Gold traded near $4,700 in mid-May 2026 even after the inflation shock of 2022 faded and real yields stayed positive. The clean explanation is not a resurrected CPI trade. It is a reserve-allocation trade: central banks bought more than 1,000 tonnes a year from 2022 through 2024 and another 863 tonnes in 2025, changing who sets the marginal price.
Gold near $4,700 an ounce in mid-May 2026 is not supposed to sit comfortably inside the old macro model. The October 2022 low was near $1,656. From there, bullion roughly tripled while the inflation panic cooled, the Fed completed one of the sharpest hiking cycles in modern history, and real yields stayed positive.
The old framework says that should not happen. Gold pays no coupon. Higher real yields raise the opportunity cost of holding it. Disinflation should reduce the urgency of owning it as a monetary hedge. A strong dollar should tighten the screw further. Yet the price kept climbing through conditions that would usually cap the move or reverse it.
The first correction is factual. The relevant mid-May price was not a neat $4,613 Monday close. Public daily XAU and gold-price series put gold closer to the high-$4,600s to low-$4,700s around May 11-13, with $4,613 appearing in earlier May trading. The exact print is less important than the regime: gold was still trading around $4,700 after a move from the 2022 low that no simple real-yield regression can explain.
The second correction is analytical. This is not proof that real yields no longer matter. They still matter at the margin. It is evidence that the marginal buyer changed. The Western ETF-and-real-yield model is describing one part of the market. Since 2022, another part has been large enough to set the price.
Get analysis like this delivered daily. No account required.
This analysis was produced by the Convex Research Desk from live economic data and is for informational purposes only. It does not constitute financial, investment, or legal advice. See our editorial standards and terms of service.
Convex combines institutional macro research with AI-powered signal generation. Access live trading signals, portfolio analytics, and strategy backtests.
The real-yield model worked best when the price-setting flow was Western investment demand: futures, ETFs, macro funds, and dollar-sensitive investors using gold as a hedge against inflation or monetary disorder. In that market structure, a higher TIPS yield can pull gold lower because the same investor choosing between bullion and real rates controls the marginal dollar.
That has not been the dominant structure since 2022. The more important flow has been official-sector demand. World Gold Council data show central banks bought 1,082 tonnes in 2022, 1,037 tonnes in 2023, 1,092 tonnes in 2024, and 863 tonnes in 2025. The 2025 number cooled from the prior three years, but it was still far above pre-2022 norms.
Those flows are different from ETF flows. A reserve manager does not trade gold like a macro hedge fund. A central bank adding bullion to diversify reserves is solving forliquidity, sanctions resilience, domestic custody, and long-run balance-sheet composition. It is not trying to beat next week's payrolls print.
That distinction explains why the tape looked wrong to investors using the old model. Western ETF holdings could soften, real yields could remain elevated, and gold could still rally if official buyers were absorbing metal for reasons that had little to do with the next CPI release.
The Sanctions Shock
The catalyst was not mysterious. After Russia's February 2022 invasion of Ukraine, the United States, European Union, United Kingdom, Canada, Japan, and allies immobilized roughly $300 billion of Russian central-bank reserves. The legal mechanics were complicated. The message to other reserve managers was simple: reserves held inside another sovereign's financial architecture can become political instruments.
That did not make dollar reserves useless. It did change the risk calculation. Treasuries are liquid, deep, and income-producing, but they sit inside a system controlled by the issuer and its allies. Euro assets carry the same problem. Yuan assets are not deep or open enough to replace the dollar at scale. Bitcoin is too volatile and too small for reserve management in size.
Gold is different. It is no one's liability. It can be stored domestically. It can be moved or pledged outside the correspondent-banking system. It does not solve every reserve problem, but it solves one problem that became more important after 2022: it reduces dependence on a foreign custodian's political permission.
That is the variable the old spreadsheet does not have. The model can handle inflation, real yields, and the dollar. It struggles with sanctions risk on reserve assets because that risk was treated as close to zero for most large sovereigns before 2022. Once it moved above zero, the demand curve shifted.
Why CPI Did Not Kill the Bid
The inflation data complicate the story but do not overturn it. Headline CPI fell from the 9.1% peak in June 2022 to 2.4% in February 2026, then reaccelerated to 3.3% in March and 3.8% in April. Core CPI was 2.8% year over year in April. That is not a clean disinflationary glide path anymore, but it is also not the 2022 panic.
If gold were only a CPI hedge, the move would be hard to defend. It rose while inflation fell, and it remained bid when real yields were positive. But if gold is also a reserve-diversification asset, the CPI path becomes secondary. The key question is not whether inflation is 2.4%, 3.3%, or 3.8% in a given month. The key question is whether central banks still want more bullion on their balance sheets.
The answer, through 2025, was yes. The pace slowed, but it did not normalize. Buying above 800 tonnes in 2025 is still an unusually large official-sector bid. That keeps pressure under the market even when Western investors would normally fade the move.
April CPI still matters. At 3.8% headline and 2.8% core, it makes it harder for the Fed to validate an aggressive easing path. Higher real yields can produce tactical gold drawdowns, especially if ETF holders sell into a stronger dollar. But a tactical drawdown is not the same thing as a regime break. To break the post-2022 gold regime, the market needs evidence that official-sector demand is no longer absorbing Western liquidation.
What Would Actually Falsify the Thesis
The clean test is not one CPI print. It is the interaction of price, ETF flows, and central-bank reserve data.
If GLD and other Western gold ETFs see redemptions while spot gold holds near the highs, the reserve-bid thesis strengthens. It means the selling is being absorbed somewhere else. If Shanghai premiums widen during U.S. real-yield selloffs, that also supports the thesis because it points to Asian physical demand rather than Western futures positioning. If Poland, India, China, Turkey, or other repeat official buyers keep adding at elevated prices, the mandate-driven explanation remains intact.
The bearish version is just as clear. If real yields push materially higher, the dollar firms, ETFs liquidate, and official-sector buying slows at the same time, gold can finally reconnect with the old model. A reserve buyer can be price-insensitive for a while. It is not price-insensitive forever. Turkey has sold gold during domestic funding stress before. Other reserve managers can pause when the local political cost of buying expensive metal rises.
That is why the 2025 slowdown matters. Central banks still bought a large amount, but the decline from more than 1,000 tonnes to 863 tonnes says price and liquidity conditions are not irrelevant. The reserve bid is powerful. It is not infinite.
The Market Is Not One Buyer
The practical mistake is treating gold demand as one homogeneous pool. It is not. Jewellery buyers are price-sensitive. ETF buyers are performance-sensitive. Futures traders are rate-sensitive. Central banks are balance-sheet-sensitive. Each group can dominate at different points in the cycle, and the headline price is the clearing level between all of them.
That is why a high gold price can coexist with weaker jewellery demand, mixed ETF flows, and continued official buying. Those are not contradictions. They are separate demand curves. A household in India may delay purchases when the local price spikes. A reserve manager may still add because the allocation target is multi-year and the sanction-risk argument has not changed. A macro fund may short gold against real yields for a month. A central bank can absorb that flow without caring about the fund's stop-loss.
Once the market is read that way, $4,700 gold becomes less mystical. It is not saying inflation is about to return to 9%. It is saying that the non-yield motive for holding gold has become large enough to compete with the yield motive for selling it.
The Right Way to Read $4,700 Gold
The useful conclusion is not that gold is cheap or expensive at $4,700. It is that the reason for the price has changed.
Before 2022, the cleanest shorthand was real yields. When real yields rose, gold usually struggled. When real yields fell, gold usually caught a bid. That relationship is still part of the tape, but it is no longer enough. A model that ignores official-sector buying will keep treating the rally as impossible, then be forced to explain why the impossible keeps happening.
Gold's mid-May 2026 level is therefore best understood as a market-clearing price between two forces. Western macro investors see positive real yields and ask why a zero-coupon metal should be this expensive. Reserve managers see sanctions precedent, fiscal stress, geopolitical fragmentation, and dollar concentration risk, then keep allocating to an asset that is not anyone else's liability.
That is the anomaly. Not that gold rose during disinflation. That can happen in a crisis. The anomaly is that it kept rising after the initial crisis logic faded, because the buyer base had changed in a way the old model was not built to capture.
The next signal is not a single CPI release. It is whether elevated prices finally slow central-bank buying enough for real yields to regain control of the tape. Until that happens, gold around $4,700 is not a mistake in the data. It is the market telling you which buyer matters most.