France's favorite savings product is, in American eyes, almost never insurance. What it is instead, a PFIC wrapper or a foreign trust, decides years of tax and reporting.
No conversation in our Franco-American practice repeats itself as often as this one. The client, an American in France or a French family with an American child, holds an assurance-vie. The French banker praises its tax privileges, its succession clause, its eight-year clock. And the American analysis must begin with the sentence nobody wants to hear: for US tax purposes, this contract is almost certainly not insurance, and everything the French adviser loves about it is irrelevant to the IRS.
What the contract is instead is the genuinely difficult question, and the two candidate answers lead to different, equally demanding regimes: a transparent wrapper holding passive foreign investment companies, taxed under the punitive default rules of section 1291, or, on a minority analysis, a foreign grantor trust with the Forms 3520 and 3520-A exposure we have mapped elsewhere. The two paths share one certainty, FBAR reporting of the cash value, and one conclusion: the assurance-vie and the US person are a mismatch by construction.
Part I explains why American law refuses the insurance label. Part II details the PFIC path, the majority analysis. Part III examines the grantor trust thesis. Part IV sets out the triage we apply to existing contracts, including the French cost of exiting.
I. Why American law sees no insurance
A. Section 7702: a definition, not a label
US tax law does not take insurance on faith. Section 7702 of the Internal Revenue Code defines the life insurance contract for the whole of the tax code: a contract must be insurance under applicable law and pass one of two actuarial tests, the cash value accumulation test, capping the cash surrender value at the net single premium funding the future benefits, or the combination of the guideline premium requirements and the cash value corridor, which forces a genuine death benefit to stand permanently above the savings (IRC section 7702(a) to (d)). The French assurance-vie is built on the opposite philosophy: it is a savings vehicle wearing a thin layer of insurance, with a death benefit that typically tracks the account value rather than exceeding it through any corridor. Most contracts, and virtually all the standard bank-distributed ones, fail these tests, not marginally but by design.
B. Investor control, the second lock
Even a contract that squeezed through the actuarial tests would face the investor control doctrine, developed by rulings and case law: where the policyholder directs the investments inside the wrapper, selecting funds, arbitrating between supports, the American analysis treats the policyholder as owning those investments directly, and the wrapper evaporates. The assurance-vie multisupports, whose commercial appeal is precisely the client's hand on the allocation, is the textbook case. The consequence of failing either lock is the same: the United States looks through the contract, and the question becomes what the American taxpayer directly owns.
II. The PFIC path: the majority analysis
A. What the wrapper contains: PFICs, almost by definition
Looking through the contract, the American taxpayer owns units of French and European funds, OPCVM, SICAV, fonds euros structures. A foreign corporation is a passive foreign investment company where 75% or more of its gross income is passive or 50% or more of its assets produce passive income (IRC section 1297; IRS, Instructions for Form 8621). Investment funds meet these tests by their very purpose. The American holder of an ordinary assurance-vie multisupports therefore holds a portfolio of PFICs, exactly as if the funds sat in a taxable securities account, and each of them is a separate reporting unit on Form 8621.
B. Section 1291: the default regime is a punishment
Absent timely elections, each PFIC falls under section 1291: distributions exceeding 125% of the trailing three-year average, and all gain on disposition, are excess distributions, allocated ratably across the holding period, taxed at each year's highest ordinary rate, and charged interest under section 6621 on the resulting deferral (IRS, Instructions for Form 8621, rev. Dec. 2025). The regime converts long-term, patiently compounded savings, the very virtue the French product sells, into ordinary income plus an interest meter running from the first year. The escape hatches, the QEF election requiring fund-level American accounting that French funds do not publish, and the mark-to-market election requiring regularly traded status, are rarely available for the funds actually held inside French contracts. In practice, the longer the contract has run, the worse the arithmetic: time, the ally of the French saver, is the enemy of the American one.
C. The compounding layer: reporting
The tax is only half the burden. Each PFIC wants its Form 8621 annually in the relevant cases; the contract's cash value is a reportable foreign financial account on the FBAR, as an insurance policy with cash value under the regulation we detailed in our FBAR inventory; and Form 8938 duplicates the disclosure above its thresholds. A modest contract can thus generate more American paperwork than the rest of the taxpayer's estate combined, with penalty exposure on each layer, independent of any tax due, the pattern we described among the ten structural mismatches.
III. The grantor trust thesis
A. The argument
A minority of the American practice analyzes certain assurance-vie contracts not as transparent wrappers but as foreign grantor trusts: the policyholder settles assets with a French institution, retains control and revocation rights, and designates beneficiaries, a functional architecture that echoes the trust the Code knows. On that reading, the American policyholder is the grantor of a foreign trust, taxed on its income as it arises and, decisively, exposed to the information regime of Forms 3520 and 3520-A, whose penalties, the greater of 10,000 dollars or 35% of contributions and distributions, and 5% of the trust's value for the annual statement, we detailed in our double reporting trap analysis.
B. Our reading
We do not adopt the grantor trust characterization as a default: the contract is a bilateral insurance agreement under French law, not a trust instrument, and the majority practice taxes it as a transparent wrapper. But intellectual honesty obliges two admissions. The question has no authoritative published answer specific to the assurance-vie, so a conservative filer facing a large, old contract may reasonably protect the position with trust-style disclosure, the cost being paperwork rather than tax. And whichever characterization one retains, the economic outcome converges: current taxation of the inside build-up, punitive treatment of the funds, and a reporting stack. The debate changes the forms, rarely the conclusion.
IV. Triage: what to do with an existing contract
A. The four-question grid
Our triage runs on four questions. Who holds it: the US person directly, a non-American spouse, a child about to become American by residence? What is inside: fonds euros only, multisupports, dated portfolios with heavy embedded gains? Since when: pre-immigration contracts with modest American-period gains differ radically from contracts fed throughout US person years. And has it been disclosed: an undeclared contract is first a disclosure problem, cured through the streamlined procedures where non-willful, before it is an investment problem. Only after these four answers does the keep-or-surrender arithmetic begin.
B. The exit and its French price
Surrendering has a French cost that must enter the same spreadsheet: the gains are taxed at surrender, with the regime of Article 125-0 A of the CGI, reduced rates for contracts past eight years and an annual allowance of 4,600 euros for a single taxpayer or 9,200 euros for a couple on the taxable products (CGI, art. 125-0 A, version in force since 1 Jan. 2022). For many mid-sized contracts, the French exit cost is modest against the American cost of staying: the section 1291 interest meter does not stop, the compliance fees recur annually, and the succession clause the product was bought for produces, on the American side, none of its French magic. Our default for the compliant US person is therefore exit and reinvestment in American-compatible assets, with three standing exceptions: the non-American spouse's own contracts, which should stay strictly out of the American orbit; old contracts with small embedded gains, where a clean surrender is nearly free; and situations where imminent life events, a return to France, a succession, change the holder analysis itself.
C. Our position
The assurance-vie is not a bad product; it is the wrong product for a US person, sold by a distribution network that has no reason to know it. The correct sequence never varies: characterize the contract, disclose what should have been disclosed, run the two-country exit arithmetic, and rebuild the savings on instruments both systems tolerate. What we refuse is the comfortable middle path, keeping the contract undeclared because the French adviser insists it is insurance: that path converts an investment mistake into a penalty problem, and we have written enough about those to know how they end.
Conclusion
Through American eyes, the French assurance-vie fails the definition of insurance at section 7702 and the investor control doctrine, and dissolves into what it contains: a portfolio of PFICs under the punitive default of section 1291 for the majority analysis, or a foreign grantor trust with its 3520 exposure for a minority one. Both paths tax the build-up currently, both stack reporting on the FBAR baseline, and both turn the product's French virtues, time, capitalization, the succession clause, into American liabilities.
The instruction for the US person in France is accordingly blunt: never subscribe, and for existing contracts, run the triage now, disclosure first, exit arithmetic second, French surrender cost included. The instruction for the Franco-American family is subtler and more valuable: keep the assurance-vie where it belongs, in the hands of the non-American members, and give the American ones a portfolio built for their passport. The product is excellent; the match is the problem.
Frequently asked questions
My French bank says my assurance-vie is life insurance, so why would the IRS disagree?
Because American law applies its own definition. Section 7702 requires actuarial tests, a genuine death benefit standing above the savings, that standard French contracts are not designed to pass, and the investor control doctrine disregards wrappers whose holder directs the investments. Failing either, the United States looks through the contract and taxes what is inside, whatever the French label says.
What actually happens tax-wise if I keep my contract as a US person?
The funds inside are PFICs in the majority analysis: gains and large distributions are thrown back across your holding period, taxed at each year's top ordinary rate with an interest charge, and each fund wants its Form 8621. The cash value is FBAR-reportable every year, and Form 8938 may apply. The longer you hold, the worse the throwback arithmetic becomes, which is the exact inverse of the product's French logic.
Should I surrender my assurance-vie, and what does that cost in France?
Often yes for a compliant US person, but only after the arithmetic. France taxes the gains at surrender under Article 125-0 A of the CGI, with reduced rates and an annual allowance of 4,600 or 9,200 euros once the contract has passed eight years. That cost is frequently modest against the recurring American cost of staying: the section 1291 interest meter, annual compliance fees, and a succession clause that produces no American benefit. Pre-immigration contracts and a non-American spouse's contracts follow different logic.
I never declared my contract to the IRS. In what order do I fix this?
Disclosure first, investment decisions second. An undeclared contract is a reporting failure, FBAR at minimum, usually income and PFIC reporting too, and the streamlined procedures repair it at zero penalty for non-willful taxpayers. Surrendering before disclosing solves nothing and creates a taxable event on top of an unresolved compliance problem. Characterize, disclose, then run the exit arithmetic with counsel.