10 August 2015Actuarial & underwriting

A brief history of the (captive) world—and its ongoing battle with the IRS

The captive industry has been around for almost 50 years and for almost that long has been fighting with the Internal Revenue Service (IRS) in seeking to have captive insurance companies receive the favourable tax treatment allowed by the US tax code for insurance companies.

The normal tax treatment of business expenses is that they are deductible when paid. Insurance companies get to deduct expected losses when they are incurred. The public interest in ensuring that insurance companies have sufficient funds in reserves to meet their future obligation to pay claims when they are settled means that insurance companies are able to deduct uncertain future commitments (subject to some discounting factors), giving them a substantial deferral of income until known claims are settled out.

Self-insured companies are not able to deduct loss reserves and incurred-but-not-reserved amounts, they can deduct losses only when they are paid. Thus if a corporation could form an insurance company that met the known rules to achieve insurance company tax treatment it could achieve a significant tax deferral benefit. The benefit is a timing issue as the insurance company treatment is allowing the acceleration of the loss deduction to year one instead of receiving the tax benefit as the claims are paid out, say over 10 years for workers’ compensation.

If expected workers’ comp losses for 2014 were $10 million, a captive passing muster would be able to deduct the full $10 million as an expense, as opposed to a self-insured which would probably be able to deduct $2 million in actual claims payments made in 2014. The remaining $8 million in claims would be deducted from 2015 to 2024 as they were paid. The benefit of the $8 million acceleration depends on internal rates of return but actuarial firms have ballparked the benefit as between 6 percent to 8 percent, or $500,000 to $650,000, annually in the above example.

"Multiple entities or sufficient unrelated risk can be structured to follow the court cases of revenue rulings and create a tax position that is strongly defensible."

A simple captive would cost approximately $80,000 in operating costs and while there may be many other reasons for forming this company, the tax efficiency remains an often decisive factor in the formation of a captive. (Other reasons can include lack of markets, access to lower cost reinsurance, a profit centre, policy wording flexibility, collateral relief, a cost allocation tool and other insurance or business problems.)

The IRS long held to the “economic family” argument which took the position that a captive was not acting as an insurance company when it was a controlled entity of its parent company. The basic argument was that the captive did not meet the two basic tests for insurance: risk distribution and risk transfer.

Risk distribution is the law of large numbers at work. The IRS believes that insurance should “spread” the risk of loss among “many” entities or exposures. This grey area has been the biggest battle ground for the captive industry.

Risk transfer means that the insurance company must be able to make either a profit or a loss. The issue here has always been what constitutes a “loss”. The industry for many years used a rule of thumb of a 10 percent chance of losing 10 percent (for property and casualty exposures). This position has been challenged but there is, to date, no clarity to this issue.

Part of the issue was that some risks (eg, workers’ compensation) have a long payout pattern and an insurance company could easily recoup the 10 percent risk margin through investment income, which might call into question whether the captive is really “making a loss”.

Important court decisions

Helvering v LeGierse, 312 US 531 (1941) established that both risk transfer and risk distribution are required for a contract to be treated as insurance.

Throughout the 1970s and 1980s the IRS held to its doctrine of the economic family and courts disallowed deductions for premiums paid to captives in Carnation v Com’r, 71 TC 400 (1978), Stearns-Roger v Com’r, 577 F.Supp 833(d.Cob 1984) and Clougherty Packing v Com’r, 84 TC 948 (1985).

Sears, Roebuck v Com’r 972 F.2d 858 (7th Cir. 1992) took this position to an extreme. Sears paid $14 million to Allstate, a wholly owned company. Allstate wrote over $5 billion in premiums so the Sears premiums constituted less than 0.5 percent of the total premium written. The IRS contended this did not meet risk transfer or risk distribution and was simply moving money between related companies. The court disagreed, indicating that there was clearly risk distribution among numerous entities and that Allstate could suffer a loss from the Sears business.

The captive industry started to win, beginning in the late 1980s with a now seminal case.

In Humana v Com’r 881 F.2d 247 (6th Cir. 1989) the 6th circuit court of appeals held that a “brother-sister” captive insurance structure constituted insurance for federal income tax purposes and thus premiums from the captives brother-sister entities were deductible. The court also held that premiums from the captive parent were not deductible. The decision was based on a “balance sheet” approach, whereby “risk shifting” is recognised if the value of a loss is transferred from a brother-sister’s balance sheet to the captive.

“The captive insurance company, in turn, can elect under a separate section of the tax code to be taxed only on the investment income from the pool of premiums, excluding taxable income of up to $1.2 million per year in net written premiums." IRS website

Brother-sister refers to multiple separate legal entities. The structure in Humana was approximately 30 companies under a holding company and the holding company was the owner of the captive.

This was followed in Kidde Industries v US, 40 Fed Cl.42 (1991) while The Harper Group v Com’r 96 TC 45 (1991) found that risk transfer and risk distribution were achieved when the captive received 29 to 32 percent of its premiums from unrelated parties.

Amerco and a number of its subsidiaries purchased insurance policies from Republic Western Insurance Company and deducted the premiums for income tax purposes. In Amerco et al v Com’r, 9th Cir. (1992) the insurance business from the Amerco group constituted from between 26 percent to 48 percent of Republic’s business; the remaining insurance business was unrelated. Because Republic was a subsidiary of Amerco, the IRS determined that the transactions did not constitute “insurance” for income tax purposes, and disallowed the deductions. Amerco petitioned the Tax Court for a redetermination and the Tax Court held that the arrangement between Amerco and Republic constituted insurance for federal income tax purposes.

Following on from these defeats in Harper, Kidde, Harper and Amerco, the IRS issued three revenue rulings in 2002.

Rev Ruling 2002-89 indicated that 50 percent unrelated risk qualifies as sufficient risk distribution for insurance company tax treatments.

Rev Ruling 2002-90 indicated that 12 or more subsidiary companies, where no one sub comprised more than 15 percent of the total premium, also achieved sufficient risk distribution for tax treatment as insurance

Rev Ruling 2002-91 indicated that seven or more unrelated insureds in a group captive also achieved risk distribution and risk transfer.

There was no explanation as to why seven entities worked for unrelated companies but it needed 12 subsidiaries to achieve risk distribution. It is worth noting that the “safe harbour” rulings are set much higher than some of the court cases discussed previously.

In Rent-A-Center and Affiliated Subsidiaries, 142 TC 1 (2014) the court held that in order to have risk distribution, the insurer needs to insure a large enough pool of unrelated risks. The Tax Court said a captive may achieve adequate risk distribution by insuring only subsidiaries within its own affiliated group. There were a sufficient number of statistically independent risks in the Rent-A-Center case given the large number of stores, employees and vehicles.

This decision is considered significant in that the court seemed to look at total exposure units more than number of entities in determining risk distribution and also that it did not place much import on an intercompany loan, a fact the IRS had previously argued was evidence of a lack of risk transfer.

Most practitioners are unwilling to go out on a limb and argue that Rent-a-Center has moved the bar. It is still seen as prudent to look for seven or more subsidiaries or to try and find 30 percent or more unrelated risk if a captive owner wants to achieve favourable tax treatment. These levels are still in the “grey” area which could be challenged by the IRS as being outwith their safe harbour but are seen as very defensible based on the court cases discussed in this article.

It is also worth noting that the IRS issued proposed guidance on risk transfer and risk distribution in September 2011 which would have “looked through” to a segregated cell and each cell would then be looked at to see if it met risk distribution and risk transfer tests as if it were a standalone entity. There has been no follow-up to this guidance but many in the cell industry have acted as if it is the expected IRS position, even though it has not been formalised.

Other federal tax issues

Small insurance company election

There has been a huge increase in the formation of captives that write premiums below $1.2 million and elect to pay tax only on their investment income under IRS code section 831(b). The aggressive use of this vehicle by estate and tax planners has resulted in the IRS including captive insurance in their “dirty dozen” list of tax scams.

To quote from the IRS website:

“Another abuse involving a legitimate tax structure involves certain small or ‘micro’ captive insurance companies. Tax law allows businesses to create ‘captive’ insurance companies to enable those businesses to protect against certain risks. The insured claims deductions under the tax code for premiums paid for the insurance policies while the premiums end up with the captive insurance company owned by same owners of the insured or family members.

“The captive insurance company, in turn, can elect under a separate section of the tax code to be taxed only on the investment income from the pool of premiums, excluding taxable income of up to $1.2 million per year in net written premiums.

“In the abusive structure, unscrupulous promoters persuade closely held entities to participate in this scheme by assisting entities to create captive insurance companies onshore or offshore, drafting organisational documents and preparing initial filings to state insurance authorities and the IRS. The promoters assist with creating and ‘selling’ to the entities often times poorly drafted ‘insurance’ binders and policies to cover ordinary business risks or esoteric, implausible risks for exorbitant ‘premiums’, while maintaining their economical commercial coverage with traditional insurers.

“Total amounts of annual premiums often equal the amount of deductions business entities need to reduce income for the year; or, for a wealthy entity, total premiums amount to $1.2 million annually to take full advantage of the Code provision. Underwriting and actuarial substantiation for the insurance premiums paid are either missing or insufficient. The promoters manage the entities’ captive insurance companies year after year for hefty fees, assisting taxpayers unsophisticated in insurance to continue the charade.”

Cascading federal excise tax

Premiums ceded to insurers or reinsurers outside of the US are subject to a 4 percent federal excise tax (FET) on direct premiums and 1 percent on reinsurance premiums. The IRS held that this tax “cascaded” and was due if an offshore company reinsured to another non-US company and so on. This was an area of audit effort in recent years looking for foreign schemes and additional revenue. Validus Reinsurance, a Bermuda company, decided to challenge this tax.

Validus filed suit in the US District Court challenging the IRS’ determination and assessment of tax on the grounds that the FET does not apply to transactions between two non-US reinsurers that occur outside the US. Validus also contended that the IRS lacks the power to tax transactions between non-US parties.

The District Court ruled in favour of Validus. It noted that Validus’ transactions are retrocessions and reasoned that the plain language of the statute does not impose a FET on retrocessions.

953(d) Election

In 1989, the IRS published Notice 89-79, which provides substantive and procedural rules regarding an election under section 953(d). Section 953(d) allows a controlled foreign corporation engaged in the insurance business to elect to be treated as a US corporation for US tax purposes. A controlled foreign corporation that makes this election will be subject to tax in the US on its worldwide income but will not be subject to the branch profits tax or the branch-level interest tax imposed by section 884. Further, the FET imposed under section 4371 on policies issued by foreign insurers will not apply.

The need for a closing agreement and letter of credit (LoC) to cover possible tax obligations to the US has become more contentious recently. The IRS has been using a more aggressive formula in determining required amounts as the formula says “10 percent of the electing corporation’s gross income”. When a captive wrote $1 million in premium in December of (say) 2014 the IRS, in several cases, argued that the corporations gross income should be $12 million ($1 million in December would be held to be $12 million annualised). This would require posting a LoC for $1.2 million instead of $100,000. There is some thought this is being done to make life more difficult for offshore companies to take the 831(b) election.

Where we are today

The captive world has had a string of court successes that today create a reasonable road map for a company to follow to achieve insurance tax treatment for a captive subsidiary. Multiple entities or sufficient unrelated risk can be structured to follow the court cases of revenue rulings and create a tax position that is strongly defensible. However, the explosion of 831(b) small insurance company captives has put captives back in the crosshairs of the IRS.

It is clear the IRS is still looking to challenge captive insurance arrangements where it sees an operational fact pattern that is unusual for an insurance company or a fact pattern that has had questions raised in court decisions.

Watch out for RVI Guaranty and Subsidiaries v Com’r which is being tried currently. It is addressing another unclear insurance issue—what constitutes insurance risk as opposed to business risk. This case could have major impact on many of the esoteric covers being written in some 831(b) captives and whether the IRS will have some court guidance as to what is business risk as opposed to insurable risk.

The fight goes on!

Gary Osborne is president of USA Risk Group. He can be contacted at:

More on this story

23 November 2021   Almost four decades in the industry.
12 October 2021   Hartley Hartman will help drive strategy in growing market.
4 September 2021   Gary Osborne of Risk Partners is confident that captives will remain an important tool for business for decades to come and those who join the industry will be rewarded with a long and prosperous career.