Competing trajectories affecting the price of Gold and Silver

Markets rarely move for one reason. Several forces act at once, some aligned and some opposed. Understanding a market is mainly a matter of sorting which ones carry weight.

That sorting is often done badly, because explanation drifts into prediction. A statement about what has happened becomes a statement about what will happen. The second kind claims knowledge no one holds. The future is unknown, and “if” is a way of speculating against that unknown instead of admitting it.

The approach set out here refuses the drift. It works only from the historical record up to the present, and it treats each significant force as a trajectory: a direction and a strength, traceable in data, that has held over an identifiable period.

Each trajectory is built from the record alone, without anticipation. Its evidence is shown. A trajectory may be strong or weak, structural or episodic. Its ending is always the same move: on the record available, this is what the force has been doing.

The forces are then read together. No single force explains a market; the relevant object is the pattern they make in combination, which are aligned, which offset, and which carry the greater weight. Weight here is not fixed. It shifts between periods, and a force that once explained most of the movement can decline to nearly none of it, as real yields did for gold after 2021.

The result is a statement of the weights as they stand, not a forecast, and of what the record shows when the weights sit in that arrangement. The record does not show whether the weights will hold, and the gaps it leaves remain unstated.

Prediction is refused at every step.

Trajectory 1 — Structural demand is rising

The most durable change in the gold market since 2020 sits on the demand side. It comes from buyers whose decision horizon is measured in years, and it is documented across three series: central bank purchases, the motives behind those purchases, and the investment flows that have followed them.

The central bank series marks the step change plainly.

Between 2010 and 2021, official buyers averaged 473 tonnes a year. In 2022 the run jumped to 1,136 tonnes, the largest annual total in records running back to 1950. The following two years held close to that level, and 2025, at 863 tonnes, was still the fourth-largest year on record.

Scale matters more than the path from year to year. Mine output reached roughly 3,670 tonnes in 2025, itself a record, so official buyers have absorbed roughly a quarter or more of annual mine output every year since 2022. The bid has also been one-sided: reductions remained rare and small, the largest reported cut of 2025 amounting to 15 tonnes. A buyer cohort that takes a large share of new supply and returns almost none of it changes the way the market clears.

The breadth of the bid is at least as telling as its size. Twenty-two institutions added a tonne or more in 2025. The National Bank of Poland was the largest buyer for the second consecutive year, adding 102 tonnes to reach 550 tonnes, 28 per cent of its reserves, and raising its target allocation to 30 per cent. The People’s Bank of China reported additions in every month of the year, lifting holdings to 2,306 tonnes, still under 9 per cent of its reserves. Kazakhstan posted its largest annual purchase on record, and Brazil returned to the market after a four-year absence, adding 43 tonnes between September and November.

Official buyers stepped back as valuations ran, but the bid did not reverse. Purchases fell 21 per cent in 2025 while the price set 53 record highs, yet the fourth quarter was the strongest of the year at 230 tonnes, executed at record price levels. The World Gold Council’s 2025 reserve survey drew 73 central bank responses, its broadest sample since the survey began, and found 95 per cent expecting global official gold holdings to rise over the following year and 43 per cent planning additions of their own.

The motive behind the bid is reserve diversification away from custody risk. In 2022 the foreign-held reserve assets of a central bank were immobilised by the jurisdictions where they were booked. Conventional reserve assets carry custody-jurisdiction risk that activates precisely in the scenarios reserves exist to cover. Allocated physical gold does not, because it is no institution’s liability and its value depends on no other balance sheet performing. PIMCO attributes part of the record official-sector accumulation since then to diversification away from that exposure, and the World Gold Council’s reserve surveys record the same motive set: diversification, performance in crisis periods, and gold’s standing as a politically neutral asset. In September 2026 China announced a plan to serve as custodian for foreign sovereign gold reserves, a move read as creating fresh demand from emerging-market holders seeking protection from sanctions, rather than shifting existing holdings out of London or New York. By the European Central Bank’s 2025 assessment, gold at market value had passed the euro to become the second-largest reserve asset globally, behind only the dollar, after fifteen consecutive years of net accumulation.

Much of this bid is not visible when it happens. Reported figures flow through IMF statistics and central bank disclosures, but the World Gold Council’s estimate of total official demand consistently exceeds what identified buyers account for. For 2025, roughly 57 per cent of estimated purchases could not be attributed to a disclosed buyer. Monthly statistics understate the bid in real time and revise history after the fact. A buyer that is large, persistent and partly invisible is what a pre-2022 fair-value model could not see, and it is the main reason those models now run a standing error.

Investment flows are the faster, more visible layer of the same trajectory. Physically backed gold ETFs took in US$89bn in 2025, the largest year on record in dollar terms, with holdings rising 801 tonnes, a tonnage gain second only to 2020. North American funds took US$51bn, close to 57 per cent of the total; Asian funds added US$25bn, more than all their prior years combined; European funds converted two years of outflows into US$12bn of inflows. Collective holdings ended near 3,930 tonnes, an all-time month-end high, and assets under management more than doubled over the year. Against a 2024 monthly average inflow of US$292mn, the 2025 pace ran an order of magnitude above it.

On the record available, the structural-demand trajectory points up. That is a description of documented behaviour, not a forecast. For four consecutive years official buyers have absorbed more than 800 tonnes annually, survey intent points the same way, and investment flows reached record scale in 2025. Whether the trajectory continues is a question the record cannot yet answer.

Trajectory 2 — Price is at extremes

The price record shows two features working at once. Gold has more than doubled from the levels reached in mid-2020, and it has corrected sharply several times on the way.

The year 2025 produced 53 record highs and a 44 per cent rise in the annual average price. From roughly $1,957 in July 2020, the spot price reached $5,405 in January 2026.

The correction is already on the record. From the January peak to the July low, the price fell about 26 per cent, more than a quarter, in roughly six months. The second quarter of 2026 was gold’s worst quarter since the second quarter of 2013. By early September the price had recovered to about $4,400, still around 19 per cent below the record and a little over 10 per cent above the July low.

The demand record shows what an extreme price does to different buyers. Jewellery, the price-elastic segment, contracted in 2025 as the price set records. Bar and coin purchases reached a twelve-year high over the same period: investment-driven buying persisted while price-sensitive consumption retreated. Both responses appeared in the same year.

Regional premiums printed the divergence. On the Shanghai Gold Exchange the premium over the international quote flipped to a discount through 2025 even as the global benchmark set records, a direct read of weak Chinese physical demand at high prices. India moved the other way, local premiums reaching $7 an ounce ahead of the festival season, the highest since November 2024, while local prices stood at all-time highs.

The fast layer matters at extremes because positioning is crowded. Options market activity was a leading contributor to gold’s 2025 return, and crowded long positions unwind quickly. The week around Jackson Hole in August 2026 took 4.72 per cent off the price. The record holds several such episodes: the price climbed and repeatedly cut back.

On the record available, the trajectory at these levels is volatile in both directions. Fifty-three record highs in one year, a drawdown of more than a quarter within the following six months, then a partial recovery. The corrections are documented, not hypothetical. The record does not show where the price settles next.

Trajectory 3 — Real yields remain elevated

Real yields are the force that used to drive the gold price, and they remain elevated. The 10-year Treasury nominal yield broke above 5 per cent in September 2026, a level last seen briefly in 2023 and before that in 2007. The inflation-adjusted 10-year real yield has held above 2 per cent.

The mechanism is opportunity cost. Gold pays no coupon and costs money to hold; the return a holder forgoes is measured against an asset with no credit risk, normally the 10-year TIPS yield. A basis point more real yield raises the carrying cost of a non-yielding position, and a basis point less lowers it. Research from the Federal Reserve Bank of Chicago formalises the logic: gold trades as a long-duration real asset, so its price carries a strong inverse relationship to long-term real rates.

For two decades that one variable did most of the explaining.

Through 2021 the channel was tight enough to build models on. Real yields fell across the financial crisis and stayed low or negative for much of the period, and gold rose with remarkable consistency: higher real yields meant weaker gold, lower or negative real yields meant stronger gold.

The break came in 2022. Real yields rose roughly 250 basis points that year, the largest annual increase on record against a prior maximum near 150, and the dollar gained more than 8 per cent. A model estimated on the previous two decades maps that combination to a deep drawdown in gold. The drawdown did not occur; gold closed the year with a small gain. The explanatory power of real yields collapsed to 3 per cent.

The collapse has persisted. Since 2024 the figure is about 7 per cent, with real yields above 2 per cent while the price kept rising. The level of the gold price is no longer anchored to the TIPS curve, and a fair-value model estimated on pre-2022 data runs a standing error.

The channel has not disappeared, only narrowed. Real yields still transmit at short horizons: a repricing of rate expectations moves the quote the same day, and the relationship reasserts when flow-driven demand goes quiet. The September 2026 break above 5 per cent on the nominal yield carried both ingredients raised at the outset: inflation concerns and expectations of tighter policy on one side, questions about bond-market confidence on the other. A yield rise driven by growth tightens the opportunity-cost channel; a yield rise that itself expresses the risk being hedged does not.

On the record available, real yields exert downward pressure on gold, and that pressure is weaker than at any point in the measured history. The force is real, operates at the edges, and has not governed the price level since 2022.

Trajectory 4 — Supply grows slowly

The supply side of the gold market is the slowest-moving force on the list, and the 2025 record shows why it cannot arbitrage a demand shift.

Mine production reached a record of roughly 3,672 tonnes in 2025, yet total supply grew only about 1 per cent over the year.

The gap is structural. New mining capacity takes years to bring on, so a demand shift cannot be answered with new ounces on any short horizon. The one fast-responding component is recycling, and recycling is bounded by something outside the miner’s control: how much of the existing stock holders are willing to release at a given price.

Price therefore clears demand against a nearly fixed short-run supply. Demand can move in weeks; mine supply cannot. That imbalance would matter less if new mine output were flowing freely to the market, but since 2022 official-sector buyers have absorbed roughly a quarter or more of annual mine output each year, and the bid is one-sided. The effective quantity of new metal available to the wider market is smaller than the production record suggests.

On the record available, supply is growing slowly and inelastically. It does not offset the demand trajectories already documented; at any horizon the market actually trades on, supply follows the price rather than setting it.

Trajectory 5 — Geopolitical risk is elevated

Geopolitical risk is the force whose direction the record cannot project. It is elevated now, it explains a leading share of gold’s recent return, and its timing is unforecastable.

The demand it generates references balance-sheet security rather than yield or price. Gold is no institution’s liability: it has no issuer and no counterparty, and its existence depends on no payment or settlement infrastructure. Held in allocated physical form, it is the one reserve asset whose value does not require another balance sheet to perform.

In 2022 those properties acquired a price. The foreign-held reserve assets of a central bank were immobilised by the jurisdictions where they were booked. The episode changed reserve management because it exposed a structural feature: conventional reserve assets carry custody-jurisdiction risk that activates precisely in the scenarios reserves exist to cover. PIMCO attributes part of the record official-sector accumulation since then to diversification away from that exposure, and the World Gold Council’s reserve surveys record the same motive set: diversification, crisis performance, and gold’s standing as a politically neutral asset.

The current elevation is documented. The World Gold Council’s December 2025 commentary attributed roughly 60 per cent of gold’s full-year return to explicit model variables, led by geopolitical risk and options market activity, particularly from August to October. The World Bank, in the same month, described gold’s record highs of October 2025 as coming amid rising geopolitical tensions and strong investor demand, supported by central bank purchases.

The World Gold Council also records the episodic pattern across cycles: the three strongest years for gold ETF demand on record each coincided with a period of acute global stress.

The timing is what sets this force apart. Hedging demand arrives as a step increase when a triggering event lands, shows up in flow and premium data within days, and can leave as quickly when the event recedes. No historical series predicts the next trigger; the record shows only that the episodes come in clusters and that gold responds to them.

On the record available, geopolitical risk is high and its trajectory is unforecastable. It speaks to the direction of gold demand only while the risk is live, which is why it sits alongside the structural drivers without matching any of them in duration.

Trajectory 6 — Silver carries industrial exposure

Silver is the metal whose trajectory diverges most clearly from gold’s. Tthe reason is structural: roughly half of silver demand is industrial, while gold’s industrial share is a small fraction of its total.

The 2025 record shows the difference in magnification. Silver rose about 147 per cent over the year, against a smaller rise in gold, and the spot price sat near $63 an ounce in September 2026 after trading in the low $30s a year earlier. The leverage runs both ways: corrections of 15 to 25 per cent are normal within silver’s bull moves.

The drivers explain the divergence. The central bank bid that anchors gold is nearly absent in silver; official reserves hold negligible silver. What anchors silver instead is the energy and electronics demand behind the industrial half: solar photovoltaics, electrical contacts, and the manufacturing cycle more broadly. When that cycle runs, silver gains more than gold; when it slows, silver falls harder.

The safe-haven role also operates differently. Silver receives a safe-haven bid in stress periods, but a severe slowdown cuts industrial demand at the same time as fear demand rises, leaving two forces pulling in opposite directions. Gold does not face that collision, because reserve and investment demand dominate its total.

On the record available, silver’s trajectory is tied to the industrial cycle in a way gold’s is not. That gives silver a higher ceiling in booms and a lower floor in downturns, and it makes the two metals’ trajectories correlated, not interchangeable.

Synthesis — the six trajectories read together

The six trajectories describe one market from six angles. Read together they explain the shape the price record already shows: an upward path with violent corrections.

Three forces point the same way on demand. Structural central bank buying has absorbed roughly a quarter or more of annual mine output every year since 2022. De-dollarisation and geopolitical hedging feed the same demand and move investment flows with it. The three read as one chain: the 2022 reserve freeze exposed custody risk, which drove reserve diversification, which became the structural bid, which pulled record ETF inflows behind it.

The counterweights are uneven. Real yields, the variable that once explained 84 per cent of gold’s price variance, now explain about 7 per cent and have not governed the level since 2022. Supply grows too slowly to offset a demand shift on any horizon that matters. The one live downward force is the price itself: at record levels, price-sensitive buyers retreat and crowded positions unwind.

The net reading follows. The aligned demand forces set the direction; the weakened headwinds and extreme valuations set the character of the path.

The record shows fifty-three record highs in 2025, a drawdown of more than a quarter into July 2026, then a partial recovery. The synthesis adds one thing: on the record available, the upward forces are stronger and more persistent than the downward ones, so the price trends up, while corrections come fast and can be large.

The record does not show whether those weights hold. Central bank buying has decelerated. Real yields could reassert if the regime shifts. The next geopolitical trigger has no series that predicts it. The synthesis states the weights as they stand now and stops there.