The captive insurance subsidiary can provide insurance to the various brother-sister members of the group and/or write insurance policies to third-party customers outside the group. From a business perspective, this type of arrangement may provide a better method for managing group risks as well as potentially reducing the premiums needed to insure against those risks. (This is because the insurance company within the group may have lower overhead costs than a third-party insurer, or the group overall may have less risk of loss and would pay higher rates to a third-party insurer that also covered riskier insureds.) From a tax perspective, the captive insurance arrangement provides the benefit of a deduction for loss reserves. Also, insureds are allowed to deduct insurance premiums paid.
Case law has developed a four-factor test for determining whether a putative captive insurance arrangement qualifies as a Sec. 831 insurance company (and, thus, can derive the tax benefits above): (1) The risks being insured must be insurable; i.e., there must be a probability of a future risk-of-loss event's occurring and a quantification of possible loss (e.g., if a loss event has already occurred and the only issue is the extent of the damages, that is not insurance); (2) the arrangement must be insurance in the commonly accepted sense; thus, the putative captive company should operate like a true insurance company—a "Potemkin village" arrangement, in which the captive does not act in substance like a true insurer, will not do; (3) risk shifting must occur; i.e., the risk of loss must shift off the balance sheet of the insured to the insurer; and (4) there must be risk distribution, the focus of this item.
To have risk distribution, the number of independent risks being insured in a pool must be large enough for the law of large numbers to operate. The law of large numbers is a statistical concept relating to probability that in the context of insurance means that the greater the number of units that are individually exposed to risk of loss, the greater the likelihood that the actual loss that occurs due to that exposure will equal the expected loss. Suppose a given calamity has a probability of occurrence of one in a million; however, the loss from such a calamity is $10 million. If this can occur only once, the result is binary: either nothing is lost or all $10 million is lost. However, if there are numerous independent potential loss events, the expectation for each such event is the probability multiplied by the loss, or $10.
In a brother-sister arrangement, the IRS has traditionally focused on the number of tax-recognized entities in the structure and the policies written to each such member. Two situations examined in Rev. Rul. 2005-40 provide important guidance in this area. In Situation 3, a parent-captive arrangement had 12 limited liability companies (LLCs) that were considered disregarded entities for federal income tax purposes as the insureds, no one of which accounted for less than 5% or more than 15% of the total risk assumed. In Situation 4, the facts were the same, except the 12 entities were subsidiary corporations. In the former situation, the IRS held that no risk distribution was present, as the risk was only shifted but not distributed, focusing on the number of insureds from a federal income tax perspective. In the latter situation, the IRS held that there was risk distribution.
This was a troubling position, as the focus appeared to be on the number of insureds being the key to risk distribution, as opposed to the underlying risks themselves. Moreover, both the disregarded LLCs and corporations were respected as legal business entities, so their mere tax status's differentiating between a valid and invalid arrangement was also troubling.
On the other hand, the courts, albeit in dicta, appear to indicate that a single insured could be a valid captive arrangement if there are enough insurable risks to satisfy statistical risk distribution. In Malone & Hyde, Inc., T.C. Memo. 1993-585, the Tax Court said that "[w]hen an insurer has a sufficiently large number of risks such that great variations in aggregate losses are unlikely, and the premiums received plus its capital make it a viable risk bearer, one can say that risk distribution is present regardless of the number of insureds covered" (quoting Winslow, "Tax Avoidance and the Definition of Insurance: The Continuing Examination of Captive Insurance Companies," 40 Case Western Reserve L. Rev. 79, 152 (1989–90)). Note that this case was decided in 1993, before the revenue ruling. A similar statement was made in dicta in Gulf Oil Corp., 89 T.C. 1010 (1987).
From 2005 until 2014, it was uncertain whether a putative captive arrangement that fell short of the Rev. Rul. 2005-40 safe harbor based solely upon the number of insureds would fail to qualify. Then the Tax Court in Rent-A-Center held that "[r]isk distribution occurs when an insurer pools a large enough collection of unrelated risks" and, "[i]n analyzing risk distribution, we look at the actions of the insurer because it is the insurer's, not the insured's, risk that is reduced by risk distribution." The court looked at the presence of a "sufficient number of statistically independent risks" and not the number of insureds. In Securitas Holdings, the Tax Court stated that "[w]e evaluate risk distribution through the actions of the insurer" and cited the language in Rent-A-Center. The Securitas Holdings decision even more plainly stated that "[r]isk distribution incorporates the law of large numbers which has been described as follows: 'As the size of the pool increases, the chance that the loss per policy during any given period will deviate from the expected loss by a given amount (or proportion) declines' " (citation omitted).
Thus, it appears taxpayers are now on firmer footing when it comes to captive arrangements that fall outside the Rev. Rul. 2005-40 safe harbor. While the IRS may still challenge such arrangements, it appears the Tax Court is consistently applying a standard that does not focus on the number of tax entities within the brother-sister group.
Greg Fairbanks is a tax senior manager with Grant Thornton LLP in Washington.
For additional information about these items, contact Mr. Fairbanks at 202-521-1503 or firstname.lastname@example.org.
Unless otherwise noted, contributors are members of or associated with Grant Thornton LLP.