Originally published in Captive Review, September 2010.
Phillip England of law firm Anderson Kill & Olick and Randall Beckie of Frontrunner Captive Management outline how insurance companies should look at risk distribution as a hallmark of capital efficiency
Risk distribution, a defining characteristic of an insurance company, continues to be a topic of US tax controversy for captives. Still today, a quarter of a century after the controversy took root, controversy lingers because risk distribution has not been defined in a consistent way by the courts and the Internal Revenue Service (IRS). What has been absent from both IRS and judicial guidance is the context of why risk distribution is a defining characteristic of an insurance company.
Risk pooling
We might suggest that the context should be: by pooling risk, an insurance company becomes able to fund adverse-loss scenarios with less capital than if risks were not pooled. Risk pooling enables an insurance company to use capital efficiently. Long-standing tax jurisprudence has defined an insurance company as a corporation whose capital and efforts are primarily devoted to the earning of income through premiums charged for the issuance of its contracts of insurance, of whatever nature they may be. By evaluating an insurance company's tax status according to how its capital is used, the courts have opened the door to measure the significance of risk distribution in terms of the effect on an insurance company's need for capital. Measuring risk distribution for an insurance company's capital efficiency would imply that risk distribution might not necessarily require a large diversity of insureds but would instead require a critical mass of covered riskexposure units.
An example of identifying risk distribution by reference to an insurer's capital efficiency is: assume an insurer issues 100 policies, each of which covers a single potential claim limited to $10, and the premium is $1 for each policy. Aggregate premiums total $100, and aggregate exposure to loss totals $1,000. At a conventional 3:1 premium to surplus ratio, the insurer would carry $33 of capital to issue these policies, which amounts to an efficient use of capital to support $900 of maximum net-underwriting losses in the aggregate. The phenomenon of risk distribution is evident by the fact that the insurer needs $33 of capital rather than $900 of capital.
Multiple insureds: how many is enough to make a risk pool?
In the IRS' view, risk distribution has two main definitional characteristics:
- Multiple insureds.
- A multitude of covered riskexposure units.
The IRS says a captive must demonstrate both characteristics. From the position of the IRS, it seems correct to observe that court decisions suggest that insurance presupposes pooling of risk among more than one insured. However, there is scant authority (other than rulings issued by the IRS itself) for the proposition that risk distribution requires more than two insureds. Revenue Ruling 2002- 90 suggests that a risk pool requires 12 insureds, but there is no authority for requiring 12. Revenue Ruling 2002-91 suggests that a group captive requires pooling among at least seven insureds, but the authority for requiring seven appears to be based on what we call the 'IRS' 15% limitation', and this limitation in turn overreaches in applying relevant authority.
The IRS' 15% limitation posits that no single insured may lead to more than 15% of the insurer's premiums or 15% of the insurer's aggregate exposure to loss. The theory behind the 15% limitation is that an insured must not in a significant part pay for its own risks. IRS rulings attribute this notion to the Humana case, although the Humana case did not reach any such holding. Humana did say (in dicta) that risk distribution exists where there are several insureds (not quantifying 'several'), although the IRS' formal position in Rev. Rul. 2002-90 ignores that.
There is authority for the notion that two insureds (specifically, two hospitals) are enough to comprise an insurance company, at least in a situation where covered risk-exposure units are multitudinous. In its General Explanation of the Tax Reform Act of 1986, the Joint Committee staff, addressing the reasons for amending Section 501(m) of the Internal Revenue Code, stated: "The providing of insurance benefits by an organisation otherwise described in sec. 501(c)(3) generally is considered a commercial activity that does not meet the requirements for tax-exempt status. For example, if two or more unrelated tax-exempt organisations pool funds for the purpose of accumulating and holding funds to be used to satisfy malpractice claims against the organisations, the organisation holding the pooled funds is not entitled to tax exemption because the activity (for example, the provision of insurance) is inherently commercial in nature."
To say that two insureds are enough to make an insurance company is to define an insurance company according to its distinction from an arrangement of self-insurance. Perhaps that explains why risk distribution presupposes multiple insureds, because if an insurer covers merely one insured, the insured's premiums would have to fund entirely such insured's own losses (assuming that the insurer operates at breakeven or better). However, when the premiums of two insureds are co-mingled to fund claims from either one, the arrangement economically no longer has the character of self-insurance. Broadly speaking, the case law on the topic of 'what is insurance' is concerned with distinguishing insurance from self-insurance. It would seem reasonable to define risk distribution in the same context as the tax law defines an insurance company in general, that is, by contrast with self-insurance.
By viewing two insureds as the minimum needed to demonstrate sufficient risk distribution, we would resolve the contraction inherent in the IRS' position. In Rev. Rul. 2002- 90, the IRS suggests that risk distribution would be undermined if any single insured accounts for more than 15% of the captive's premiums. Yet in Rev. Rul. 2002-89, the IRS says that risk distribution would be sufficient where one insured accounts for 50% of the captive's premiums, provided that the remaining premiums come from multiple insureds. How much premium (as a portion of the captive's total premiums) is any single insured permitted to pay?
Tax-court precedent
Arguably, the tax court has already addressed this question. With Harper Group, the tax court viewed the parent company and subsidiaries as a single economic family, yet held that such an economic family achieved risk distribution in its captiveinsurance arrangement where the economic family originated 71% of the premiums and unrelated parties (namely customers) originated 29% of the captive's premiums. Logically it follows that if risk distribution can exist where one insured accounts for 71% of the premium; risk distribution could exist among as few as two insureds (provided that covered riskexposure units are multitudinous).
This interpretation of Harper Group has been advanced at least twice: in technical advice memorandum 200323026 and in private-letter ruling 201025077. In both private rulings the IRS ignored the fact that the tax court viewed the parent and subsidiaries as a single economic family. It is difficult to imagine how the IRS could get away with ignoring the precedent in Harper Group and Humana and tax-legislative history if a taxpayer were to vigorously defend its captive insurance arrangement. Our thought here is that a taxpayer's defence could be invigorated by advancing a theoretical concept of risk distribution that is more consistent with the analytical frameworks that the courts have already developed.
Phillip England is a shareholder of the law firm of Anderson Kill & Olick, where he heads the tax practice and the captive advisory practice.
Randall Beckie, CPA, provides insurance-tax consulting support to Anderson Kill & Olick's captive practice and also runs Frontrunner Captive Management.
About Anderson Kill & Olick, P.C.
Anderson Kill practices law in the areas of Insurance Recovery, Anti-Counterfeiting, Antitrust, Bankruptcy, Commercial Litigation, Corporate & Securities, Employment & Labor Law, Real Estate & Construction, Tax, and Trusts & Estates. Best-known for its work in insurance recovery, the firm represents policyholders only in insurance coverage disputes, with no ties to insurance companies and no conflicts of interest. Clients include Fortune 1000 companies, small and medium-sized businesses, governmental entities, and nonprofits as well as personal estates. Based in New York City, the firm has offices in Newark, NJ, Philadelphia, PA, Ventura, CA and Washington, DC and Greenwich, CT. For companies seeking to do business internationally, Anderson Kill, through its membership in Interleges, a consortium of similar law firms in some 20 countries, assures the same high quality of service throughout the world that it provides itself here in the United States.
Anderson Kill represents policyholders only in insurance coverage disputes, with no ties to insurance companies, no conflicts of interest, and no compromises in it's devotion to policyholder interests alone.
The information appearing in this article does not constitute legal advice or opinion. Such advice and opinion are provided by the firm only upon engagement with respect to specific factual situations
