In two opinions issued just four days apart, the IRS has won two more cases involving risk-pooled 831(b) captive insurance companies, known as microcaptives. The opinions were issued in Kadau v. CIR, T.C. Memo. 2025-81 (July 31, 2025), which you can read here, and CFM Insurance, Inc. v. CIR, T.C. Memo. 2025-83 (August 4, 2025), which you can read here.
The Kadau case involved an engineering company that paid premiums to its affiliated captive. The captive manager was RMC, which I wrote about in my article, Captive Manager Avoids 6700 Penalties After Jury Trial (April 13, 2024). The actuary was Marn Rivelle, a very nice guy but also who seems to feature frequently in these losses to the IRS. Because the underlying business could not otherwise meet the risk distribution requirements for a captive insurance company, it participated in a risk pool known as RMC Property, which of course was affiliated with the RMC group of companies.
The Kadau captive failed for tax purposes for pretty much the same reasons that we see in all of these captive cases, i.e., the “reinsurance” agreement through the RMC risk pool provided little more than a circulation of the funds back to the owner of the business paying the premiums, the captive was undercapitalized for the risks that it was purportedly taking on, the premiums were not properly calculated in any realistic actuarial sense, the policies issued were kind of a joke once the terms were scrutinized, and were unreasonable, and the like. Worse, the captive’s primary investment was a life insurance policy on the taxpayer, which indicated that the real purpose of the arrangement was for the owners of the operating business to dodge taxes. However, the taxpayers were able to dodge additional taxes because the IRS lost on its argument that the captive’s 953(d) election (for a foreign insurance company) had failed.
Nonetheless, the Tax Court found that the taxpayers did not reasonably rely upon RMC and Rivelle so as to avoid penalties, although the Tax Court indicated that it would defer a ruling as to whether the taxpayers would pay the 20% inaccuracy penalty or the 40% gross valuation misstatement penalty.
The other opinion, as indicated above, was from the CFM Insurance case. Owners of a Chicago grocery store chain created a Utah captive to cover a variety of risks related to the company, which risks were arguably more legitimate than those found in probably all the other microcaptive cases to date. The problem was that to handle these risks, the owners of the grocery stores ended up using Artex Risk Solutions, a division of Arthur J. Gallagher, to set up their captive â and Artex has also appeared in more than its fair share of these microcaptive cases where the taxpayers got slammed.
Nonetheless, the captive seems to have been more legitimate than most microcaptives. There wasn’t a bogus risk pool involved to make things easy for the Tax Court. The actuarial calculations bore some resemblance to reality within the insurance market place, the grocery store chain really did have substantial risks that needed to be covered by insurance, and there were enough of these risks to get the captive over the “points of insurance” a/k/a “risk exposures” threshold, which pretty much never happens with a microcaptive. Indeed, the Tax Court found that there was approximately 4.5 million such exposure units which was sufficient to get the captive’s underwriting well into the that Valhalla of insurance tax law known as the Law Of Large Numbers. Even the premium calculations were held to be reasonable, which is a first in these microcaptive cases. “So far, so good!”
Said the man falling from a skyscraper as he passed the 30th floor plummeting downwards. Alas, the captive here also ended with a thud and a bloody mess on the pavement.
Even though the premiums were reasonable in amount, the way the premiums were paid was shoddy at best. The premiums charged by the captive were not paid by the grocery store chain until December of their year of issuance, which violated the terms of the policies issued (except for the 2015 premium, which although also paid in December did not violate the policies since by then the requirement had been eliminated).
But the payment of the policies paled in significance to when the policies were issued. For 2012 and 2015, the policies were not issued by the captive until the policy period had ended, and for 2013 the policies were not issued until only four days before the policy period expired. For 2014, there were five months remaining in the annual policy period before the policies were issued, but the Tax Court did not find this to be timely. While there were binders issued for the policies, these expired after 150 days and that time seems to have passed before the policies were issued.
The drafting of the captive policies was at best sophomoric. Key terms were frequently left undefined and often criteria was missing to determine if a covered claim had occurred at all. The key supplier policy did not define who as a key supplier. The reputational risk policy did not specify the instances when a reputation was damaged. They key employee policies did not define a category of key employees. Continuing on, the Tax Court noted:
“And then there was the business-interruption DIC policy. The 2014 policy provided coverage for a long list of events but did not define any of the terms used, including ‘economic sanctions’ and ‘denial of access’ in the 2014 policy. Most alarming is that the 2014 and the 2015 business-interruption DIC policies purported to cover losses from terrorism, pollution, and dishonest acts by employees, but also included a blanket exclusion for claims based on those very conditions.” (Emphasis added).
Although rating it as a neutral factor, the Tax Court also found that the captive’s claims-handling practices were also shoddy.
The next set of facts reviewed by the Tax Court were also pretty ugly. One such set of facts went to the owner’s understanding of the captive relationship. As the Tax Court noted:
“While CFM was managed by Artex, Presta was the 50% owner and president of the company. He testified at trial that he knew little of the operationsâhe even forgot that he had appointed himself as CFMâs president. We donât think that outsourcing the operation of a captive undercuts in all cases the characterization of a company as an insurance company, but when the president of the company doesnât even know that he is the president, something is off.”
Ultimately, the Tax Court found that the captive arrangement here was “a much closer call than is usual in microcaptive cases,” but it still fell short of being an insurance arrangement for tax purposes. Because of that, the 831(b) election made by the captive would be held ineffective and the premium payments made to the captive would not be deemed as “insurance”. Further, the taxpayers would not be allowed to recharacterize the payments made to the captive as a loss reserve, since the apparent purpose of the captive arrangement when it was formed was to avoid the taxes that would have been payable for a mere loss reserve.
So the taxpayers in CFM lost their deduction, but what about penalties? The Tax Court held that the taxpayers had relied on their own independent CPA for advice about the tax treatment of their captive and that their reliance was reasonable under the circumstances. Thus, no penalties.
ANALYSIS
The difference in the two cases is that in Kadau the taxpayers simply relied upon the promoter of the captive (and thus promoter of the tax shelter) and its hired actuary as to the proper tax treatment, whereas the taxpayers in the CFM case undertook the wise decision to ask their CPA about the taxation of the arrangement. The Kadau taxpayers ended up paying penalties and the CFM taxpayers did not. That is the primary difference of these two cases.
Otherwise, both cases are alike in the sense that it is not so much the formation of the captive and how it is licensed that matters so much as it is how the captive is operated after that. The Kadau captive arrangement involved a risk pool, which has repeatedly proven to be pretty much an automatic loss in all of these cases, and also invested in life insurance policies which make no sense for a captive. While the CFM captive arrangement surprisingly (if not shockingly) got the premium payments right and had sufficient risk points to meet risk distribution, the shoddy if not outright shambolic running of the captive in terms of policy issuance and payment was its doom.
Thus, a loss is a loss is a loss, although the CFM taxpayers avoiding penalties was sort of a consolation prize for getting many things right and not just outright imploding like so many of these deals.
Can any microcaptive arrangement get past the Tax Court? Not holding my breath, but the CFM opinion indicates that there are some such arrangements out there which have a better chance than others.
On the other hand, and as numerous decisions including now Kadau indicate, 831(b) risk-pooled microcaptives are simply dead on arrival at the U.S. Tax Court.