Starting with the 2015 iteration of its annual "Dirty Dozen" list, the IRS began an unprecedented crack down on suspected tax shelters using captive insurance companies. In the time since, Congress has amended the relevant statutory provision to preclude a particular captive use that avoids estate and gift taxes, but left unaddressed a plethora of other "off label" and arguably abusive uses. Most recently, the IRS issued Notice 2016-66 officially classifying certain captive uses as "transactions of interest." Such classification requires taxpayers to disclose their participation in a transaction similar to the one described in the Notice. Material advisors may also have disclosure and list maintenance obligations. Failure to disclose a relevant transaction exposes the taxpayer to penalties under Code Sec. 6707A, though other penalties may also apply upon audit.

But how did we get here? This article provides a primer on (1) what a captive insurance company is; (2) how the Code has traditionally provided favorable treatment to captives; (3) the seminal standards applied by the IRS and courts to test whether a captive qualified for such treatment; (4) how taxpayers used that version of the law for tax avoidance; (4) IRS and Congressional responses to curb perceived abuses; and (5) things to consider going forward.

What Is a "Captive" Insurer and Why Form One?

"Captive" insurance companies are ones formed as subsidiaries to insure the risks of and affiliated with their corporate parents. Their usage spans from behemoth corporate conglomerates with hundreds of subsidiaries to small, independent farmers, doctors, and other closely-held businesses. Among other benefits, captives allow businesses to insure risks that they otherwise would not, or could not, insure. In the latter case, some businesses face perils for which commercial carriers do not offer coverage products. Without captives, these businesses either would have to forego coverage or self-insure. If the prospective insured believes the risk is significant enough to warrant coverage, then foregoing that coverage is unacceptable. And though better than foregoing risk mitigation altogether, self-insurance technically is not "insurance" at all, and that option also deprives the insured of legitimate tax benefits. Even where commercial insurance is available, though, the coverage may be prohibitively expensive. Insuring with a captive can be more affordable for those businesses, and because a captive's coverage is customizable, the insured has ability to negotiate unique coverage products.

Additionally, by insuring with a captive, an insured pays its premiums as usual, but instead of those amounts going to enrich an unrelated party, they instead stay within the same corporate structure and, in that regard, the insured gets the best of both worlds. Captives also allow their owners to access the reinsurance market. Finally, and most relevant, there are tax benefits to using captives, including deductions of premiums paid by the insured and of unearned premiums received by the captive. And those captives that qualify for and make an election under Code Sec. 831(b) are permitted to pay tax only on their annual investment returns, leaving their net written premium income untaxed. Until recently, a captive could earn no more than $1.2 million in annual net written premium in order to qualify for the election; hence the moniker "micro" captive. This 831(b) election is the particular benefit that has led to much of the perceived abuse.

"Insurance" for Tax Purposes

Although micro captives are uniquely small vis-à-vis their larger cousins formed by corporate conglomerates, the IRS applies the same rules to both. The central question for a captive, regardless of size, is whether it provides actual "insurance." If it does, then the IRS generally will honor the consequent tax benefits to both. If, however, it concludes that no real "insurance" exists, the IRS instead will disqualify the captive and assess tax liabilities against the captive, the insured, and possibly the individuals who own the related business entities. As shown below, however, it is significantly easier for larger captives to satisfy these standards, meaning that the IRS's "one-size-fits-all" approach has a regressive and disproportionate impact on micro captives.

The Tax Code does not define "insurance." The governing analysis, therefore, traces its origins to Supreme Court precedent. In Helvering v. Le Gierse, the Court explained that "[h]istorically and commonly insurance involves [both] risk-shifting and risk-distributing ... That these elements of risk-shifting and risk-distributing are essential to a[n] ... insurance contract is agreed by courts and commentators."1 Although modern-day standards have evolved and added additional nuance, it remains true that risk-shifting and distribution are the primary hurdles to persuading the IRS that actual insurance is present.

Risk-shifting focuses on whether the insured "transfers some or all of the financial consequences of the potential loss to the insurer, such that a loss by the insured does not affect the insured because the loss is offset by the insurance payment."2 Risk-distribution, on the other hand, focuses on the insurer. When present, risk-distribution "allows the insurer to reduce the possibility that a single costly claim will exceed the amount taken in as premiums and set aside for the payment of such a claim. By assuming numerous relatively small, independent risks that occur randomly over time, the insurer smooths out losses to match more closely its receipt of premiums."3

For decades, the IRS took the position that risk-shifting and distribution could not occur within the same economic "family." The simplest demonstration is a corporate parent with a captive as its sole subsidiary. In that context, according to the doctrine, neither risk-shifting nor distribution takes place where that parent purchases insurance from its captive. Because the parent is the 100% beneficial owner of the captive, it would suffer an offsetting financial loss if the captive ever paid a claim. Thus, the parent never really "shifts" the risk because it will continue to bear the economic burden in the event of a loss. The economic family doctrine also held that risk distribution lacks in that context, since the captive would have assumed risks solely from its parent.

The economic family doctrine was conceptually defective in that it ignored the legitimate, primary motivation taxpayers had in forming captives in the first place—to insure risks of the captive's affiliates. In this way, the IRS punished captives for doing precisely what they were designed to do. Though the economic family doctrine became antiquated in the face of more sophisticated captive arrangements, the IRS was slow to abandon its position. For some time the IRS continued to apply the doctrine to disqualify insurance transactions between brother-sister business entities. But despite—and perhaps because of— its dogmatic adherence to the economic family theory, the IRS began suffering a series of losses in the courts.

Seminal precedent condoning brother-sister insurance took place in the 1980s, but not until the new millennium did the IRS formally abandon the economic family doctrine. This watershed moment reflected that, although easy for the IRS to apply, the "cookie cutter" nature of the economic family doctrine did not account for the nuances that separated different captive structures. But the IRS's shift created a guidance vacuum. To satiate taxpayers' understandable yearning for predictability, the IRS created a series of "safe harbors" that, if met, would allow captive transactions within the same economic family to qualify as "insurance."

Two separate Revenue Rulings in particular created the most important safe harbors, each of which represent a different path to achieving risk-shifting and distribution. Rev. Rul. 2002-90 addressed insurance transactions between subsidiaries of a common parent. The case involved a domestic holding company that owned all of the stock of its 12 domestic subsidiaries, each of which had "a significant volume of homogeneous risks." The holding company formed an adequately capitalized, licensed insurance subsidiary to insure directly the professional liability risks of each of the 12 subsidiaries, no one of which amounted to more than 15% or less than 5% of the captive's overall assumed risk. The IRS held that, on those specific facts, there was sufficient risk-shifting and distribution.

Although the "bright line" takeaways from Rev. Rul. 2002-90 were the number of subsidiaries and the proportion of risk ceded by each, the IRS equally emphasized that "[i]n all respects, the parties conduct[ed] themselves in a manner consistent with the standards applicable to an insurance arrangement between two unrelated parties." The IRS noted with approval that (1) there were no financial guarantees between the entities; (2) the captive charged arm's-length premiums; (3) the captive was fully licensed; (4) the captive faced actual hazard; and (5) the parent formed the captive for "a valid non-tax business purpose." Thus, while the safe harbor undoubtedly required 12 entities, the IRS emphasized as just as vital that the captive pooled the premiums "such that a loss by one operating subsidiary is borne, in substantial part, by the premiums paid by others." In other words, it not only looked like insurance but it operated like it as well.

Around the same time, Rev. Rul. 2002-89 set the standard for those circumstances where parent-subsidiary arrangements can qualify as insurance. The IRS considered two similar, though vitally distinct sets of facts. In both scenarios, a domestic corporation formed a wholly-owned subsidiary to insure the parent's professional liability risks, either directly or as a reinsurer. In the first scenario, the premiums derived from the parent constituted 90% of the total premiums earned in a year. In turn, the captive insured 90% of the corporation's total risks. The IRS held that this arrangement did not constitute insurance because it lacked adequate risk-shifting and distribution, and therefore, the premiums were not deductible. In the second scenario, however, the captive derived less than 50% of its annual premiums from its parent. Likewise, the liability coverage the subsidiary provided the corporation was less than 50% of the risks assumed by the captive. The IRS held that, under those facts, there was adequate risk-shifting and distribution.

As noted, though, risk transfer and distribution, while necessary, may not be sufficient. A recent Chief Counsel Memorandum, for example, concluded that an otherwise sound captive arrangement did not qualify as "insurance" because the policies covered "investment risk" rather than "economic loss."4 The IRS explained that "[n]ot all contracts that transfer risk are insurance policies even though the primary purpose of the contract is to transfer risk. For example, a contract that protects against the failure to achieve a desired investment return protects "MICRO" CAPTIVE INSURANCE COMPANIES against investment risk, not insurance risk." Citing 1950 precedent from the Second Circuit, the IRS continued that "[i]nsurance risk requires a fortuitous event or hazard and not a mere timing or investment risk. A fortuitous event (such as a fire or accident) is at the heart of any contract of insurance." Thus, even if a particular captive arrangement complies with all other requirements, careful planners must ensure that the insured risk is of the variety deemed acceptable by the IRS. It is unclear exactly where the IRS draws this line.

Reconciling Governing Standards with Micro Captives

A fair question is whether the IRS adequately considered micro captives when creating the safe harbors. No doubt the captive of a Fortune 500 company can act like a large, sophisticated, and independent insurance company, with hired actuaries, underwriters, adjusters, and the like. Such entities have the resources to plan their operations in exquisite detail, and their captives can easily write sufficient coverage volume to meet the IRS's exacting standards for risk-shifting and distribution. Moreover, the associated conglomerates are oftentimes large and diverse enough that the captives can show risk-shifting and distribution without having to write a single unaffiliated policy.

Micro captives, conversely, face a much more difficult time qualifying for the safe harbors. After all, micro captives typically have parents that are closely-held, small businesses. Such businesses often have limited resources, narrow expertise, and lack the affiliated risk to allow the captive to write coverage solely within the same economic family. Consider, for example, a small, closely-held construction company. While such owners undoubtedly have construction and general business expertise, they likely only have superficial familiarity with the insurance industry. Even assuming that the owners are not dissuaded by the regulatory trappings of forming and operating an insurance company (and that they are able to do so without unwittingly attracting unwanted IRS attention), how is that company going to show risk-shifting and distribution? Most often such closely-held businesses do not have 12 subsidiaries, meaning that if the construction company wishes to enjoy the predictability created by the IRS's safe harbors, the newly-minted captive would have to assume unrelated risk. But it is not as simple as setting up shop and marketing insurance products to the public; such activities are subject to even stricter regulation. And if construction company wants to take advantage of the Code Sec. 831(b) election, it must incur all of these compliance costs and then limit itself to the relatively small annual premium cap. At that point, is it even worth it?

This is where the IRS arguably is reaping what it sowed. Through Congress' inclusion of the Section 831(b) election in the Code, it made the policy decision to incentivize the formation of micro captive insurers. At the same time, however, the IRS designed the safe harbors solely with large captives in mind and never bothered to develop separate, realistic standards for micro captives. Because the "one-size-fits-all" approach contemplates larger companies, it has a regressive impact on small insurers and has created a regulatory environment that invites creative thinking if micro captives are to fit themselves within the IRS's exacting safe harbors.

Generating Risk-Shifting and Distribution Through Captive Managers

No doubt these daunting hurdles have caused many taxpayers to forego formation of much-needed captives. There remain enough taxpayers with a strong enough need, however, to create a marketable demand for third-party expertise in forming and operating successful, tax compliant captives. Meeting this demand are so-called captive managers and "turn-key" operations. Such services attract small businesses because they allow those businesses to delegate the hassle of creating their captives and navigating the regulatory framework.

Captive managers, in exchange for a periodic service fee, actively operate captive insurance companies on behalf of the parent. This arrangement allows a small business the benefit of owning a captive without having to hire a staff or otherwise deal with the burdens of operating an insurance company. Capable managers can perform those tasks efficiently and reliably due to economies of scale. A group of closely-held companies may also coordinate the hiring of a single manager to operate a so-called "group captive," which insures and is owned by that unaffiliated group of small businesses.5 "Turn-key" services are similar to and can overlap with managers, but such providers essentially offer "off-the-shelf" captives for purchase. Just as an affordable, boilerplate will offer an easy estate planning option to a low net-worth individual, "turn-key" services are attractive to small businesses.

But just as surely as there are many reputable managers, there also are those who market abusive uses of captives that have only the outward appearance of compliance with IRS standards. Thus, those contemplating the formation of a captive must be wary of those holding themselves out as reputable managers and touting their ability to satisfy the risk-shifting and distribution requirements. Many legitimate managers employ risk pools to generate risk shifting and distribution, but bona fide use of pools requires captive owners to assume the risk of actually having to pay a third-party claim. Managers are able to achieve these goals more easily than could a closely-held business because managers have access to perhaps hundreds of insureds. Promoters, on the other hand, have conceived of ways to structure such pools so as to remove all likelihood of having to pay third-party claims. Outward appearances, therefore, can be deceiving and a close inspection is paramount if one is shopping for a captive structure that will withstand audit. If the IRS investigates, it will inquire whether these arrangements have substance, as opposed to the mere appearance of substance.

No doubt some well-intentioned taxpayers (and managers for that matter) will find themselves lumped together by the IRS with those of lesser scruples. But one reliable way to identify a promoter is to analyze its marketing materials. If those materials advertise primary uses of captives unrelated to insurance, it is best to keep shopping. While acceptable to consider some off label uses as "fringe" benefits, a taxpayer had better be prepared to prove that its primary purpose was insurance.

Promoted Schemes and Abuses

Though Congress recently revised the statute in order to curb this particular abuse, the IRS is particularly bothered by taxpayers who use captives as a means to circumvent gift and estate taxes. These taxpayers transfer "premiums" to the captive without taxation and place title of the captive's stock in the names of the intended gift recipients. But this is by no means the only way captives can allow for abuse. Another example has the captive purchase a life insurance policy on the parent's owner, which effectively allows that owner to deduct otherwise non-deductible life insurance premiums. Other taxpayers may simply choose to park their untaxed money in the captive and use those assets for investing. Other captives engage in loan backs, where the parent pays premiums to the captive, both sides take their deductions, and the captive "loans" the money back to the parent. At the conclusion of the transaction, the parent retains the beneficial use of the money it paid as premiums but avoids paying tax on those amounts. In lieu of a "loan," some owners tap into the captive's reserves by having the captive invest much or all of their premium income in the parent, affiliates, or in other locations that benefit the parent.

In each case, though, tax avoidance—not insurance— is the goal, so any significant claims expenses would be counterproductive. In order to reduce or eliminate that risk, "piggybank" captives might assume implausible risks, especially relative to the premium. For example, few businesses can reasonably claim to form a captive to insure against terrorism risks, especially when priced expensively. Similarly, a company in the Midwest typically does not view hurricane damage as a reasonably foreseeable peril. Many use such implausible risks to make a captive arrangement appear motivated by insurance, where tax savings are the real prize.

Promoters also design their captive arrangements to have only the appearance of risk-transferring and distribution. For example, the parent might purchase a policy on the open market but then have that insurer cede the entire risk to the captive in exchange for a bulk of the premium. In other cases, the promoters might create supposed risk pools. The captives appear to reinsure part of the pool's intermingled exposure, thereby acquiring third-party risk. To avoid the prospect of paying third-party claims, the promoter might have parent corporations indemnify any claims made to the pool. Other common tactics include having excessively high deductibles, which make it unlikely that claims will ever trigger the pool's liability. There are many ways to achieve it, but the desired outcome is for the parents to retain, rather than shift, the risk of loss, though to make it look otherwise. At bottom, the IRS will invalidate a captive insurer if it lacks economic substance.

The captive itself also must have facial legitimacy. This means that the captive should be adequately capitalized and the formation and policy papers should suggest that the captive put care into protecting itself. Similarly, the captive should independently underwrite all risks and develop a premium defendable under arm's-length market conditions. If the insured fails to pay premiums without consequences, that too will suggest inadequate independence, as will a one-sided claims history. No one fact will be determinative, but the goal should be for it to appear to anyone examining closely enough that the captive is arm's-length from its parent.

Congressional Revision to Code Sec. 831(b)

On December 18, 2015, President Obama signed into law the Protecting Americans from Tax Hikes Act of 2015 (the "Act"). Congress included in the Act revisions to the Code Sec. 831(b) election. The amendments, which are comprised of dense—if not impenetrable—prose, are far from a panacea. They leave unanswered certain questions and unaddressed a host of widely-known abuses of "MICRO" CAPTIVE INSURANCE COMPANIES micro captives. Though the IRS may issue interpretative regulations to address some of the questions raised by the amendments, it is nevertheless evident now that a statutory solution to abuses outside of the estate planning realm is not forthcoming.

The amendments, which became effective January 1, 2017, raised the net written premium cap to $2.2 million and added an inflation adjuster to allow for the continued increase. In this regard, the amendments threaten to make the abuse of micro captives all the more tempting by increasing the potential to shelter income. Congress did, however, add a "diversification" requirement that makes it more difficult to qualify for the 831(b) election. There are two ways a captive can meet the diversification requirement. The first is relatively straightforward: "An insurance company meets the [diversification] requirement[] if no more than 20 percent of the net written premiums ... of such company for the taxable year is attributable to any one policy holder." This diversification test borrows from the facts in Rev. Rul. 2002-89, which created one of the first "safe harbors" for meeting the widely-accepted definition of "insurance" for tax purposes, discussed above.

By way of reminder, Rev. Rul. 2002-89 held that where a captive's premiums from its parent reflected no more than 50% of the company's total assumed risks, there was adequate risk shifting to qualify as "insurance," but when that number increased to 90%, no real risk shifting took place. Thus, the first diversification test under the revisions to Code Sec. 831(b) 50% safe harbor under Rev. Rul. 2002-89. As of January 1, 2017, therefore, any micro captive that previously met all published IRS standards by insuring not more than 50% of its risks of its parent will have to reduce that exposure to 20% if it wishes to qualify for the Code Sec. 831(b) election. Note, however, that this 20% standard merely borrows from, but does not supplant, the insurance test. Thus, if a micro captive continues to insure between 20% and 50% of its parent's risk, then while it may not qualify for the Code Sec. 831(b) election, the IRS still should treat it as "insurance." This distinction can be vital for the deductibility of the parent's payment of premiums. The second diversification test is more complicated and ambiguous. The following is the statutory language, in relevant part and with emphasis supplied:

An insurance company meets the [diversification] requirement[] if ... no person who holds ... an interest in such insurance company is a specified holder who holds ... aggregate interests in such insurance company which constitute a percentage of the entire interests in such insurance company which is more than a de minimis percentage higher than the percentage of interests in the specified assets with respect to such insurance company held ... by such specified holder.

The statute defines the emphasized terms. "Specified assets" are "the trades or businesses, rights, or assets with respect to which the net written premiums ... of such insurance company are paid." In the prototypical example of a parent-subsidiary captive arrangement, the "specified assets" presumably are those belonging to the parent and transferred to the captive as premium. A "specified holder," by contrast, is "any individual who holds ... an interest in such insurance company and who is a spouse or lineal descendant ... of an individual who holds an interest in the specified assets." Finally, "deminimis," for these purposes, means "2 percentage points or less."

This diversification test is designed to combat the known estate planning abuse where, for example, a husband and wife own equal shares of a family business. If the husband could place a 100% ownership interest of the captive in the wife's name, then the husband could effectively gift portions of his ownership interest in business in the form of tax free "premium" payments to the captive. In that example, the "specified assets" are those in the business that the husband wants to transfer to his wife, but on which he wishes to avoid taxation on the transfer. The wife is the so-called "specified holder." In order to meet the second diversification test, therefore, the wife's interest in the captive cannot exceed by more than 2% her interest in the business. Thus, because the wife has a 50% interest in the business from which the "premiums" are paid, the captive will fail diversification if she owns any more than 52% of the captive. With that example, it is easy to see how Congress believes that this rule will stymie further use of captives as a means to transfer ownership of assets without taxation.

Yet the statute does not address what happens where there is more than one specified holder and more than one pool of specified assets. For example, assume that a captive's premiums come from two separate bank accounts, the first held jointly by a husband and wife, and the second held jointly by the same husband and his son. Both mother and son, who each have a 50% interest in their respective accounts used to pay some of the premium, are "specified holders." But to what pool of assets would the IRS look to determine the maximum interest such wife could hold in the captive? If only to the account in which she has a 50% interest, then she can safely hold up to 52% of the captive without threatening diversification. But if the amendment instead aggregates the assets in both accounts, then she can own no more than some amount decidedly less than 50% of the total "specified assets," which would impact significantly the maximum interest each could have in the captive. The statute is silent on this issue, though the IRS may later choose to clarify the ambiguity with a regulation.

Notice 2016-66

As noted in the introduction, Notice 2016-66 formally classified certain uses of captives as "transactions of interest," requiring disclosure. Section 2.01 of the Notice describes the attributes of the transactions requiring disclosure. They include a person who owns a business entity acting as the insured. A captive also owned by that person, the insured, or related persons enters into purported insurance or reinsurance transactions with the insured. The captive makes an election under Code Sec. 831(b), allowing it to exclude net written premium from its income. The person, the insured, or related persons own at least 20% of the captive's stock. Finally, one or both of the following is true:

  • Total liabilities assumed by the captive for insured losses and administrative expenses during the "Computation Period" are less than 70% of:
  • Premiums earned by the captive during the Computation Period, less
  • Policyholder dividends paid by the captive during the Computation Period; or
  • The captive has made a loan or otherwise conveyed assets to the person, the insured, or a person related to such person or insured in a manner claimed to be tax free.

The Notice defines the "Computation Period" as the five most recent taxable years of the captive or, if the captive has been in existence for less than five taxable years, the entire period of the captive's existence.

To paraphrase, therefore, the Notice targets any Code Sec. 831(b) captive at least 20% of which is owned by the insured, the owner of the insured, or a related person, and which, over the prior five-year period, had a combined ratio that is less than 70% (or, alternatively, which made a tax-free transfer of reserves). Note, however, that depending on the specific line of insurance at issue, a 70% combined ratio would indicate a financially healthy captive and is a target for which many companies would aim. In other words, such a claims history should not, on its own, suggest that the insurance company has been levying unrealistic premiums, assuming implausible risks, or otherwise suspiciously avoiding significant claims experience. Combined with the low 20% ownership threshold, the IRS has opted to cast a wide enough net so as to require the disclosure of countless above-board captive arrangements.


While the statutory amendments may have adequately addressed one perceived abuse of micro captives, there are several additional known tax strategies using these entities. For uses outside of the gift and estate context, the same standards that pre-date the amendments will continue to apply in testing the bona fides of a particular arrangement. Thus, because of Congress' narrow focus, promoters will continue to market micro captives as an effective means of tax savings, meaning that taxpayers must remain vigilant in distinguishing between sound, legitimate uses of captives from those that the IRS has made a policy decision to combat. Undoubtedly, audit and litigation activity in these areas will continue over the coming years.

For those taxpayers with a good faith, bona fide interest in captive insurance coverage, great care is mandatory at the planning stage if they wish to take advantage of a Code Sec. 831(b) captive, and even then, they must be prepared at all times to defend an audit. But a taxpayer otherwise dissuaded from the use of captives because of Congress and the IRS's recent actions need not abandon the idea of a captive entirely. Such taxpayer should instead consider forming a captive and foregoing the Code Sec. 831(b) election, making it a so-called "831(a)" captive by default. While such an insurance company would not get the benefit of avoiding tax on net written premium, the company and its owner still would enjoy many of the same advantages as 831(b) captives, and as well as some additional advantages unique to 831(a) captives. For example, the captive's owner still gets the benefit of deducting its premium payments where the Code otherwise allows; it still can influence the type of insurance products available (so long as the actuary analysis justifies the premiums), and it has the benefit of an ownership interest in its insurance company. It tends to reduce the "sting" of paying premium to a profitable insurance company when one actually owns that company. Moreover, an 831(a) captive still is not taxed on unearned premium and is not subject to the $2.2 million dollar net written premium limit. Most importantly, an 831(a) captive is not subject to disclosure under Notice 2016-66, meaning it enjoys a certain level of obscurity. The one word of caution is to remember that captives of all stripes must meet the IRS's definition of "insurance." One must continue to use care, therefore, to make sure that one's captive meets those standards in an economically-substantive way.


1. Helvering v. Le Gierse, SCt, 41-1 USTC ¶10,029, 312 US 531, 539, 61 SCt 646 (1941).

2. Rev. Rul. 2002-90, 2002-2 CB 985.

3. Id.

4. CCA 201511021 (Dec. 1, 2014).

5. Through the cooperation of these unrelated groups, both risk-shifting and distribution are more easily met. See Rev. Rul. 78-338, 1978-2 CB 107.

Originally published by Journal of Tax Practice & Procedure.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.