The Empire Strikes Back
This is Part 1 of a series on the recent Avrahami decision.
By Sean King, JD, CPA, MAcc
Last week the United States Tax court issued the long-awaited opinion in the Avrahami captive insurance case. And after more than 30 years of losing most every captive insurance case of consequence, the IRS finally got a win.
The Avrahami case involved a small, foreign-domiciled captive insurance company (“CIC”) that had made both the 953(d) election (to be taxed as a US entity) and the 831(b) election (giving it favorable US tax treatment on its underwriting profits). The CIC in question insured the Avrahami’s operating businesses, primarily jewelry stores in Arizona, against a wide variety of risks. Importantly, the Avrahami’s businesses were the sole insured entities under most policies issued by the CIC. The only exception was terrorism insurance: The Avrahami’s CIC did act as a reinsurer for terrorism policies issued by a fronting company to many unrelated insureds. These reinsurance premiums accounted for 30% of the total premiums received by the Avrahami’s CIC. Captives managed by CIC Services, LLC achieve Risk Distribution in a far more robust manner than the Avrahami captive structure.
A number of troubling factors unique to the Avrahami case make it unlike most captive insurance arrangement with which we and the captive industry are familiar. Consequently, it doesn’t offer much useful guidance to the industry. In Avrahami, most of the captive’s assets were loaned out to the Avrahamis or related entities. This is unusual. The Avrahami captive issued policies with seemingly unclear and/or contradictory terms (also unusual). The court found that it charged unreasonably high premiums (though this determination is questionable as noted below and in subsequent parts of this series). The Avrahami captive pooled only terrorism risk rather than all or most risks (most captives today book all or most risks). Avrahami’s lawyer had (arguably) directed the actuary to arrive at a certain pre-determined price for some policies, suggesting that the transaction was motivated more by tax considerations than risk management concerns.
The court’s ruling centered primarily upon the concept of “risk distribution.” Risk distribution is inherent in the very definition of insurance. Risk distribution is the ability of the insurance company to spread its risks among multiple insureds. If the insurance company can’t use Peter’s premium to help pay some of Paul’s claim, then no real insurance exists. Without risk distribution, the money set aside as “premiums” is little more than a nondeductible loss reserve.
Remember that under all but the Avrahami’s terrorism policy, there were no other insureds save the Avrahami’s own entities, and those entities only numbered three or four. The court first ruled (no surprise really) that spreading risk among only three or four entities owned by the same owner is not sufficient to achieve adequate risk distribution (at least when those entities don’t themselves contain a critical number of “exposure units”). This meant that any risk distribution achieved via the terrorism reinsurance arrangement was all that stood between the Avrahamis and defeat.
Judge Concluded Insufficient Risk Distribution on Terrorism Policy
After scrutinizing the terrorism policies issued by the fronting carrier (Pan American), a portion of the risk of which was reinsured to the Avrahami’s CIC, the court concluded that Pan American was not a bona fide insurance company:
We won’t condemn Pan American solely for its atypical fee structure, but that structure came combined with excessive premiums, an ultralow probability of a claim ever being paid, and payments of a circular nature. We have to make a finding of fact on this. Is such an entity engaged in insurance or is it just part of a tax-reduction scheme papered to look like an entity engaged in insurance? The answer is that more likely than not, Pan American is not a bona fide insurance company.
Because Pan American wasn’t a bona fide insurance company, the terrorism policies that it issued did not qualify as “insurance”. Because they were not insurance and the Avrahamis had no other way of demonstrating risk distribution, the court ruled that none of the policies issued by the captive qualified as insurance. Because there was no real risk distribution, the deduction taken for “insurance premiums” to the captive were denied.
The Determinative Factors
The court emphasized that it was a combination of factors that lead it to conclude that Pan American was not bona fide. No single factor was determinative. Let’s consider the key factors in turn.
(Note: Unlike with Avrahami, the pricing of policies issued by most of our clients’ captives have been independently reviewed and approved by state regulators and multiple third party actuaries.)
By far the most important factor in its decision was the court’s conclusion that the terrorism premiums were “grossly excessive”. The court arrived at this conclusion by simply comparing the pricing of the terrorism policies issued by the CIC to terrorism policies available in the commercial insurance marketplace. However, as the court itself reluctantly conceded, this was an apples-to-oranges comparison at best: The coverage offered by the CIC was far more extensive than that offered by the benchmark commercial policy.
Specifically the captive insurance terrorism policy included coverage for certain chemical and biological attacks while the commercial policies against which its pricing was benchmarked excluded those risks entirely. Even more importantly, the captive insurance protected the Avrahami’s businesses against loss of revenue as a consequence of certain terrorist attacks. By contrast the “comparable” commercial policies covered only property damage arising from certain attacks.
In analyzing the reasonableness of the Avrahami’s terrorism premiums, it’s important to understand that most commercial insurance carriers won’t issue coverage for chemical or biological attack at any price, meaning that there is no premium sufficient to induce most commercial carriers to assume this risk. Consequently, a premium for coverage that insures this risk may reasonably be hundreds or thousands of times greater than premiums for policies that exclude it. Regrettably, the court barely considered this factor.
Furthermore, the odds of a given business’s income being interrupted by a terrorist attack are thousands of times greater than the odds of its property being damaged in such an attack. To give just one example among many, taking down the power grid for weeks or months is unlikely to damage the Avrahami’s jewelry store properties but (given Arizona’s scorching heat) is very likely sufficient to keep customers out of them. Depending upon the nature of the business, the value of the lost revenue could significantly exceed the value of its property.
Given the value of revenue often exceeds the value of property and that the former is thousands of times more likely to be damaged in a terrorist attack, it’s entirely reasonable for a policy that insures against loss of income and property damages (as the CIC’s policy did) to cost hundreds or thousands of times more than one that insures only against property damage (as the benchmark commercial policies did), especially if the former also covers loss of income resulting from chemical and biological attacks that are excluded by the latter.
Despite the clear logic of the above, the judge in Avrahami concluded otherwise. Noting that the CIC’s terrorism policies were 80 times more expensive than the benchmark commercial policies by one measure and between 1,125 times and 1,875 times more expensive by another, the court simply concluded that the captive premiums were “grossly excessive,” even if we adjust for such differences in in the scope of coverage.
But, of real concern, the court never actually adjusted for such differences in scope as part of its analysis, and it never cited the testimony or expert opinion of anyone else who did. Quite simply the court substituted its unqualified, subjective judgement on this matter for the professional work of an experienced actuary. The court’s opinion that the CIC terrorism policies were overpriced even after taking into account the differences in coverage was anchored in…nothing.
More troubling is the fact that the work of the actuary in question has been scrutinized innumerable times–by state insurance regulators, by independent actuaries, and even by expert actuaries who have offered their professional opinions in other court cases. One such case is the Wilson case, which is still pending before the same judge as Avrahami. None of these independent reviews have found that the actuary in question grossly overpriced policies. In some cases, quite the contrary. In fact, the expert actuary employed in the Wilson case (to review the work of the Avrahami actuary) specifically opined that the latter’s work was reasonable and was performed in accordance with accepted actuarial standards. The pending Wilson case involves the exact same terrorism policies, the same pricing and the same actuary as the Avrahami case.
The court also took issue with the fact that the Avrahami actuary did not account for industry or geography when pricing the terrorism policies, suggesting that this was evidence of unsound actuarial practice. Specifically, the rates for the terrorism policies issued by the fronting carrier (Pan American) did not vary based upon each insured’s location or industry, and this despite the fact that those businesses operating in large, metropolitan areas, or engaged in certain lines of business, are far more likely to be impacted by terrorist attack.
While I can appreciate the judge’s concern to some degree, I respectfully suggest that this criticism is at least partially anchored in an unfamiliarity with insurance. While modern insurance carriers will often get very granular in segmenting various classes of insureds into discrete rate groups (based upon geography, industry loss history, age, gender, tobacco use, health status, or any number of other factors), this was not always the case and it’s often not the case even today. Sometimes the additional segmentation simply isn’t necessary or isn’t worth the cost to the carrier. In any event, contrary to the judge’s suggestion, highly granular segmentation isn’t essential to the actuarial validity of the pricing nor to the definition of insurance.
Life insurance policies provide an easy-to-understand example of this. It’s well known that the location of an insured’s residence has a material impact on his/her life expectancy. Life expectancies in some counties can be decades longer than in others. And yet life insurance companies don’t rate policies by the location of the insured’s residence.
As an additional example, many life insurance policies (especially in an employer-employee context) are issued at “unisex” rates, meaning males and females pay the same rate even though females live far longer than males and males have a far greater variance in life expectancy. Is the pricing of unisex policies actuarially unsound because the carrier fails to charge males more than females (thereby “grossly” overcharging the latter)? Of course not.
As a final example, consider that life insurance carriers also once only offered “standard” and “nonstandard” rates. However some carriers started segmenting the “standard” class with more granularity and eventually developed a separate rate group for “preferred” (that is, better-than-standard) risks. Other carriers have gotten even more granular and now offer “select preferred” or even “ultra-preferred” rates. But not all carriers have followed this trend, at least not to the same extent. Are life insurance carriers who fail to create multiple rate classes within the “standard” category engaging in actuarially unsound practices? Obviously not.
In short, despite the judge’s criticism, there was nothing actuarially unsound about the fact that the fronting carrier failed to create separate rate classes based upon the insured’s geography or industry. The judge simply substituted his unqualified, inexperienced and subjective opinion on the pricing of the policies for the work of a professional actuary to conclude otherwise. Such is the privilege of a federal judge.
Ultralow Probability of Claim
NOTE: Unlike Avrahami, Captives managed by CIC Services, LLC and much of the industry have a proven history of paying claims.
The Avrahami court found that the premiums in question were “excessive” in part because of the “ultralow probability of a claim ever being paid”. Thus, probability of claim was the second factor that weighed heavily in the court’s decision that the fronting carrier wasn’t bona fide.
But…was the probability of claim actually “ultralow”? The court made much of the fact that the nation had never experienced a terrorist attack that would be covered under the CIC’s terrorism policies (terrorist attacks originating in large metropolitan areas, like those of September 11, 2001, were specifically excluded from coverage). This is indeed a problematic fact. Even so, anyone involved in national security will tell you that large-scale attacks against critical infrastructure located outside of key metropolitan areas (dams, bridges, the regional power grids, computer viruses, biological attacks, spread of pandemic disease, etc.) are more than a remote possibility. Such attacks have actually been planned and their effects could rapidly spread even (especially) to large metropolitan areas even though don’t originate there. Experts frequently say “it’s only a matter of time” before this type of thing happens.
Perhaps the most obvious and compelling factor suggesting that the risk of claim is far more than “ultralow” is the fact that most commercial carriers won’t insure the risk of lost revenue due to terrorist attack, and especially chemical or biological attack, for any amount of premium. If carriers believed that they could safely profit by charging meaningful premiums to insure against “ultralow” risks, they certainly would do so. But…they don’t. Why? The answer is obvious: The risk of large claims is simply too great, not ultralow. This strongly belies the court’s subjective opinion to the contrary. Even so, given that no such an attack has never happened, it’s hard to be too critical of the court’s logic on this point.
Circular Flow of Funds
The next factor that caused the court to question the legitimacy of the terrorism insurance fronting company was the “circular nature” of the money flows. Premiums were paid by entities controlled by the Avrahamis to the fronting company owned by third parties. The fronting company then paid the same amount of premium to the Avrahami’s CIC, owned by Mrs. Avrahami, in the form of reinsurance premiums, in exchange for the CIC assuming a portion of the fronting company’s overall risk, including the risk of terrorism losses by unrelated companies also insured by the fronting carrier. Because money went from an entity controlled by both Avrahamis to another entity controlled by Mrs. Avrahami, the court considered the flow of money to be “circular.”
There are obvious problems with this conclusion. First, the entities had different but overlapping ownership, so the flow was not completely circular.
Second, the payment of premium to Mrs. Avrahami’s captive insurance company came in exchange for the latter assuming real economic risk, though admittedly the court previously found that risk to be “ultralow”. Had the risk been “real” in the court’s mind, its hard to see how such payments could be deemed “circular”.
Third and most importantly, money flows in captive insurance arrangements are almost invariably “circular” by definition, in the sense that the insurer and insured are typically controlled by the same economic interests. That’s the very nature of most captive arrangements, and the court has recognized common ownership to be acceptable ever since it rejected the IRS’s “economic family” doctrine decades ago. Based on prior precedent, circular flows of the type in Avrahami simply are, by themselves, not problematic in the absence of additional complicating factors (like excessive premiums and ultralow risk of claim).
Although the fact pattern was frustratingly unique and the court made some errors of logic in its analysis (this case may very well be appealed as a result), we must nonetheless do our best to learn what lessons we can from the ruling. What might those lessons be?
First the obvious: If your captive insurance company issues “ultralow” risk policies in exchange for “grossly excessive” premiums, if those policies contain confusing or contradictory terms, if those premiums are then mostly loaned back to you or to related businesses, and if your captive has never paid a claim (and isn’t likely to ever do so), then you’ve got problems that need addressed. Fortunately, this doesn’t apply to most captives.
Second, we must recognize that it was the combination of these factors described above (excessive premiums, an ultralow probability of claims and circular flows) that led the court to the conclusion that the fronting company (Pan American) was not bona fide. Presumably, based on prior court precedent, any one of these factors alone (except for maybe excessive premiums) would be insufficient to disqualify the insurance company. For instance, we know from precedent (in the “economic family doctrine” decisions, among others) that a quasi-circular flow of funds without the presence of overpriced policies and/or an ultralow probability of loss is not problematic.
Third, because quasi-circular flows will always be present in most any captive insurance arrangement, we must therefore ensure that premiums in general, and especially premiums that count toward achieving risk distribution, are fairly priced. We must also ensure that there is at least a very real possibility of claims (and much preferably an actual history of receiving, processing and paying claims).
The importance of accurate and defensible actuarial pricing cannot be overstated. It may not be enough going forward that the actuary be experienced and sufficiently qualified (as the Avrahami one most assuredly was). Rather, the actuary may need also be articulate and eloquent (qualities for which actuaries are not particularly famous) enough to help a judge understand the complexities of the actuarial sciences. In some cases, having two or more independent, qualified actuaries price all or some policies might be advisable. In the alternative, having one actuary price certain types of risks and another price others might prove useful in some cases.
Additionally, we can conclude from Avrahami that it’s not advisable to rely on only one line of coverage to establish risk distribution. If the pricing for that single line proves defective in the judge’s mind for some reason, or if a judge simply substitutes his/her subjective judgement on a given line of coverage for the professional determinations of an experienced and qualified actuary (as the Avrahami judge did), then all is not lost (from a risk distribution perspective) if other lines of risk were sufficiently distributed and priced accurately. Therefore pooling all or virtually all risks, and not just one or two, seems highly important going forward.
Pooling all or virtually all risks has the additional benefit of also ensuring (provided the risk pool is sizable enough) that insured losses will occur, claims will be filed, and claims will be paid each year. With a sufficiently large and diversified risk pool, no judge will be able to assert, as the Avrahami judge did, that there was an “ultralow probability of a claim ever being paid.” With such pools, paying at least some claims each year is a foregone conclusion.
The Avrahami decision is based upon atypical facts. It likewise contains errors of logic by the judge that could result in an appeal. Consequently, it has little precedential value to the rest of the captive insurance industry. Even so, there are a few important things to be learned, primarily the importance of rock-solid risk distribution, defensible actuarial pricing and a reasonable expectation of claims (and preferably a history of actual claims).
After concluding that the Avrahami captive failed for lack of risk distribution, the judge went on to consider other factors that, in the judge’s opinion, caused it to fail on other grounds also. Those factors will be the subject of our next installment in this series on the Avrahami decision.