Why Central Bank Gold Demand No Longer Sets the Price

Gold just posted its worst weekly stretch in months, sliding through a third consecutive weekly decline. In the same window, China’s central bank recorded its heaviest single-month gold purchase since late 2023.

That dissonance is the puzzle worth sitting with. A softening price should, on the simplest reading, cool the appetite of the world’s most price-sensitive buyers. Instead, the People’s Bank of China (PBoC) has now bought gold for 22 consecutive months, adding 3.93 million troy ounces to reach a record 76.73 million troy ounces. This is not an isolated impulse; it sits against a wider backdrop where aggregate central bank gold demand has moderated globally, and where JPMorgan has flagged exchange-traded fund (ETF) inflows as the new marginal force setting the price.

Two forces are pulling in opposite directions. If sovereign buying is structurally positive but no longer the dominant price driver, where does that leave your positioning? That is the question this piece is built to answer, and the answer starts in Beijing.

China’s 22-month accumulation streak just reached a new high

In August 2026, the PBoC added 650,000 troy ounces to its reserves, roughly 20.2 tonnes. It was the largest single-month addition since late 2023, and it extended a buying streak that resumed in November 2024 and has run unbroken ever since.

Take that streak apart and the scale becomes harder to dismiss. Twenty-two months of consecutive purchases have delivered a cumulative 3.93 million troy ounces into China’s reserves. The pace has not depended on gold being cheap; the buying continued through rallies and through the recent decline alike.

That price-indifference is the tell. This is not a central bank timing entry points around the edges of its reserves. It is executing a programme large enough to shift its monetary exposure in a deliberate direction, and analysts broadly read that direction as a structural reconfiguration rather than opportunistic accumulation.

The rationale, on the consensus view, rests on three connected motives:

• De-dollarisation: reducing reliance on US dollar assets as a share of total reserves
• Renminbi internationalisation: building a gold backdrop that supports the currency’s wider global role
• Sanctions insulation: holding an asset that cannot be frozen or seized by another government

Some commentators note a tactical thread running through it too, pointing out that the PBoC has taken advantage of price weakness before. The dominant framing, though, is that long-term monetary autonomy overrides short-term price speculation for a buyer of this kind.

That distinction matters more than it first appears. Structural programmes do not reverse on price signals, which makes this demand price-insensitive in a way that ETF flows never are. For anyone trying to model a floor under the gold price, that is the difference between a durable support and a fickle one.

The three motives behind China’s buying, reduced dollar dependency, renminbi internationalisation, and sanctions insulation, are not uniquely Chinese; the de-dollarisation trends reshaping reserve allocation across emerging markets in 2026 reflect the same structural logic operating at sovereign level well beyond Beijing.

Anchor figure Following the August purchase, China’s official gold holdings reached a record 76.73 million troy ounces, the highest level the PBoC has ever reported.

The record holding tells you China is not hedging cautiously. It is committing at a scale that meaningfully changes its exposure, and scale of that kind sets a floor that does not flinch when the price does.

What Turkey’s selling streak reveals about the other side of official-sector behaviour

If China is the case for structural accumulation, Turkey is the case that complicates it. In August 2026, Turkey’s gold position fell by 8 tonnes, the steepest monthly decrease since November 2021, according to Bloomberg. That is the mirror image of Beijing’s programme, and it happened in the same month.

The driver was domestic. QNB economists attribute the August decline to domestic investors treating elevated gold prices as a selling opportunity, with the central bank’s position shifting alongside that behaviour. Turkey has long used gold sales as a lever to counteract downward pressure on the lira, and 2026 has followed that pattern: the country sits among the largest year-to-date official sellers, at roughly 83 tonnes of net sales as of June 2026.

A brief note on the numbers is worth flagging. Sources diverge on Turkish gold volumes, largely because total national stock figures and official central bank reserve figures are measured differently. The official reserve framing is the one used here, and even within it the picture is not one-directional.

Because despite the August drop, Turkey’s holdings remained roughly 95 tonnes higher on a 12-month basis. A country can be a net accumulator over the year and a sharp seller in a single month, and both facts can be true at once.

The lira pressure that flips policy direction

The mechanism behind Turkey’s selling is domestic monetary stress rather than any change of heart about diversification. When the lira comes under pressure, gold sales become a tool for managing the currency, and long-term reserve intentions give way to immediate fiscal need.

What Turkey tells you is that sovereign gold demand comes with political and fiscal conditions attached. A country’s trajectory can flip from major accumulator to major seller inside a single year when domestic stress overrides diversification intent. That makes the aggregate official-sector figure a composite of individual decisions that can diverge sharply, and a less reliable single signal than the headline “central banks are buying” implies.

The aggregate official-sector figure is a composite of individual sovereign decisions, each shaped by distinct central bank reserve functions including currency stabilisation, sanctions insulation, and long-term diversification, which is why the headline ‘central banks are buying’ can obscure directional divergence as sharp as the China-Turkey split visible in August 2026.

How ETF inflows displaced central banks as the marginal price driver

Here is where the mechanics of the market have quietly shifted. Even as China buys at pace, JPMorgan’s analysis identifies ETF demand, not sovereign demand, as the force currently setting the price at the margin.

The flows explain why. Roughly $6.5 to $7 billion in net inflows moved into GLD, the largest gold-backed ETF, over the two months preceding mid-September 2026. According to JPMorgan, that two-month figure is the largest since November 2025.

Set against the price action, the pattern sharpens. Spot gold traded at $4,302 per ounce on 15 September 2026, after briefly touching $4,360 the previous Friday and then rolling into its third consecutive weekly decline. Central banks kept buying through all of it; the price still turned. That is the clearest sign that the near-term direction is being set elsewhere.

The two forces resolve into a dual-driver model worth holding onto:

1. The central bank floor. Consistent, price-insensitive sovereign buying, of the kind China is executing, establishes a structural baseline beneath the price. It does not chase rallies and does not flee declines.
2. The ETF accelerant. Rate-sensitive fund flows raise the ceiling when money moves in, but those flows reverse quickly when the macro backdrop shifts. They are the marginal price-setter, and they are the fragile layer.

Headline flow GLD attracted roughly $6.5 to $7 billion in net inflows over two months, its heaviest such stretch since November 2025.

JPMorgan’s framing is that rate-sensitive ETF demand grows more significant precisely as central bank buying moderates. The bank’s modelling reportedly suggests every 100 tonnes of committed-buyer demand generates around a 1.7% price increase, though that specific sensitivity figure is unverified and worth treating with caution.

The rate sensitivity risk that makes the ceiling fragile

The vulnerability is straightforward once you see it. If rate expectations shift toward higher-for-longer, ETF flows can reverse fast, and gold’s upside depends on those expectations staying accommodative.

That reversal risk amplifies downside volatility even when central bank demand holds perfectly steady. The floor does not move, but the ceiling can drop out from under the price on a single macro surprise.

The rate-sensitivity risk is compounded by the fact that ETF demand is not monolithic; regional ETF flow divergence across North American, European, and Asian funds in 2026 means a single macro surprise can simultaneously trigger outflows in rate-sensitive Western funds while Asian allocations remain sticky, adding a geographic layer to the fragility the dual-driver model identifies.

The practical read is asymmetric. The floor is durable because sovereign demand is structural; the ceiling is fragile because ETF flows are rate-sensitive and quick to unwind. That asymmetry is where the real risk in a gold position now sits.

What the historical arc of central bank buying tells investors about where this cycle sits

Step back far enough and the current wave stops looking like a novelty. It is the latest phase in a decades-long reversal of how central banks treat gold.

The broad sequence, though several of these figures are estimates rather than independently confirmed, runs like this:

• Net seller era (broadly 1989 to 2009): central banks were, on balance, offloading gold from reserves.
• Transition to net buyer (from 2010): the balance flipped, and central banks became consistent accumulators.
• Post-2022 acceleration: amid geopolitical fragmentation, annual net official purchases are estimated to have surpassed 1,000 tonnes.

The precedent for today’s sanctions-driven buying is Russia. Following sanctions after 2014, Russia accelerated its purchases sharply, reportedly amassing around 2,299 tonnes and taking gold to roughly 24.5% of total reserves.

Precedent at scale Russia’s post-2014 accumulation, estimated at around 2,299 tonnes, shows what sanctions-insulation buying looks like when a state commits to it over years rather than months.

That is the same logic now operating in China, only at larger scale and with greater geopolitical weight. The motive that drove Russia to build a sanctions-resistant reserve is the motive analysts attach to Beijing today.

The historical arc cuts both ways for investors. It confirms the structural case is real and has extended precedent, but it also shows these waves are conditions-dependent. Turkey’s 2026 reversal is the live reminder that sovereign accumulation can plateau and turn when domestic pressures override long-term intent.

The conditions driving China’s programme, geopolitical fragmentation, de-dollarisation pressure, and sanctions risk, are durable. They are not guaranteed to persist at current intensity forever. That means the floor sovereign demand builds is real, but the ceiling is never set by central banks alone.

Reading the current demand split before making a directional call

Return to where this started: a softening price against record Chinese buying. That tension is not a contradiction to be explained away. It is the market’s structure made visible.

The dual-driver framework resolves it. China’s 650,000 troy ounce August purchase anchors a structural floor that does not respond to price. The $6.5 to $7 billion GLD inflow shows the ETF layer has been active but fragile, and rate expectations, not sovereign demand, are setting the near-term direction from a price of $4,302 per ounce.

Turkey’s 8-tonne August decline stands as the counterweight to any tidy bullish narrative. The aggregate “central banks are buying” story conceals country-level reversals that arrive fast and hard.

For positioning, three forward variables carry the weight:

• Rate trajectory: the single biggest input into whether ETF flows sustain or reverse
• ETF flow direction: the marginal price-setter, and the layer most likely to move first
• China’s monthly pace: whether the PBoC holds its cadence or signals any moderation

The takeaway is a reframing. A correction in the gold price is not evidence the structural case has broken; it is far more likely a signal that ETF positioning has shifted on rate expectations, while the sovereign floor sits undisturbed beneath it. Investors who hold that architecture in mind read price moves more accurately than those watching a single headline number.

For readers wanting to place the current cycle in the broadest structural context, our deep-dive into the monetary shift driving sovereign gold demand traces how geopolitical fragmentation after 2022 accelerated the transition from a net-seller era to the sustained accumulation wave central banks are executing today.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results, and financial projections are subject to market conditions and various risk factors.