What Actually Drives the Gold Price

The Core Model: Gold as a Zero-Coupon, Infinite-Duration Bond

The most influential academic framework for gold pricing comes from Robert Barsky and Lawrence Summers' 1988 paper "Gibson's Paradox and the Gold Standard" (Journal of Political Economy). They proposed treating gold as analogous to a very long-duration, zero-coupon, riskless bond. Because gold produces no income — no dividend, no coupon, no rent — the only cost of holding it is the yield you forgo by not holding an interest-bearing asset instead. That forgone yield should be measured in real terms, because gold, like any physical commodity, should track the general price level over the very long run.

This single idea explains most of what looks confusing about gold's short-term behaviour. Rising inflation does not mechanically push gold up — rising nominal rates that outpace inflation (i.e. rising real rates) push gold down, and falling real rates push gold up, regardless of the inflation print itself. That is why gold fell through much of 1980–2000, a period of generally positive but declining inflation, while Paul Volcker's Federal Reserve pushed real interest rates sharply positive — and why gold surged from 2001 to 2011 and again from 2018 to 2020 as real rates fell toward and through zero.

The Empirical Record: Real Rates vs Gold, Cycle by Cycle

The theory is only useful if it survives contact with the historical data. It largely does — with the clearest test being the introduction of Treasury Inflation-Protected Securities (TIPS) in 1997, which gave researchers a direct, market-priced measure of real interest rates to compare against gold in real time.

The 2022 episode is the most interesting deviation from the simple model, and it points to the second major driver discussed below. Real rates rose faster than in almost any prior period, which the Barsky-Summers framework predicts should have hit gold hard — yet gold ended 2022 roughly where it started. The most commonly cited explanation, including from the World Gold Council's own research, is a structural shift in official-sector demand that began offsetting the usual rate sensitivity.

The Second Driver: Official-Sector (Central Bank) Demand

Central banks became large, persistent net buyers of gold starting around 2010, reversing decades of net selling that followed the end of Bretton Woods. Purchases accelerated sharply from 2022 onward, with the World Gold Council recording central bank demand above 1,000 tonnes per year in 2022, 2023, and 2024 — roughly double the pre-2022 decade average. This is a large enough flow, relative to annual mine supply of approximately 3,000–3,700 tonnes, to move price independent of the real-rate cycle.

The dominant explanation in the research and among reserve managers themselves is de-dollarisation risk management following the freezing of a portion of Russia's foreign exchange reserves in 2022 — an event that made visible, to every central bank holding dollar reserves, that those reserves carry geopolitical counterparty risk that gold, held domestically, does not. The World Gold Council's annual central bank surveys since 2022 consistently rank "geopolitical risk" and "reserve diversification" as the top two stated motivations for gold purchases, ahead of inflation hedging.

The Dollar Channel

Gold is priced globally in US dollars, which creates a mechanical inverse relationship independent of any macro story: when the dollar weakens against other major currencies, gold becomes cheaper for non-dollar buyers, demand tends to rise, and the dollar price of gold tends to rise to compensate — and vice versa. Researchers typically measure this against a trade-weighted dollar index (such as the DXY or the Federal Reserve's broad dollar index). The historical correlation between gold and the DXY has averaged around -0.4 to -0.6 over rolling multi-year windows, meaningful but weaker and less consistent than the real-rate relationship — the dollar channel amplifies or dampens moves more often than it initiates them.

Crisis and Tail-Risk Demand

Gold's behaviour during acute market stress is well documented and distinct from its behaviour in normal conditions. During the 2008 Global Financial Crisis, gold fell alongside almost every other asset in the initial liquidity-crunch phase of September–October 2008 (forced selling to raise cash), then rallied strongly through 2009–2011 as the policy response — near-zero rates and quantitative easing — pushed real yields deeply negative. During the COVID-19 shock of February–March 2020, gold likewise dipped briefly on a dash for cash before rallying to a then-record high by August 2020.

What the Data Does Not Support

Two commonly repeated claims do not hold up well under scrutiny. First, the idea that gold moves in lockstep with headline CPI inflation is not well supported over periods shorter than a decade — see our dedicated gold and inflation article, which covers the Erb and Harvey (NBER, 2013) research on this specifically. Second, the idea that broad money-supply growth (M2) directly and immediately drives gold prices is intuitive but empirically weak on its own; money-supply growth matters mainly to the extent it shows up in the real-rate and inflation-expectation channels above, not as a standalone predictor with a short, tradable lag.

Putting It Together

No single-variable model fully explains gold's price. The most defensible synthesis from the academic and institutional research is a layered one: real interest rates set the macro backdrop and explain the largest share of medium-term variance; official-sector (central bank) demand has become a large enough flow since 2022 to structurally offset or reinforce that backdrop; the dollar channel amplifies moves that originate elsewhere; and crisis demand adds a distinct, policy-response-driven layer on top, rather than a simple instantaneous flight-to-safety spike.