Avrahami and its Impact on 831(b) captives
Benyamin and Orna Avrahami, the owners of an 831(b) captive that insured the couple’s Arizona-based jewelry business and real estate holding companies, commenced an action in U.S. Tax Court in early 2015 in response to the issuance of notices of deficiency by the IRS totaling nearly $2.5 million in additional taxes and penalties. The IRS’s preliminary statement contained in its October 2015 answering brief emphasizes that Avrahami is the first case to address a section 831(b) transaction and that the Court’s opinion may impact other participants in the risk distribution pooling arrangement at issue. But the impact could go well beyond that, since the structure of the 831(b) in the Avrahami case is not unusual.
If the court decides in the IRS’s favor, it is possible that 831(b) captives which operate in a manner similar to that in which the Avrahamis’ captive (allegedly) operated (i.e. as a “device” to lower federal income and estate taxes) will be shut down. In disallowing the Avrahamis’ captive insurance expense deductions, the allegations that the IRS is making are not new (that the captive and pool are not true insurance arrangements as there is a lack of economic substance, risk shifting, and risk distribution) but the context of an 831(b) captive represents a new twist on an old argument.
Summary of the Facts
The Avrahamis were introduced to captive insurance by their CPA, who recommended that they consult with an estate planning attorney by the name of Neil Hiller and a New-York based captive insurance lawyer, Celia Clark. According to the IRS, Hiller and Clark worked together to set up a St. Kitts-based captive, Feedback Insurance Co., Ltd., as part of the Avrahamis’ estate plan and as a scheme to lower the Avrahami’s personal federal income taxes. Clark allegedly had established first a cross-insurance (from 2007 to 2008) and then a risk pooling arrangement (from 2009 to 2010) for many of her clients engaged in a variety of unrelated businesses. According to the IRS, the arrangements were designed for no reason other than to meet the risk distribution requirements of insurance for federal income tax purposes.
Under the arrangement developed by Clark, the captives first issued direct policies to the captives’ related insureds, and they did not share that risk with any other captive. Under the cross-insurance program, Clarks’ clients would then swap insurance premiums with each other’s captive insurance companies and operating businesses. Client A would have its business pay client’s B captive in exchange for client B’s business paying client A’s captive. As a result of the swap, each of Clark’s clients could have their businesses pay $1.2 million in premiums and claim associated insurance deductions while each of their captives would receive a $1.2 million in premiums to be excluded under section 831(b), with $360,00 or 30% derived from Clark’s other clients’ unrelated businesses.
At the heart of the cross-insurance program was terrorism policy to provide coverage in excess of the insured’s TRIA backed commercial policies. Clark allegedly told her clients “because this policy has been designed for use in a cross-insurance program, several safeguards are built in reducing the actual risk to the insurer.” For example, in the event of a loss, if an insured wished to repair damage at a cost in excess of the deductible amount, the written consent of the insurer was required. Because the policies had no deductible, the insured would have to seek written consent of the insurer for all repairs. Another safeguard put in place was that losses occurring in cities with populations over $1.5 million were excluded under the policy. Although the term “city” was not defined in the policy, according to the IRS, nearly half of the insureds were based in locations that could only be construed as “cities” by any definition – Los Angeles, New York, Houston, Cleveland, Miami, Phoenix, and Las Vegas, to name a few. These ambiguities, the IRS said, acted as safeguards to protect the insurers (Clark’s clients) from any one of her other clients submitting claims that an insurer did not want to pay.
In 2009 Clark set up a reinsurance arrangement for all of her clients through Pan-American, another St. Kitts entity, purportedly to generate more fees for herself and to create the appearance of risk distribution. Under the arrangement, the client’s captives would enter into a quota-share reinsurance agreement obligating the captive insurance company to assume a percentage of the risk from the policies that Pan-American issued to all operating businesses participating in the arrangement. Each client’s participation was at $360,000, calculated as 30% of their target premiums for each year ($1.2 million). Because Pan-American was a St. Kitts insurer, there was no reserve requirement. St. Kitts only required reserves to be maintained on net premiums, and Pan-American ceded 100% of its premiums to Clark’s clients. However in the interests of building a comfort level for her clients, it did maintain 2.5% of premium payments as loss reserves during the policy year, at the end of which the 2.5% was distributed back to the clients.
Not insurance in its Commonly Accepted Sense
In its brief, the IRS has asserted that the Feedback and Pan-American arrangements are not insurance in the commonly accepted sense. The Avrahamis learned about captive insurance through their CPA, not an insurance broker. Premiums were not actuarily determined but rather suggested by Hiller (the estate planner) who in correspondence said that 20% of the premiums should come from American Findings and the remaining 80% from real estate operations, and by Clark who allegedly communicated to the actuary what the premiums should be. Feedback did not prepare financial statements or hold meetings of its board of directors. Feedback allegedly did not have any employees or contract with experienced insurance professionals to conduct insurance functions such as underwriting, setting of premiums and preserves, investment management, and claims administration.
In fact, it was Clark who took on the role of a quasi-management company by drafting the policies and generating invoices, though she did not perform any underwriting services for Feedback, manage investments or set reserves. Feedback paid claimed losses without establishing the validity of those claims. Feedback and Pan-American were not regulated by St. Kitts and did not follow regulations. They solicited business in Arizona, provided insurance to Arizona residents, and therefore should have followed Arizona regulations, which they did not. The policies were confusing and sloppily drafted, combining features of both an occurrence and claims made-policy. Furthermore, Feedback’s policies did not insure insurable risks. For example, Feedback’s loss of key employee policy insured against the departure of Mr. and Mrs. Avrahami. Because the Avrahamis are shareholders of American Findings, they protected themselves through the policy from voluntarily departing. The IRS also points out that the policies under the Pan America arrangement were not arms’ length because of the “safeguards” put in place to ensure that no insurer would be required to pay a claim in the event of a loss.
No Risk Distribution
As those in the insurance world know, in order for an arrangement to constitute insurance for federal income tax purposes, the risks must be distributed among a sufficient number insureds so that the statistical law of large numbers permits the insurer to smooth out its losses. Exactly how many insureds are required has been debated. For many years, the rule of thumb was 12 subsidiary companies, with no one insurer representing less than 5% nor more than 15% of the total risk (see Revenue Ruling 2002-90). However, in 2014, the Tax Court call these rules into question when decided in favor of the taxpayers in the Rent-A-Center and Securitas cases. The taxpayers in both cases won the risk distribution argument in large part because the captives insured a sufficient number of “exposure units” —- stores, employees, cars, etc.
Both Avrahami and the IRS have discussed “exposure units” in making their arguments. Avrahami has attempted to show that there is a sufficient number of exposure units among its four subsidiary insureds by virtue of having different types of policies, different parties, including a (rental) management company, tenants, and tenants customers. The IRS in its answer brief points out that in Rent-A-Center there were over 2,000 stores which operated in 50 states, had over 14,000 vehicles and operated over 7,000 vehicles. The IRS also points out that in Securitas the captive received premiums from over 25 separate entities in one year, 45 separate entities in another year, and its insured had hundreds of thousands of employees in over 20 countries and operated 2,250 vehicles.
The IRS states that “those large pools of statistically independent risk exposures—regardless of how Petitioners define risk exposure—are not present here.” So it appears that, in the IRS’s eyes at least, a taxpayer must have many thousands of employees or vehicles, and double digit entities, in order to win the “exposure unit” argument. The IRS pointed out that the Avrahami real estate entities had no employees, engaged the same property management company, and each held one retail shopping center or piece of land in Arizona. American Findings (the jewelry store) had 35 employees, owned no real estate, and operated three retail jewelry stores in space that is leased in the Phoenix area. The IRS then concludes “Because Feedback failed to insure 35 affiliated entities or otherwise pool a sufficient number of statistically independent risk exposures amongst its affiliates entities, it did not distribute risk through its affiliated entities.”
The IRS makes the same argument with respect to the Pan-American arrangement, pointing out that the pool only had 102 insureds in 2009 and 142 in 2010. Moreover, the risks that were insured under the pool were not homogenous.
The IRS also cites the Harper case in asserting that there was no risk distribution with the Pan-American pool. Harper, decided in 1991, has long stood for the proposition that risk distribution has been met where there is 30% unrelated business. In response to the Avrahamis’ assertion that Pan-American was a true insurance arrangement because 30% of the risk insured by its captive through the arrangement was third-party, the IRS states that the 30% rule of Harper does not apply, because the risks insured in that case were all homogenous, and the risks insured through Pan-American were diverse. The IRS goes further to say that all of its revenue rulings addressing risk distribution assume that pooled risks are homogenous. It will be interesting see how the court rules on the issue of risk homogeneity, as this issue has not (until now) been at the forefront of captive litigation.
No Risk Shifting
In its answer brief, the IRS also asserts that the Avrahami entities did not shift risk to Feedback because Feedback was not financially capable of meeting its obligations, and the entities failed to submit claims for losses that appears to be covered by the policies. According to the Answer Brief, Feedback never had more than $1.5 million in cash or cash equivalents on hand during the years at issue, and the promissory notes totaling over $2 million on loans made by Feedback to affiliate entities were not secured. Despite this thin capitalization, Feedback’s maximum total exposure under the policies was allegedly as high as $26 million in 2010. Citing again to Rent-A-Center, the IRS claims that “although the captive in Rent-A-Center was thinly capitalized and initially failed to satisfy Bermuda’s minimum solvency and liquidity provisions, to compensate, the parent company issued a parental guarantee to cover the obligations of the captive.” Ironically, the IRS has historically disqualified captive insurance arrangements where a parental guarantee exists. It appears the IRS has abandoned this approach, at least within the context of 831(b)s.
Conclusion
If the IRS wins Avrahami, it could mean that 831(b)s and their managers will need to revisit their insurance programs to ensure that 831(b) captives do not resemble Feedback and Pan-American. This would entail confirming the that policies are written by insurance specialists, premiums are actuarially determined, and embody all of the accoutrements of a true insurance policy. What is more challenging will be meeting the requirements of risk shifting and risk distribution. Smaller insurance companies that have made the election under 831(b) will not be able to rely on having three of four subsidiary insureds to meeting risk distribution. Nor will they be able to meet risk distribution by taking on 30% unrelated third party risk, unless the risk is homogenous. Risk shifting could also become more challenging with the IRS’s renewed focus on capitalization. Hopefully, as the IRS has requested, the Tax Court will reach a decision soon so that existing 831(b)s can plan accordingly.
