If you follow gold commentary you will hear that gold "hedges inflation", "moves against the dollar" and "falls when rates rise". These are all partly true and all slightly misleading, because they describe symptoms of one underlying mechanism.

The mechanism: gold has an opportunity cost

Gold pays nothing. It generates no interest, no dividend and no rent. A bar sitting in a vault produces exactly one bar of gold a year later, minus storage.

Every other liquid store of value competes with that by paying a yield. So the real question an investor faces is never "is gold good?" but "what am I giving up by holding gold instead of something that pays?"

That giving-up is gold's opportunity cost, and the thing that determines it is the real interest rate — the nominal rate minus expected inflation.

  • When real rates are high, safe bonds pay a meaningful return after inflation. Holding gold costs you that return. Gold becomes less attractive.
  • When real rates are low or negative, those bonds lose purchasing power. Gold's zero yield stops being a disadvantage — a zero return beats a negative one.

This is why the strongest observed relationship in gold is not with inflation, or growth, or the stock market, but with real yields, and it runs inverse: real yields down, gold up.

Why "gold hedges inflation" is only half right

Inflation matters, but through real rates rather than directly.

If inflation rises and central banks raise nominal rates faster, real rates go *up* and gold can fall despite high inflation. If inflation rises and rates lag behind it, real rates go *down* and gold typically rises.

So the accurate statement is narrower: gold tends to do well when inflation is running ahead of interest rates. Not simply "when there is inflation". Periods where high inflation coincided with aggressive rate rises have been poor for gold, which is exactly what the mechanism predicts.

Why the dollar matters

Gold is quoted in US dollars worldwide. That produces two distinct effects, and they are often conflated.

The arithmetic effect. If the dollar strengthens against other currencies and gold's dollar price is unchanged, gold has become more expensive for buyers in those currencies — which tends to soften demand. The reverse holds too.

The shared-driver effect. The same conditions that lift real US yields typically also strengthen the dollar. So the dollar and gold often move oppositely not because one causes the other, but because both are responding to rates.

This is why the same metal can be at a record high in one currency while well below its peak in another. A currency that has weakened against the dollar shows a higher local gold price for exactly that reason — the metal did some of the work, and the currency did the rest.

The other real drivers

Real rates explain a great deal but not everything.

Central bank demand. Central banks hold gold as reserves, and sustained official buying is genuine, price-insensitive demand that does not respond to yields at all.

Crisis and risk demand. In periods of financial stress or geopolitical shock, gold attracts flows regardless of what rates are doing. These moves are sharp, and they are the ones that break the rate relationship for a while.

Jewellery and physical demand. Seasonal and cultural buying — wedding seasons in South Asia, festival demand — is a large share of physical consumption. It shapes local premiums more than the global price, but it is not nothing.

Supply. Mine production changes slowly and recycling responds to price with a lag, so supply rarely drives short-term moves. Over a decade it matters more.

What this means practically

If you want to understand a move in gold, the first question worth asking is what happened to real yields — not what the headline said.

And it explains why gold's behaviour looks inconsistent to casual observers. It rose during some inflationary periods and fell during others; it rallied in some crises and not others. That is not gold being random. It is one variable — the real cost of holding a non-yielding asset — being pushed around by several forces at once.

None of this predicts anything. Understanding the mechanism tells you what to watch, not what will happen. Our projection pages model a range of outcomes from measured volatility precisely because the direction is not knowable.