There are two honest ways to talk about the gold price. The first is to name a number and hope. The second is to name the mechanisms, describe how they interact, and state plainly what would prove you wrong. This edition takes the second route, and its subject is the shift that has done most to confuse forecasters since 2022: the arrival of the official sector as a persistent, price-insensitive buyer alongside the traditional macro driver of real interest rates. Nothing here is investment, tax or legal advice.
The old sentence: real yields, not inflation
The pre-2022 relationship was durable enough to be treated as a rule. Gold pays no coupon, so its opportunity cost is the real yield available on a safe alternative. When inflation-adjusted yields on US Treasuries fall, holding a non-yielding reserve asset costs less and the price tends to rise; when real yields climb, the reverse. This is why gold frequently disappoints during high-inflation episodes in which central banks raise policy rates aggressively — nominal inflation is not the mechanism, the real rate is.
That relationship has not been repealed. It still explains most of the week-to-week and quarter-to-quarter variance, and it is still the cleanest way to understand why speculative and investment demand ebbs and flows. What changed is that the residual — the part real yields could not explain — stopped being noise and started having a direction.
The new clause: the official sector
Central banks have always held gold. What is different in this cycle is the composition and the motive of the buying. Reserve managers in emerging and middle-income economies have been accumulating bullion as a share of reserves at a pace that has no modern precedent, and they have been doing it for reasons that are indifferent to price.
A commercial buyer asks whether gold is cheap. A reserve manager asks whether a reserve asset can be frozen, sanctioned, defaulted on or inflated away by a foreign legislature. Gold's answer to all four questions is unusually attractive: it is a bearer asset with no issuer, no counterparty and no jurisdiction, provided it is held domestically. That is not a view on the gold price. It is a view on the reliability of the alternatives.
The consequence for market structure is subtle but important. Price-insensitive buying does not create a trend on its own — the flows are too small relative to the total above-ground stock to overwhelm investment demand. What it does is absorb supply on weakness. It raises the floor. Periods that in an earlier cycle would have produced a deep drawdown now produce a shallower one, because a bid arrives that does not care about momentum.
Two questions before you quote a tonnage figure
Official-sector numbers are the most casually misused statistics in this market, so we hold ourselves to two checks before repeating any of them. First: is the figure reported to the IMF or estimated? A meaningful share of accumulation in recent years has been inferred rather than declared, and inference is a range, not a datapoint. Second: is it a purchase or a repatriation? Metal moved from a foreign vault to a domestic one changes the custody risk profile without changing the quantity of gold anyone owns, and it is regularly reported as if it were demand.
Apply both filters and the picture stays strong but becomes less dramatic than the headlines. That is usually the correct outcome of doing the work.
What China actually tells us
China is the case study everyone reaches for and few handle carefully. Its official reserves remain a modest share of total reserves by the standards of European holders, its disclosure cadence has varied, and its domestic market absorbs large volumes of privately held metal that never appears in official accounts. The reasonable reading is directional rather than quantitative: a large reserve holder with an unusually low gold share, an explicit interest in reducing exposure to a single foreign balance sheet, and the fiscal capacity to buy for a very long time without needing a favourable entry price.
The unreasonable reading is a specific tonnage target on a specific date. We have not seen a credible one, and we would treat any confident claim to the contrary as marketing.
De-dollarisation, precisely
The term is used to mean two very different things, and conflating them produces bad analysis. The strong version — the dollar ceasing to be the primary invoicing and reserve currency — has little supporting evidence and no plausible near-term mechanism, because no alternative offers comparable market depth or legal infrastructure. The weak version — reserve managers reducing the share of reserves held in dollar-denominated claims while the dollar remains dominant — is measurable, gradual, and already underway.
Only the weak version is needed for the gold argument to work. A few percentage points of reallocation across the global reserve pool is an enormous flow relative to annual mine supply, and it can proceed for a decade without anything resembling a monetary regime change. Investors who need the strong version to be true are, in our view, holding a correct position for an incorrect reason.
How the two drivers interact
The useful mental model is a floor and a ceiling. Official-sector accumulation sets the floor: persistent, slow, insensitive to price, and unlikely to reverse quickly because the motive is strategic rather than tactical. Federal Reserve policy sets the ceiling: while real yields are positive and the policy stance restrictive, investment demand faces a genuine opportunity cost, and rallies meet resistance.
Both can be true simultaneously, which is precisely why the last few years have frustrated single-driver forecasters. A market with a rising floor and a firm ceiling produces higher lows without producing runaway highs — until the ceiling lifts. The asymmetric case for gold is not that official buying will drive the price; it is that official buying has changed the shape of the downside while leaving the upside dependent on a policy cycle that will eventually turn.
The four series we track
We keep the dashboard deliberately small and entirely public. US ten-year Treasury inflation-protected security yields, as the cleanest proxy for real opportunity cost. IMF International Financial Statistics reserve tables, for declared official holdings by country. Reported holdings of the major physically backed exchange-traded products, as the best available high-frequency read on Western investment demand. And regional physical premiums or discounts in the large consuming markets, which reveal whether physical demand is confirming or contradicting the paper market.
Four series, all free, all published on a stated cadence. If a fifth is needed to make an argument work, the argument is usually the problem.
What would prove this wrong
A framework that cannot be falsified is a slogan, so here are the conditions we would treat as genuine refutation. Sustained net official-sector selling across multiple unrelated reserve managers over consecutive quarters, which would mean the strategic motive has weakened. A durable regime of high positive real yields alongside a rising gold price, which would mean the opportunity-cost mechanism has broken and our ceiling logic is wrong. Persistent deep physical discounts in the major consuming markets during a price rally, indicating the flows are entirely financial. Or a credible, liquid, sanction-resistant reserve alternative emerging, which would remove gold's distinctive advantage rather than merely dilute it.
We review these conditions quarterly and will say so in this newsletter if one of them starts to bite.
On the programme
Reserve accumulation, the official-sector data problem and the mechanics of monetary diversification are the spine of the central banking and reserves track at the Congress on 10 October. Sessions are curated first; speakers are announced individually as contracts complete, and we do not list names before then.
Housekeeping
General Admission remains free and includes every live session plus ninety days of replay access. Pro and VIP places are arranged by manual invoice after registration. This edition is editorial and educational only — not investment, tax or legal advice, and no part of it is a recommendation to buy or sell any asset.