Tax Shelter Litigation

IRS Loses Captive Insurance Case on Good Taxpayer Facts

Written by: F. Hale Stewart and Beckett G. Cantley

In 2002 the IRS issued several revenue rulings that established two safe harbor rulings for the appropriate structure of a captive insurance company.1 Those rulings followed — and were the result of — a long case history that saw taxpayers successfully rebut an IRS argument that a captive insurance company was not a bona fide insurer and therefore was not eligible for deductions related to the paid-in insurance premiums. 

The IRS losses began in the landmark Humana case,2 when the court held that insurance premiums paid by brother-sister subsidiaries were deductible. Those losses continued in AMERCO,3 when the Service unsuccessfully challenged a captive company established by U-Haul International Inc., arguing that Republic Western Insurance Co. was not a viable, stand-alone insurer. The final loss was handed down in UPS,4 in which the government lost on appeal. 

For about a decade after the 2002 rulings (until approximately 2010-2011), the Service was relatively quiet regarding captive insurance companies. However, the recent Securitas Holdings case5 — which followed the Rent-A-Center6 decision earlier this year — resulted in another loss for the Service in its attempts to challenge captive insurance companies established and operated by larger companies.

Securitas AB is a Swedish company that established SHI Holdings in 1999. SHI then acquired Pinkerton’s Inc., which had approximately 48,000 employees and 250 worldwide offices. The subsidiary purchased additional security companies in the early 2000s, eventually providing the services through nearly 100,000 employees and 2,500 vehicles. Other business lines included cash handling, alarm installation, and monitoring. 

SHI also acquired Protectors Insurance Company, a Vermont licensed captive, in 2000. While Protectors had no employees, it maintained other indicia of separate corporate existence, including a unique bank account, individualized financial records, regulation by Vermont’s Department of Captive Insurance, and the holding of the annual board of directors meetings. In 2002 the Securitas corporate group formed another captive, Securitas Group Reinsurance Ltd. (SGRL), an Irish reinsurer. By 2004 SGRL had more than $77 million in funds. Like Protectors, SGRL observed all the requisite corporate formalities, including keeping separate books and records and maintaining a separate bank account. 

SHI subsidiaries had large insurance expenses, which were purchased from third-party companies. As the insurance market hardened in the early 2000s, SHI purchased a deductible reimbursement policy from Protectors. The captive program allowed SHI to better predict insurance expenses, lower costs, and centralize risk management. Protectors reinsured its risks through SGRL. Protectors paid premiums totaling approximately $77 million in 2003 and $56.9 million in 2004, as verified by outside actuaries. 

In 2002 SHI purchased another captive insurance company, Centaur. That captive had section 501(c)(15) status, meaning it could operate tax-free if it received less than $600,000 in premiums per year.7 However, that amount is based on the total premiums paid by all insurance companies in the corporate group.8 To prevent the Service from revoking Centaur’s section 501(c)(15) designation, SHI signed a parental guarantee for Protectors, allowing SHI to argue that for federal tax purposes, Protectors was not an insurance company.9 

Protectors underwrote a deductible reimbursement policy — a standard captive insurance coverage. It had ample corporate substance, evidenced by a unique paper trail, a separate tax return, and regularly held the board of directors meetings. Outside actuaries verified the level of insurance premiums. There is no mention of tax motivation for the transaction. The only questionable fact is the existence of a parental guarantee, which proved fatal in earlier captive cases — and the existence of which led to the Service’s challenge of Rent-A-Center’s captive. 

The court used a modified version of the Harper10 test to determine if Protectors was a valid insurance company (while Harper usually has three factors, the court added a fourth): 

1. the arrangement must involve insurable risks; 

2. the arrangement must shift the risk of loss to the insurer; 

3. the insurer must distribute the risks among its policyholders; and 

4 the arrangement must be insurance in the commonly accepted sense. 

Because both parties stipulated that the arrangement contained insurable risks, the court next turned to the issue of risk shifting, which simply means that in the event an insurance risk occurs, the loss is shifted from the insured to the insurer. That is accomplished by the indemnity payment that makes the insured financially whole. In analyzing that factor, the court used the balance sheet (or net worth) analysis, which looks at the insured’s financial position before the claim is filed and after the company has received an indemnification payment. The Service, relying on previous cases with a similar guarantee, argued that the guarantee also prevented risk-shifting. The Service’s reasoning was based on the potential three-step circular cash flow that could result from the guarantee: (1) The parent would file a claim with an undercapitalized captive; (2) the parent would make a payment on its guarantee to the captive; and (3) the captive would then make an indemnification payment composed of the guarantee funds, thereby returning the parent’s payment. However, that hypothetical fact pattern assumes the captive is undercapitalized, triggering the guarantee. In Securitas, the captive was adequately capitalized, making the possibility of a guarantee payment remote. Also, the primary purpose of SHI’s guarantee was not to supply the captive with funds but to allow Centaur to maintain its tax-exempt status. That factor, combined with Protectors’ adequate capitalization, allowed the court to distinguish the case from others in which the parental guarantee proved to be fatal. 

The court then turned to the question of risk distribution, which is viewed from the insurance company’s perspective and occurs when the insurer ‘‘pools a large enough collection of unrelated risks, those that are not generally affected by the same circumstance or event.’’11 Central to the Service’s argument is the requirement that the insurer bear risks from other companies.12 But here, SHI employed more than 100,000 employees and operated more than 2,000 vehicles. And the reinsurer, SGRL, received premiums from more than 25 companies in 2003 and 45 in 2004. That pooling of a large number of exposures — regardless of their ultimate source — was sufficient to achieve risk distribution. 

Finally, the court held on whether or not the arrangement was for insurance in its commonly accepted sense, and found adequate facts to supporthat factor. Both Protectors and SGRL: 

were governed by various governmental authorities; issued policies; charged reasonable premiums; had those premiums verified by outside actuaries; paid claims; kept their own books and records; maintained their own financial accounts; and •were adequately capitalized.

Those factors combined to create viable, legally recognizable insurance companies. In our estimation, the IRS has lost the recent cases because there were facts favorable to the taxpayer. The IRS is now turning its attention to section 831(b) cases with much less favorable taxpayer facts in which a court is not as likely to see things as the taxpayer does. It is understandable that the IRS would have a difficult time with large taxpayer/ widely distributed risk situations. However, when it gets to the small business sector, the Rent-A-Center and Securitas precedent will not be of much assistance. We expect the next wave of captive litigation to be more fruitful for the IRS. As such, we send the following warning to practitioners: Be careful out there. 

Back to Top