Gold hit $5,589 per ounce on January 28, 2026. It was the kind of number that forces a rethink — not just of the asset, but of the assumptions underneath it. Four months later, gold trades near $4,600. That is an $989 correction, roughly 18% off the all-time high, compressed into fewer than 100 trading days. The dollar is at a three-year low. Central banks are buying at a pace that rivals the 2022 record. Real yields, which supposedly drive gold inversely, are sitting at 1.94% — elevated, not collapsing. And yet gold is down.
That paradox is not noise. It is the signal. Something structural has shifted in how gold prices. Our analysts have been tracking the breakdown for weeks, and the conclusion is uncomfortable for anyone running a simple commodity model: the correlations that governed gold for the better part of two decades no longer hold.
Gold's $989 Drop From ATH: Why the Standard Explanations Don't Hold
The textbook narrative for a gold correction goes like this: yields rise, the dollar strengthens, risk appetite returns, and gold sells off as an opportunity cost becomes too high. Run that checklist against May 2026 and it falls apart at nearly every point.
The DXY — the dollar index — sits at 98.2, a level last seen in mid-2023 when gold was trading in the $1,900s. A dollar at three-year lows should be fuel for gold, not a headwind. The inverse dollar-gold relationship has been one of the most reliable macro correlations over the past 30 years, with a coefficient routinely running between -0.6 and -0.8. That relationship is currently muted at best, inverted at worst.
Central bank demand is the second anomaly. Poland added 80 tonnes to its reserves in recent months. China has been a buyer for 15 consecutive months. The World Gold Council estimates 755 to 850 tonnes of central bank purchases for full-year 2026 — a figure that, if realized, would match or exceed the record year of 2022. These are not marginal buyers. Central banks represent price-insensitive, strategic accumulation.
North American ETF outflows offer part of the story. Retail and institutional investors in the US and Canada pulled roughly $13 billion from gold ETFs in March 2026 alone. The outflows reflect profit-taking, asset reallocation toward equities after the April recovery, and in some cases forced selling by multi-asset funds managing volatility targets.
What we are left with is a correction that is real, that has legitimate technical and flow explanations, but that is occurring against a macro backdrop — weak dollar, active central bank buying, geopolitical stress, and a Hormuz premium in oil — that should, by historical precedent, be constructive for gold. The fact that it is not tells us something has changed.
The Broken Gold-Yield Correlation: From -0.82 to +0.18 and What It Means
The gold-yield correlation is where the clearest structural break lives. For most of the post-2008 era, the relationship between 10-year TIPS real yields and gold was reliably inverse, with a correlation coefficient running near -0.82 at peak. The logic was elegant: gold is a zero-yield asset, so its opportunity cost rises when real yields rise, and falls when they decline.
The current 10-year TIPS real yield sits at approximately 1.94%. By the old model, this should be deeply bearish for gold. And yet gold was trading above $5,000 earlier this year with real yields at similar levels. The correlation has not just weakened; it has flipped. The current coefficient is running near +0.18 — effectively zero, with slight positive tilt.
Our analysts see three drivers behind this regime change.
De-dollarization as a structural bid. The dollar's share of global central bank reserves has fallen from 71% in 1999 to below 58% today. That is not a cyclical fluctuation — it is a 13-percentage-point structural shift over 25 years that is still in motion. When reserve managers diversify away from dollars, they need somewhere to put the capital. Gold is the only asset with sufficient liquidity, no counterparty risk, and no political attachment. This structural demand is not sensitive to real yields.
Geopolitical uncertainty as a persistent premium. Brent crude is trading between $108 and $116 with a visible Hormuz risk premium. When oil carries a war premium, so does gold — and that premium does not respond to yield movements.
The breakdown of consensus macro playbooks. Institutional gold positioning was heavily model-driven through 2021 and 2022 — algorithms that sold gold when real yields rose beyond certain thresholds. As those models have failed to predict gold's behavior, their weight in overall positioning has declined. What replaces them is less systematic and more discretionary, which means the yield signal carries less price impact than it did during the era of high-frequency macro momentum.
This is not a temporary noise in the data. Our analysts have run rolling 12-month correlations and the breakdown has been persistent since mid-2024. The market is repricing gold under a new framework, one where reserve diversification and geopolitical risk premium carry more weight than real yield differentials.
Silver, Crude, and J.P. Morgan's $5,000 Target: How to Position Now
Silver hit $121 per ounce at its 2026 high. It currently trades between $73 and $76, a correction that exceeds gold's on a percentage basis. The gold-to-silver ratio sits near 61:1 — historically elevated relative to the low-40s ratio during the recent gold bull run.
Silver's correction is larger than gold's for a familiar reason: silver carries both monetary and industrial characteristics. When risk appetite wanes and industrial production concerns surface — as they have with the oil price spike feeding into input cost fears — silver gets sold twice. Once as a monetary hedge and once as an industrial commodity.
On crude, the Hormuz premium is not theoretical. Brent at $108 to $116 reflects a market pricing meaningful probability of supply disruption, not a temporary spike. For commodities portfolios, elevated energy prices validate the broader geopolitical risk narrative that supports gold, and also compress discretionary consumer and corporate spending in ways that eventually feed into recessionary signals — which historically favor gold over risk assets.
J.P. Morgan's Q4 2026 target of approximately $5,000 per ounce deserves engagement. The bank's commodity team is not projecting a return to the January high; they are projecting recovery to a level that was previously seen as a ceiling. Our analysts find the framework credible for three reasons. First, the structural drivers — de-dollarization, central bank accumulation, geopolitical premium — are not reversing. Second, the ETF outflow story is a positioning cycle, not a fundamental demand collapse. Third, the $4,600 level is holding as technical support through multiple tests, which suggests real demand at these prices.
Gold miners offer leveraged exposure at a discount. Many producers trade at historically compressed multiples relative to the gold price, partly because the rally from $2,000 to $5,589 happened faster than earnings revisions could track.
What our analysts are not doing is treating this correction as the end of the commodity cycle. The correlations have broken. The old playbook is less useful than it was. The structural buyers — central banks, Asian physical markets, geopolitical risk allocators — are still accumulating. And J.P. Morgan's $5,000 target, far from being ambitious, may turn out to be conservative if the Hormuz premium persists and the dollar continues its structural decline.
The gold-yield correlation that defined a generation of macro investing is broken. The question is not whether to own gold. The question is which framework you are using to understand it — because the old one no longer works.