Are Micro-Captives Legal? What the IRS Actually Targets

It’s a fair question, and it deserves a straight answer: yes, micro-captives are legal. Captive insurance is a long-established, legitimate way for a business to finance its own risk through an insurance company it owns — and a micro-captive is simply a small one that may elect a particular tax treatment. What the IRS pursues is not captives as a category. It pursues abusive arrangements that wear the insurance label without doing the insurance work. The whole question is which one you’re looking at, and this piece is about telling them apart.

That distinction matters because the search for “are micro-captives legal” usually comes loaded with a half-heard rumor — a story about an IRS crackdown, a court loss, a deduction that got disallowed. All of that is real. None of it means captives are illegitimate. It means the IRS has been precise, and consistent, about what it considers a sham: an arrangement that isn’t real insurance and exists mainly to produce a tax result.

The short answer, and why it isn’t a dodge

A captive is an insurance company you own. Instead of paying an outside carrier for risks you understand better than they do, your business forms its own licensed insurer, writes real policies, sets premiums, and pays its own claims. That’s a mainstream risk-financing tool, used at every scale of business. A micro-captive is a small one — small enough that its annual premiums fall under a statutory ceiling, which makes a specific tax election available.

That election is Section 831(b) of the Internal Revenue Code. It lets a qualifying small non-life insurance company elect to be taxed only on its investment income rather than on its underwriting income, provided its premiums stay at or below the limit — $2.9 million for the 2026 tax year (Rev. Proc. 2025-32, indexed annually). Read the order carefully, because it’s the heart of the legality question: the election is available to a company that already qualifies as an insurance company. It is not a structure you build to reach a tax result; it’s a tax treatment that becomes available once a captive operates as genuine insurance. The election follows the insurance — never the reverse.

This is general educational information, not legal or tax advice — outcomes depend on the specific facts, and any election should be reviewed with your own qualified advisors. See our disclosures.

So the honest answer to “is this legal” isn’t a dodge. It’s a redirect to the real question: is it genuine insurance? If yes, it stands on solid, decades-old legal ground. If no, no tax election rescues it.

What “real insurance” actually means

Federal tax law has no statutory definition of insurance. The controlling source is the Supreme Court’s holding that an arrangement is insurance only if it involves both risk shifting and risk distribution (Helvering v. Le Gierse, 312 U.S. 531 (1941)). The IRS and the Tax Court apply that through a working framework: there has to be genuine insurance risk, the risk has to actually shift from the insured to the insurer, the insurer has to distribute that risk across a pool of independent exposures, and the whole thing has to look like insurance in its commonly accepted sense — arm’s-length pricing, real policies, claims actually handled, adequate capital, an entity run like an insurer.

Two of those do the heavy lifting. Risk shifting means the financial consequence of a loss genuinely moves off your business and onto the captive. Risk distribution means the captive pools enough independent, uncorrelated exposures that the law of large numbers can work — the many premiums pay the few claims. For a single-owner captive, shifting is usually the easy half; distribution is where these arrangements are made or broken, because a captive insuring only its owner’s related businesses has very few independent exposures, and a handful of correlated risks don’t distribute. (This is one reason a true multi-member group captive, pooling unrelated businesses, can have a more straightforward distribution story than a single-owner micro-captive — though both rest on the same fundamentals.)

These are substantive requirements, not boxes to check. And they’re exactly the line the IRS polices.

What the IRS actually targets

Here’s the part worth being precise about, because it’s where the “captives are illegal” rumor comes from. The IRS has named abusive micro-captive insurance arrangements on its annual “Dirty Dozen” list of tax scams — first in 2015, and still on the list as recently as 2024. In its own descriptions, the arrangements it targets share a profile: they lack the attributes of genuine insurance. The risks insured are implausible or don’t match a real business need; the coverage duplicates policies the business already carries; and the “premiums” are excessive and not set at arm’s length.

Put plainly, the IRS targets the imposter, not the real thing.

The leading court cases read like a field guide to that imposter. In Avrahami v. Commissioner, 149 T.C. 144 (2017) — the first major Tax Court loss for a micro-captive — the arrangement was held not to be insurance. It failed for lack of risk distribution; the premiums jumped roughly seven-fold once the captive appeared and tracked the 831(b) ceiling rather than the risk; the pricing wasn’t actuarially supported; and a reinsurance “pool” functioned as a circular flow of funds rather than a bona fide insurer. In Reserve Mechanical Corp. v. Commissioner, T.C. Memo. 2018-86 (affirmed by the Tenth Circuit at 34 F.4th 881 (2022)), the captive failed again on two independent grounds — inadequate risk distribution and policies that weren’t insurance in the commonly accepted sense — with a pool the court again found to be a circular flow and premiums that weren’t arm’s length.

Notice what the IRS won on. Not “captives are bad.” It won because those specific arrangements weren’t real insurance. The case law and the “Dirty Dozen” description are describing the same failure from two directions: no genuine distribution, pricing reverse-engineered from a tax number, money circling back to related parties, and tax as the actual point of the exercise.

One case in this area is frequently overstated, so it’s worth a clarification. CIC Services, LLC v. IRS, 593 U.S. 209 (2021), was a unanimous Supreme Court ruling — but a purely procedural one. It held that the Anti-Injunction Act didn’t bar taxpayers from challenging an IRS reporting requirement in court before being penalized. It was not a ruling that micro-captives are legitimate, and it should never be cited as the Supreme Court blessing captives. Legitimacy still rests exactly where Le Gierse put it.

What legitimate looks like

Flip every one of those failures around and you have the profile of a captive that stands on solid ground:

• Real risk shifting and distribution. Genuine, fortuitous risk leaves the operating business, and it’s pooled across enough independent exposures to actually distribute — often by adding unrelated risk through a real pool or reinsurance arrangement that prices at arm’s length and could genuinely pay claims.
• Arm’s-length pricing. Premiums are set by an independent actuary to the actual risk assumed — not reached upward toward a tax ceiling.
• Real, needed coverage. The captive insures exposures the business actually carries and the commercial market underprices or won’t write — not duplicate cover and not implausible risks no one would seriously insure against.
• Run like a real insurer. Adequate capital, real policies, claims actually handled, annual filings and audits, every year — not a shell that files once and goes quiet.

That’s the difference between the two columns above, and it’s not subtle. A captive built this way isn’t skirting the rules; it’s doing the thing the rules describe.

The enforcement reality, without the alarm

It would be dishonest to wave away the scrutiny, so here it is plainly — and in proportion. The IRS has treated abusive micro-captives as a real enforcement priority for the better part of a decade. Beyond the annual “Dirty Dozen” listings, it stood up dedicated micro-captive examination teams (announced in 2020) and ran time-limited settlement initiatives for taxpayers already under exam (IR-2019-157 in 2019, and a second, stricter offer, IR-2020-241, in 2020). The leading cases decided on their merits — Avrahami and Reserve Mechanical — went the IRS’s way.

The disclosure rules are their own moving piece, and this is the part you should not freeze from an article. The IRS finalized disclosure regulations in 2025, but in 2026 a federal court vacated the listed-transaction designation while leaving the transaction-of-interest designation in place — and both sides have appealed. So the precise current reporting obligation for any given captive is genuinely in flux, and the right move is to confirm it with current tax counsel for the relevant tax year rather than rely on anything evergreen. (For the fuller regulatory and case-law treatment, see our compliance and legitimacy page.)

None of that should read as a warning to stay away. Read in proportion, the enforcement record says something reassuring for a business considering a real captive: the IRS has been consistent and specific about what it pursues, and it isn’t real insurance. Scrutiny is a reason to build carefully and document thoroughly — it is not a verdict on the tool.

The honest bottom line

Are micro-captives legal? Yes. Captive insurance is legitimate, decades-established, and governed by a clear body of law. The arrangements that get into trouble are the ones engineered backward from a deduction, with risk distribution and pricing treated as formalities — and the IRS targets those precisely because they aren’t insurance. No honest captive manager will promise you immunity from examination; there’s no such thing, and anyone who claims a structure is “audit-proof” is selling the same imposter the IRS is hunting. What sound structure does is put a genuine captive in a position to answer the questions if they ever come.

The way to know which side of that line your business falls on isn’t a sales pitch — it’s a feasibility study that tests the insurance case honestly before anything is formed, and is willing to tell you no. (If you want the procedural view of how that unfolds, see how to start an 831(b) captive.) If a captive genuinely fits your risk, it’s legal, defensible, and built to last. If it doesn’t, the most valuable thing anyone can tell you is that it isn’t the tool for you. Either way, you’ll know where you stand — which, for a question this serious, is the whole point.

This is general information, not legal or tax advice — see our disclosures.