Gold’s Shanghai bid refuses COMEX’s verdict

Gold was hit on Tuesday as if one firmer US inflation print had exposed a fading bullion market. We think that is too simple. The break in COMEX was a macro liquidation driven by rate fears and a stronger dollar, not proof that underlying demand has rolled over, and China’s onshore market is not endorsing that verdict.

We are not calling a shortage. Elevated exchange stocks inside China argue against that. But the evidence points to a futures selloff that ran faster than the physical market justified. For gold readers, that distinction is the whole argument: a paper flush is not the same thing as a dead bid.

Tuesday’s dump was a Fed trade

The prevailing read is that Shanghai still shows some physical firmness, while the immediate selloff was mainly a macro, paper-market move. That view is mostly right. Where we part company is in the leap from “macro drove the drop” to “physical demand must also be softening”.

The timing is plain. GoldSquawk data put COMEX gold at $4,115.00 on Nov. 9, up $39.20 or 0.96%, before it dropped to $4,030.80 on Nov. 10, down $84.20 or 2.05%. On the same day, US CPI rose 0.4% month on month against a 0.3% consensus, while core CPI rose 0.3% against a 0.2% consensus. That landed on top of a strong US labour report from Nov. 6: employment growth was 165K against a 65K forecast, unemployment was 4.1% against 4.2% expected, and wage growth came in at 0.4% against 0.3% forecast.

That is a clean macro shock. Later, on Nov. 12, Morgan Downey’s Commodity News said hotter US CPI pushed the dollar to a 19-month high and fully priced a December Fed hike. That fits Tuesday’s move. What it does not prove is that bullion demand turned down with it.

China is still taking metal

The official sector is one reason we should resist that easy read. IMF data put China’s gold reserves at 75.44 million ounces in September 2026, with the reserve-gold series rising steadily through 2026. Asharq Bloomberg reported the PBOC added 25.8 tonnes in October, extending its buying streak to 24 straight months. That is not marginal buying.

The more important point is what happened onshore after the COMEX hit. GoldSquawk data compiled by GoldSquawk showed the Shanghai premium over London at +$8.72 an ounce by Nov. 13, against a 7-day average of +$1.74. That means the premium was roughly five times its own weekly mean, using $8.72 and $1.74. GoldSquawk data put Au(T+D) open interest at 210,300 lots, while volume reached 59,700 kg, or 1.8 times its 19-day median. Delivery was 30,796 lots. And shorts paid longs for four trading days in a row.

That last point is not trivia. The Shanghai Gold Exchange says shorts pay longs when delivery-making volume is smaller than delivery-taking volume in its deferred contracts. In plain English, more participants wanted metal than wanted to make delivery. That is not what a market in demand collapse does.

Firm does not mean tight

The strongest evidence against our view is serious, and we should take it seriously. GoldSquawk data put Shanghai Futures Exchange gold warehouse stocks at 118.4 tonnes in the week to Nov. 13, unchanged on the week and the highest since 2025-11-14. Over 52 weeks, those stocks rose by 28.0 tonnes, or 31.0%, from 90.4 tonnes to 118.4 tonnes, while COMEX gold fell 5.3% from $4,094.20 to $3,878.90 over the same span.

That is not the profile of a market starved of metal. It means supply is available inside the exchange system. It also means a positive China premium should not be mistaken for outright scarcity. It can reflect firm local absorption, delivery preference and arbitrage friction without meaning bars are impossible to source.

There is another caution. On Nov. 13, the mini Au(T+D) contract had longs paying shorts, not the other way round. So even within Shanghai’s own deferred complex, the signal was not one-way. And on Nov. 12, spot gold still fell 1.4% to $3,931.54, with Morgan Downey noting that central-bank buying provided only some floor. That cuts against any dramatic squeeze story.

What gold readers should watch now

If we are right, the lesson is not that gold ignored macro. It plainly did not. The lesson is that COMEX can overshoot when the rates trade turns violent, while the underlying bullion market stays firmer than the futures screen suggests.

That matters beyond this week. The World Gold Council said on Nov. 9 that gold had held support around $3,858 an ounce, while money-manager net longs on COMEX were around 325 to 335 tonnes and the market was at a “typical” downside extreme. If paper positioning is already near a normal washout and China’s onshore market keeps paying up for metal, the risk shifts from a smooth bear trend to a harder floor and a sharper rebound when futures sellers run out.

The tests are simple. Does the Shanghai premium stay positive against London? Do deferral fees keep favouring longs? And do warehouse stocks finally start to draw from 118.4 tonnes? If those hold while gold defends the area around $3,858, COMEX’s selloff will not stand as a verdict on demand. It will stand as a paper market that sold first and asked questions later.