The claim that central banks are buying gold faster than miners can produce it, thereby driving prices, is wrong. Official-sector demand creates a durable floor. Monetary conditions determine price direction.
Official-Sector Demand Is Fragmenting
Central banks bought more than 1,000 tonnes annually from 2022 through 2024, the first three-year run above that level since the 1950s. Purchases fell to 863 tonnes in 2025, while the World Gold Council forecasts roughly 850 tonnes for 2026.
The structural change is in who is buying. China, Russia, Poland, and Kazakhstan dominated the 2022–2024 cycle. By the first half of 2025, purchases had spread across 23 central banks.
| Buyer | Reported activity |
|---|---|
| Poland | 31 tonnes in Q1 2026; targeting 700 tonnes total |
| Uzbekistan | 25 tonnes |
| Kazakhstan | 12 tonnes |
| China | 8 tonnes in April; 18 consecutive months of buying |
| Czech Republic | 38 consecutive months of buying |
| Malaysia | 3 tonnes in January; gold reached 5% of reserves |
| New entrants | Chile, Guatemala, Bolivia, Uruguay, and Uganda |
The WGC’s 2026 survey found that 89% of central bankers expect global gold reserves to increase over the next year, while 45% expect their own holdings to grow.
This is no longer concentrated geopolitical hedging. Gold is becoming a baseline reserve allocation in a multipolar system.
The market remains two-way. Russia sold 34 tonnes through June 2026, reducing holdings to 2,292 tonnes. Turkey sold 81 tonnes. Former aggressive buyers are now providing liquidity. The official sector is continuously reallocating, not buying without limit.
The Supply Constraint Is Geological
Central banks have absorbed roughly 23–30% of annual mine output since 2022. Miners still produce about 3.5–4 tonnes for every tonne purchased by central banks.
| Year | Mine production | Central-bank purchases | Purchases as share of production |
|---|---|---|---|
| 2022 | 3,612 t | 1,082 t | 30.0% |
| 2023 | 3,644 t | 1,037 t | 28.5% |
| 2024 | 3,661 t | 1,045 t | 28.5% |
| 2025 | 3,672 t | 863 t | 23.5% |
Total supply reached a record 5,002 tonnes in 2025, including record mine production of 3,672 tonnes. Yet production barely moved from 3,644 tonnes in 2023 and 3,661 tonnes in 2024, despite gold prices rising nearly 70% year over year.
The constraint is geological inelasticity: declining ore grades, permitting friction, aging mines, and multi-year development cycles. Higher prices cannot quickly produce new deposits.
Country-level disruptions expose the bottleneck:
- Grasberg in Indonesia lost 24% of output after a Q3 2025 mud-rush incident.
- Mexico fell 18% after Equinox Gold suspended Los Filos over a land-access dispute.
- Argentina dropped 27% as Barrick’s Veladero encountered lower-grade ore.
Recycled supply grew only about 3% despite record prices. Retail holders are retaining jewelry and scrap, signaling expectations of further upside.
Demand Is Larger Than Central Banks
Total gold demand exceeded 5,000 tonnes for the first time in 2025.
Investment demand from ETFs and bars and coins reached 2,175 tonnes, up 84% year over year and more than twice central-bank demand. Jewelry and industrial use added roughly 1,200 tonnes. With mine production effectively flat and recycling barely rising, the market is structurally tight.
Central banks matter because their purchases are persistent and relatively price-insensitive. Reserve managers operate under multi-year mandates. Their demand sits beneath a larger investment market and an inelastic supply base, supporting prices without controlling them.
Why Gold Fell 27% in 2026
Gold rose 65% in 2025 and registered 53 record highs. The rally continued to nearly $5,600 per ounce on January 29, 2026. By late July, gold had fallen roughly 27–29% to about $4,000–$4,070.
The reversal was monetary, not physical.
Trump’s nomination of Kevin Warsh to succeed Jerome Powell as Federal Reserve chair reduced expected rate cuts to roughly 50–55 basis points for the year. Real 10-year Treasury yields moved toward 1.9%, versus below 1% during the rally. The dollar strengthened. A May jobs report showing 172,000 jobs against 85,000 expected further reduced near-term easing expectations and briefly erased gold’s 2026 gain.
The causal chain is direct:
| Monetary variable | Transmission to gold |
|---|---|
| Higher real yields | Raises the opportunity cost of holding a non-yielding asset |
| Fewer expected Fed cuts | Extends restrictive monetary conditions |
| Stronger dollar | Increases gold’s cost for non-dollar buyers |
| Strong labor data | Reduces pressure for near-term easing |
Physical fundamentals establish the long-run floor. Real yields, Fed expectations, and the dollar control short-run prices.
Positioning
Gold is a monetary hedge with a physical supply constraint, not a simple commodity-shortage trade.
Broadening central-bank demand, reserve diversification, and geological supply limits support the long-term case. They do not prevent a 27% quarterly decline when real yields rise and policy expectations reset.
Mid-2026 institutional forecasts expose the monetary uncertainty:
| Scenario or institution | Gold outlook |
|---|---|
| JPMorgan | $6,300 by year-end |
| Deutsche Bank | Above $6,000 |
| UBS | $6,200 in Q1–Q3; about $5,900 by year-end |
| Hawkish-Fed bear case | $3,800–$4,600 |
The $3,800–$6,300 range reflects divergent monetary-policy paths, not disagreement over mine geology.
The gold reallocation is real. It establishes a structural demand floor. It does not set the marginal price. Position around real yields, Fed policy, and dollar strength accordingly.