A generational gold allocation is not a trade. It is built once, adjusted rarely, and inherited by people who did not choose it. Everything that goes wrong with these structures goes wrong administratively — not because the price moved, but because nobody wrote down where the metal is and who is entitled to it. This edition is general educational material, not advice; the structures described require local legal and tax counsel.
Decision one: sizing
The conventional five per cent allocation comes from a specific set of assumptions — a stable monetary regime, positive real yields on sovereign debt, and bonds that diversify equities. When those assumptions hold, five per cent is defensible. The relevant question is not what the textbook number is but whether the assumptions behind it still describe the world you are allocating into.
The more useful framing replaces how much should I hold with what am I insuring against, and how much cover does that require. If the exposure you are hedging is a domestic currency and a domestic banking system, the position needs to be large enough to matter if that exposure fails, and it needs to sit outside that system. That reasoning produces a range rather than a number, and it produces a different number for a family whose operating business, property and pension are all denominated in one currency than for one already diversified across three.
Whatever number you arrive at, write down the reasoning next to it. In fifteen years someone will ask why the figure is what it is, and the answer must not be because that is what it has always been.
Decision two: form
Long-horizon allocations are typically dominated by allocated, audited physical metal, with a smaller liquid sleeve for rebalancing. The logic behind that split is functional rather than aesthetic: the physical share exists to be untouchable, and the liquid share exists to be touched.
This matters more than it sounds. A position you can sell in ninety seconds from your phone will eventually be sold in ninety seconds from your phone, usually during the exact week you should not. Deliberate friction — a vault, a custodian instruction, a settlement window — is a feature of the physical leg, not a cost of it. Meanwhile the liquid leg absorbs the rebalancing trades so the core is never disturbed.
The exchange-traded sleeve brings its own considerations: whether the fund is physically backed and allocated, whether it permits redemption in metal and at what size, its expense ratio compounded over decades, and its tax treatment in your jurisdiction, which for precious-metals vehicles is frequently unlike that of other funds.
Decision three: jurisdiction
Most families under-think this and diversify geographically when they should be diversifying legally. Two vaults in two countries under substantially similar legal systems, similar treaty networks and similar political alignment are one jurisdiction wearing two hats.
The questions that actually differentiate a storage jurisdiction: is the metal held in a bailment arrangement outside the custodian's estate in insolvency; does the country have a modern history of capital controls or requisition; what is the tax treatment on storage, transfer and inheritance; what physical access rights does an owner or heir have, and how are they exercised from abroad; and how deep is the local dealer market if a sale becomes necessary at short notice.
Concentration risk here is not only about the country. Multiple accounts with the same vault operator, or metal insured by the same underwriter, is a single point of failure regardless of where the buildings are.
Decision four: succession
This is where these structures actually fail. A vault holding with no documented succession is not an inheritance; it is a puzzle left to grieving people who do not know the vocabulary.
The minimum viable documentation: the holding entity or trust and its governing law; the custodian and account references; the auditor; who is authorised to instruct, and how that authority transfers on death or incapacity; the bar list with serials and assay certificates; the insurance policy and its limits; and the written rebalancing and disposal rules, so heirs are not left interpreting intent.
That package should live in at least two places, at least one outside the jurisdiction of the metal, and someone who is not the principal needs to know it exists. Review it whenever the family, the custodian or the law changes — which in practice means at least every couple of years.
The costs, compounded honestly
A generational position has a running cost and it deserves to be written down rather than assumed away. Allocated vault storage plus insurance typically runs in the region of a few tenths of a per cent a year, dealer spreads on physical bars are wider on the way in and out than on paper, small denominations carry a higher fabrication premium per ounce than large bars, and exchange-traded wrappers charge an expense ratio that compounds silently for as long as you hold.
Over thirty years those differences are not rounding errors. A twenty-five basis point annual drag removes roughly seven per cent of the terminal position; seventy-five basis points removes something closer to twenty. The point is not that the cheapest structure wins — a poorly held cheap position is worse than a well held expensive one — but that the cost has to be a decision rather than a discovery.
Liquidity planning for the people who inherit it
The event that forces a sale is usually a tax bill, not a market view. Inheritance and estate taxes in many jurisdictions fall due before assets can be freely realised, which turns an illiquid holding into a forced sale at whatever price the calendar offers. Families who plan for this hold a defined liquid sleeve sized against the estimated liability, keep at least one holding in a jurisdiction with a deep local dealer market, and document in advance which tranche gets sold first and on whose instruction.
The alternative — heirs selling the wrong bars in the wrong market under time pressure because nobody wrote down an order of operations — is the single most expensive administrative failure we see, and it costs more than any storage fee ever will.
Three failure cases we see repeatedly
The undocumented vault. Metal held in a jurisdiction the family understands, under an account nobody but the principal can instruct. The holding is real, the claim is valid, and it takes heirs eighteen months and a foreign lawyer to establish either.
The single-operator illusion. Four accounts, three countries, one vault operator and one insurance underwriter. Diversification on the statement, concentration in the risk.
The rebalancing drift. A core sized at eight per cent, never trimmed through a decade in which gold outperformed everything else, quietly becoming a twenty per cent position that no one ever consciously decided to hold. Drift is a decision made by inaction.
The bearer-metal assumption. Home-held metal treated as part of the estate plan without any register, serial list or valuation basis. There is no registrar to write to the heirs and no institution that knows it exists; if it is not written down somewhere else, it is not inherited, it is found or it is not.
The structure nobody else can operate. A multi-jurisdiction holding company arrangement built with a single adviser who has since retired, understood by one family member who has since died. Sophistication that cannot be administered by its inheritor is a liability wearing the costume of prudence.
The review cadence
A structure built once and never reviewed is a structure that slowly stops matching the family it belongs to. We suggest a light annual check and a full review every two to three years, or immediately on four triggers: a change in the family (birth, death, marriage, divorce, relocation), a change in the custodian or its ownership, a change in tax or reporting law in either the holding or residence jurisdiction, and a drift of the position more than a stated band away from its target weight.
The annual check is short: confirm the bar list still reconciles, confirm insurance limits still cover current value rather than the value at purchase, confirm the authorised signatories are alive, competent and reachable, and confirm at least one person other than the principal can still find the documentation pack.
The point of the exercise
None of this is exciting, and that is the whole argument for it. A well-built core is the part of a balance sheet that permits the rest to take risk, and the highest praise it can earn is that the next generation never has to think about it.