Breaking Down the Avrahami Decision

After months of anticipation, the United States Tax Court has ruled in favor of the IRS in the case against Benyamin and Orna Avrahami. While the outcome is certainly disappointing to many in the captive insurance industry, it is likely not a surprise as the case turns on the fundamental concepts that might define “insurance,” if ever a true definition might be penned.


Incorporated in 2007, Feedback Insurance company, Ltd. (Feedback) was set up as a captive insurance company domiciled in St. Kitts. It filed an election under IRC Section 953(d) to be treated as a domestic corporation for federal income tax purposes, and subsequently filed an election to be taxed as a small insurance company under IRC Section 831(b).

Feedback issued various policies to passthrough entities owned by one or both Avrahamis and/or their three adult children. The policies were supposed to provide coverage for a) legal expenses arising out of administrative or disciplinary action; b) business liability related to construction defects not otherwise covered by other existing commercial policies; c) loss of business income due to reputational damage or new competition; d) fraudulent or dishonest acts of employees; e) legal expenses incurred for any reason by American Findings Corporation (American Findings), specifically; f) loss of business income due to the permanent or temporary departure of key employees (namely, Mr. and Mrs. Avrahami); and, g) tax indemnity due from an “adverse resolution” of a position taken on a tax return.

The Avrahamis engaged an attorney, Celia Clark, to facilitate the evolution of Feedback and its insurance programs. Ms. Clark was fully aware that the establishment of insurance in the federally accepted sense required sufficient risk transfer as well as risk distribution. To provide additional risk distribution to her captive clients, Ms. Clark created a Terrorism Risk Insurance Pool (TRIP) called Pan American. By its terms, Pan American would “reimburse policyholders for losses of ‘property’ and ‘expenses’ resulting directly from an ‘act of terrorism’ occurring during the Indemnity Period” but only for those occurring in a city with 1.5 million residents or less. For its services, Pan American retained a portion of the flat fee of $5,000 that Ms. Clark received from each of the pool participants.

As argued by the Avrahamis, Feedback’s participation in this risk pool is what provided third-party risk and, therefore, risk distribution. Each year, American Findings paid $360,000 to Pan American for terrorism coverage (policy limits of $5.5 million in 2009 and $5.1 million in 2010). Each year, Feedback received $360,000 from Pan American to reinsure the direct writings of the pool. This participation was strategically structured to offer third party coverage of 30% of the captive’s total premiums and thereby meet the industry’s unspoken safe harbor for satisfying risk distribution based on the Harper Group[1] case.

It seems no coincidence that this $360,000 premium was exactly 30% of $1,200,000, the premium threshold necessary for a captive to qualify for the election under 831(b) to avoid tax on its underwriting income. Further, Pan American continued to assess the same $360,000 premium each year even though the policy limits fluctuated.

At this point, it is helpful to consider that even though American Findings participated in the TRIP, it also maintained add-on terrorism coverage through the commercial market with a policy limit of $2 million and an average annual premium of $1,550 based on the two years under exam. In fact, none of the entities owned by the Avrahamis reduced any of their commercial insurance coverages after entering into contracts with Feedback or Pan American. Thus, their overall insurance expense increased, which is often contrary to why businesses establish captives in the first place.

Nevertheless, the evolution of the Avrahamis’ insurance programs continued with Ms. Clark’s engagement of an actuary for purposes of setting premiums for the TRIP policies as well as those specifically issued by Feedback. She instructed the actuary that the premiums for all other policies issued by Feedback were to be no greater than $840,000 each year. Again, this appears to be strategic since $840,000 is the remaining premium available to be earned by a captive participating in the pool at $360,000 annually and still wanting to maintain its 831(b) status.

The use of an actuary is typically a good fact when trying to determine the insurance status for federal tax purposes of a captive or its programs. “No one thinks this process lacks all subjectivity, but the work of an actuary must be reproducible and explainable to other actuaries[2].” As outlined in more detail in the 105-page Tax Court opinion, the evidence suggested the application of more subjective, experiential manipulation to setting the premium rather than any sort of analysis which could be explained, let alone duplicated, by another actuary. Inconsistent methodologies were applied to similar contracts, or to the same renewal contract in subsequent years.

Perhaps it was not entirely the fault of the actuary. One insuring agreement stated that Feedback would agree to pay for “legal expenses incurred by the insured during the Policy Period” and went on to define the Policy Period as “[e]vents occurring and reported from and after 12:01 a.m. December 15, 2009 and prior to 12:01 a.m. December 15, 2010.” When the policy language states that coverage is offered on a claims-made AND an occurrence basis, it’s no wonder there was a struggle to appropriately price the product.

Pricing was not the only issue. In a January 2013 letter from the IRS to the Avrahamis, it was noted that Feedback had never paid any claims even though it had received premiums from inception through 2010 of approximately $3.9 million. This seemed to trigger concern by the taxpayer because within two months of the IRS’ letter, the Avrahamis’ entities insured by Feedback began submitting claims on prior policy periods. Unfortunately, this attempt to mitigate the situation only exacerbated it when it was pointed out that some of the claims were filed late, yet still approved for payment.

Feedback’s investment strategy was also called into question. During 2009 and 2010 – the years under IRS examination – Feedback received just under $1.2 million each year. The defendants argue the accumulated surplus was used to invest in land, and that such investments are not uncommon holdings for captives.

The plaintiffs argued that greater consideration should be given to the mechanism by which this was accomplished: The Avrahamis’ created a partnership in 2007 called Belly Button Center, LLC (“Belly Button”). Belly Button was owned equally by the Avrahamis’ three children – each of whom testified to having no knowledge of Belly Button, their supposed ownership or of the LLC’s activity. Regardless, Belly Button purchased land for $1.96 million using $1.2 million in cash from Mr. Avrahami and a note payable to the sellers for the remainder. Using its surplus, Feedback loaned money to Belly Button which would, in turn, use the cash from Feedback to repay Mr. Avrahami, as well as the principal and interest on the land purchase loan. This happened in March of 2008, March of 2010 and again in December of 2010. The December 2010 had transaction paperwork created to match the previous transactions although the cash, this time, went directly from Feedback to Mrs. Avrahami. Each of the loans carried simple interest and a ten-year repayment period. These transactions were not disclosed to the regulators of St. Kitts until September 2014.

Insurance companies are rated by their liquidity and therefore, their ability to timely pay claims should claims occur. Neither captives nor commercial insurance companies would (be allowed to) tie up a significant portion of their funds in non-liquid investments such as real estate and mortgage loans. Yet this was exactly the situation Feedback had created.


This article began with the initial assertion that this case turns on the fundamental concepts that might define “insurance”. The absence of any one of these is enough to suggest that an entity would not qualify as an insurance company. The Courts repeatedly come back to these concepts which are summarized as follows:

  1. Insurance historically and commonly involves risk shifting
  2. Insurance involves risk distribution
  3. Insurance will involve an actual insurance risk
  4. Insurance will meet commonly accepted notions of insurance

In this case, the Tax Court focused first on the presence of risk distribution. While the parties disagree on the number of entities or risks that are covered by Feedback policies, expert witnesses for both sides argued that “more risks” are necessary to achieve a large enough pool than exists in the Avrahami fact pattern. Looking beyond the number of insured entities, the Court went on to stress the importance of the number of independent risk exposures. With only seven types of policies issued by Feedback to three or four affiliated passthrough entities covering three jewelry stores, two key employees, 35 employees and three commercial real estate properties, the Court found that this was not sufficient to meet the guidance of previously tried arguments[3] even in the context of a micro-captive program.

Let us not forget Feedback’s reliance on Pan American to provide 30% unrelated party coverage. Ms. Clark had explained to her clients that to be treated as an insurance company for tax purposes, risk distribution was required. Therefore, she created and marketed Pan American as the mechanism to strengthen her clients’ positions for this tax treatment. The TRIP premium was a one-size-fits-all premium regardless of the pool participant, and the premium was set at 30% of the IRC Section 831(b) threshold ($1.2 million in premiums). In contrast, Feedback also maintained separate terrorism coverage for which it paid premiums at a ratio to coverage of approximately .081% while the ratio for similar coverage with Pan American was nearly 7%. For every dollar that came into the TRIP program, an equal amount was distributed to the affiliated captive under a reinsurance arrangement, leaving little capital and surplus in Pan American to pay claims should it be faced with any.

These facts suggested to the Court that Pan American was not created for legitimate nontax reasons. Insurance companies are required to have actuarially determined premiums, not merely a premium advised by an actuary who was instructed to target a specific number. The flow of funds into Pan American from a Feedback affiliate and back to Feedback suggested to the Court a circular flow of funds within the Avrahami organization that could otherwise have been accomplished by direct transfer but for the attempt at the creation of insurance. Most significantly, since it was a dollar-in-dollar-out program, if a covered claim came in, Pan American would have to obtain the necessary cash from each of the reinsurers. Thus, the parties would end up paying for their own claims based on this structure. Based on the arguments presented, the Tax Court found Pan American to not be a bona fide insurance company, and thus the defendant’s reliance on unrelated party activity was moot.

Without sufficient risk distribution, the argument is over. Although the Court goes on to assess the remaining aspects of the case, it ultimately becomes irrelevant that many of the risks insured by Feedback were determined to not be insurable risks, but rather simply business risks accepted by any entrepreneur in the course of doing business. Any arguments as to the legitimacy of the real estate sale between Mr. Avrahami and Belly Button, and the affiliate loans no longer matters. The Court also found that Feedback, itself, was also not operating as a true insurance company when it consistently failed to notify and/or obtain approval from its regulators as to certain transactions. Nor was it necessarily acting as a true insurance company when it entered into contracts with terms or premiums that deemed unreasonable and not set pursuant to an arm’s-length basis.

What does matter, though, is that a captive insurance company challenged the IRS and the Tax Court, and precedent has now been set. The cumulative fact pattern was found to be unfavorable to the Avrahami case, but individual pieces of the fact pattern may very well be extant in varying context in other captive insurance companies of legitimate operation around the rest of the world. The Avrahami case is the proverbial bad apple in the barrel that will likely only serve to increase legislative scrutiny and IRS distrust of all other captive programs.

Next Steps

Whether you find similarities between this case and your captive program or simply want to ensure that too many similarities are not found, now is a good time to revisit your operating plan and have a conversation with your tax advisors. No clear-cut definition of insurance exists, so context is key and consistent, ongoing documentation is critical. If you need a partner for your captive insurance program, please contact us.


[1] Harper Grp. V Commissioner, 96 T.C. 50, 59-60 (1991)

[2] Avrahama v Commissioner, 149 T.C. 14

[3] Humana Inc. v. Commissioner, 881 F.2d 247 (6th Cir. 1989); Rent-A-Center v. Commissioner, 142 T.C. 1; R.V.I. Guaranty Co v. Commissioner, 145 T.C. No. 9 (Sept 21, 2015)

Johnson Lambert
Johnson Lambert | Author