Implementation · life insurance

Replicate the strategy in life insurance

Hold the dual-momentum portfolio inside a life-insurance contract — in France and in its European equivalents: the principle, four French contracts taken as examples, and the fund ↔ US-ETF mappings.

The wrapper is part of the strategy

A momentum strategy turns over, and every rotation sells positions that have risen. In an ordinary brokerage account, each winning sale triggers tax, which chips away at the compounding capital on every rotation; inside a French life-insurance contract, switching between funds triggers no tax — the strategy turns over fully tax-deferred, tax falling due only on withdrawal, at a reduced rate after eight years. On the tax dimension alone, this deferral lets the wrapper finish 15 to 55 % ahead of a brokerage account depending on the bracket.

But the wrapper has a cost, to be weighed against that advantage. Annual management fees take 0.5 to 0.6 % a year of the savings depending on the contract; and while the arbitrage itself is free, every ETF entry or exit carries a markup or markdown of about 0.10 % of the price (zero on a few contracts) — a real cost, simply hidden behind the “free arbitrage” wording on confirmations. On top sits imperfect replication: the investable universe is limited to the funds the contract lists, without the fine growth/value rotation or the commodity-sector futures that no contract offers — you work with proxies. Compounded over a long horizon, these fees and gaps can erode on the order of 15 % of performance; the simulator gives the net figure, contract by contract.

On a net basis, then, the wrapper's financial edge survives only when the tax saving exceeds this drag: clear at the high brackets (the flat tax), it narrows toward a wash at the low ones.

Two advantages a brokerage account cannot match remain, and often justify the wrapper on their own. Succession: on death, the savings pass to the named beneficiary under life insurance's own regime, distinct from inheritance tax and carrying specific allowances (Tax & succession). And simplicity: a percentage switch between two funds is far simpler to place — and to track for tax — than a series of buy and sell orders in a brokerage account.

Two constraints frame all this. The eight-year clock runs from when the contract is opened, not when money goes in: an old contract is an asset in itself, and moving from one insurer to another counts as a withdrawal (tax due, clock reset) — hence the choice to fit the strategy to the funds already on offer and to backtest on faithful proxies. And in a systemic crisis, the regulator can temporarily suspend withdrawals and switches, the euro fund first (Sapin 2 & euro funds): the shelter can be locked exactly when you would want to act.

Fees last checked: June 2026. Mandatory review: May 2028 at the latest.

Four universes in the picker, under the Assurance-vie group, replicate the real 70/30 inside a life-insurance contract (the four contracts taken as examples are French), using only the funds those contracts offer: 70% US equity (SPY+QQQ) and 30% real assets (world energy IXC and a broad-commodity basket; the top 2 are held, and the commodity sleeve takes shelter in short-dated government bonds SHY). Each simulation proxy is chosen to match the fund the contract actually holds — decided on what the fund contains, never on what suits the backtest. Two examples. The held energy fund is broad, non-US energy (Spirit 2's is European, STOXX 600 Energy; Swiss Life's is world energy): the proxy is therefore world energy IXC, not US XLE. And the broad-commodity fund tracks an equal-weight, ex-agriculture Bloomberg energy-and-metals index: we model it with an equal-weight blend of energy, gold, silver and base metals rather than the energy-heavy DBC (which keeps agriculture and over-weights oil).

Comparable wrappers exist in most European countries — Luxembourg life insurance with its "security triangle", Belgian branch 21/23 contracts, Portuguese seguros de capitalização (with a holding-period logic close to the French "8 years"), Italian polizze vita, the Swedish kapitalförsäkring — each with its own duration, tax and protection rules. The replication logic described here carries over: adapt the strategy to the catalogue of the wrapper you hold, rather than the other way round. Generali and Swiss Life, moreover, are pan-European insurers present in several countries. But while the principle holds everywhere, the fee schedules, fund menus and above all the taxation are specific to each contract and each country: the figures on this page are those of the French contracts, and the structure is probably close elsewhere without having been verified country by country.

The tables below map each held fund to its US simulation proxy. When the proxy is marked "built", no single US ETF tracks the fund's index, so the proxy is assembled from several US ETFs at fixed weights, rebalanced monthly — an equal-weight energy-and-metals basket for the broad-commodity sleeve.

The four contracts below are the ones the form can clone. For each: the reference contract — the one whose catalogue drove the mapping — where it is taken out, and its operational parameters. Two contracts from the same insurer offer NEITHER the same catalogue NOR the same conditions: check each holding (ISIN) and the general terms of your own contract. We have no commercial relationship with these brokers or insurers.

For a monthly-signal clone, two parameters weigh as much as the catalogue: the cost of rotation (arbitrage and transaction fees, charged on each reallocation) and the DELAY between the arbitrage order and its value date — every day of gap between the signal and execution moves the contract away from the reference strategy. The cards below summarise these parameters, as recorded on the date shown: the general terms in force prevail.

What the wrapper costs — net return, contract by contract

Earlier, we said that fees and imperfect replication can eat into performance, and that the simulator gives you the net figure. Here it is, over 2016-2026. Each contract is measured twice: first with its proxies alone (before fees — which isolates the FIDELITY drag of the catalogue), then net of management and rotation fees (the FEE drag). Against the reference: the real US-ETF 70/30.

The reference→proxies gap is the cost of the LIMITED CATALOGUE (SPY+QQQ instead of the fine IWF/IWD/IWB rotation; CMDEW+IXC instead of the four commodity sectors): gold-holding clones (Spirit 2, Suravenir) lose almost nothing on fidelity (−0.4 pts), while without gold (Swiss Life, Generali) the drag is clearer (−2.1 pts) over this window. The proxies→net gap is FEES: Generali charges no ETF buy/sell markup (free rotation), so its fee drag is near-zero (−0.7 pts) despite a costlier 0.60% management fee — its net return even exceeds Swiss Life's. The cost of rotation weighs more than management. Figures taken in June 2026 over 2016-2026 (not the long horizon "~15%" mentioned above); Suravenir is an informational clone, unfunded. This return loss is weighed against the wrapper's tax advantage (…): at high brackets the tax deferral more than offsets it, at low brackets it's tighter — situation-dependent.

All four are honestly labelled replicas, not diversifiers: they run the same logic, the same sleeves, and the same defensive discipline as the real 70/30, so their monthly returns track the US account closely. They put the same bet in a second, tax-advantaged envelope — useful for sheltering the strategy, not for spreading the risk. They are read-only here and never traded.