Gold is a political instrument that happens to trade

Gold gets modelled two ways, and both of them underperform.

The commodity framing treats it as a physical good with supply, demand and inventories. But almost all the gold ever mined still exists, annual production is small against the standing stock, and industrial consumption is a rounding error in the price. The supply-demand apparatus that works for copper has very little to grip on.

The inflation-hedge framing is more popular and empirically weaker. Over long enough horizons the relationship is defensible. Over the horizons anyone actually holds a position, gold has spent extended periods rising into disinflation and falling into inflation, and the framing survives mainly because people remember the episodes where it worked.

There is a third description that explains considerably more, and it is not a financial one.

The asset that is nobody's liability

Every other major reserve asset is someone's obligation. A Treasury is a claim on the United States government. A bund is a claim on Germany. A deposit is a claim on a bank. Each carries the credit of its issuer, and each exists inside an infrastructure that the issuer and its allies can reach. That second part is the operative one.

Gold is the only reserve asset of any size that is not a claim on anyone. Held domestically, it cannot be frozen by another state's decision, because there is no counterparty to instruct. That property is irrelevant almost all of the time and decisive occasionally, and central banks are institutions that plan for the occasional case.

Which means the largest and least price-sensitive source of demand in the gold market is not making a trade. It is making a policy decision about the resilience of its own reserves, on a horizon measured in decades, with the price as a secondary consideration at most.

Every other reserve asset is a claim on someone. That is the entire thesis, and everything else about gold's behaviour follows from it.

Why the correlations "break"

Once official demand is understood as policy-driven rather than price-driven, the recurring puzzle in gold commentary resolves itself.

The standard model says gold should fall when real rates rise, because gold pays no yield and the opportunity cost of holding it goes up. This is sound reasoning about a marginal financial buyer, and it holds when the marginal buyer is one. When a meaningful share of demand is a reserve manager acting on a diversification mandate, real rates are simply not the variable that decides the purchase. The relationship does not break; it was never load-bearing in the first place.

The same logic covers the other recurring surprise: gold rising in periods with no visible inflation problem. If the driver is confidence in the reserve system rather than the price level, then episodes that raise questions about whether reserves are safe where they are held will move gold regardless of what inflation is doing.

What it is a claim about

Read this way, gold is not an inflation hedge, an equity hedge or a recession hedge, and it has disappointed everyone who bought it as one. It is a position on whether the arrangement by which reserves are held somewhere and remain usable continues to work as assumed.

That risk is real, currently non-trivial, and difficult to hedge any other way, which is a reasonable case for owning some. It is also a risk that, if it never materialises, produces decades of an asset yielding nothing and underperforming almost everything.

Which is the awkward property of hedging a systemic assumption. Held under an accurate description, gold does exactly what it is for and looks like a failure most of the time. Held under the wrong description, it looks like a failure and gets sold at the point the description would have started to matter.