By Gold did not fall on Friday because inflation is beaten. It fell because the man who decides the price of money said the opposite. Spot gold closed 28 August 2026 at roughly $4,456 an ounce, down 3.2% on the day, after Federal Reserve Chairman Kevin Warsh used his first Jackson Hole keynote to tell the market that underlying inflation has not improved and that the Fed may still have work to do. That is the second time in seven months that a single Warsh event has repriced gold. The first was 30 January 2026, the day President Trump nominated him: gold fell 10.3% and silver 28.5% in one session, and the January high of about $5,405 has not been revisited since. Any honest gold price prediction has to start by admitting that this market’s two worst days in a year were both about one person.
Now the part that complicates the bearish reading. While gold fell through the second quarter, central banks bought 289 tonnes of it, a 62% jump on the same quarter a year earlier and the strongest second quarter in the World Gold Council’s data series. Over the same three months, gold ETFs shed 45 tonnes as Western investors marked up their rate expectations. Two groups looked at an identical falling price and did opposite things, and the official sector bought roughly 6.4 times what the funds sold. That is the single most important fact in gold right now, because it means the marginal buyer of gold is not rate-sensitive. Warsh sets gold’s price week to week. He does not set the bid underneath it. The distance between those two statements is the entire bull-bear spread.
Key facts: gold in seven numbers
Spot gold $4,456/oz at the 28 August 2026 close, down 3.2% on the day and 17.5% below the January peak of roughly $5,405, but still up 29.8% over twelve months. Sources: Trading Economics, gold-api, SPDR Gold Shares closes.
Central banks bought 289 tonnes in Q2 2026, up 62% year on year and the strongest Q2 on record, led by the National Bank of Poland at 51 tonnes. Source: World Gold Council, Gold Demand Trends Q2 2026.
Gold ETFs sold 45 tonnes in Q2 2026, with North America accounting for 61 tonnes of outflow across the first half while Asia added 70 tonnes. Source: World Gold Council.
PCE inflation is 3.7% over twelve months and 4.1% over six, meaning the Fed’s preferred gauge is accelerating, not decaying. Source: Chairman Kevin Warsh, Jackson Hole, 28 August 2026.
Gold fell 10.3% and silver 28.5% on 30 January 2026, the day Warsh was nominated as Fed Chair. That day marked the top. Sources: SPDR Gold Shares and iShares Silver Trust closes; CNN.
The gold-silver ratio is 67:1, against 46:1 at the January peaks and 88:1 twelve months ago. Silver is 43.2% below its January high; gold is 17.5% below its own. Source: FinanceFeeds calculation from ETF closes.
Every major bank 2027 target sits above spot, from UBS at $5,200 by June 2027 to J.P. Morgan at $6,300 by end-2027. There is no published bear case. Sources: bank research summaries, August 2026.
What Warsh actually said, and why it moved gold
It helps to read the speech rather than the headline. Warsh titled his remarks “In Our Time” and delivered them on his 100th day as Chairman. The passage that moved the market was not a rate signal at all; it was a standard. “Here is my standard,” he said. “We must be confident that underlying inflation is moving to our objective, clearly and at sufficient speed. Otherwise, we have work to do.”
The arithmetic behind that sentence is what should worry a gold bear and a gold bull simultaneously. Warsh noted that the twelve-month change in the PCE price index stands at 3.7%, “while the six-month change is 4.1 percent.” A six-month rate running above the twelve-month rate means inflation is accelerating. He was explicit that the summer’s better-than-expected prints “do not tell me that underlying trends have meaningfully improved,” and blunter still on accountability: “The responsibility for 65 months of sustained, elevated inflation sits squarely with the central bank. And that is where it belongs.”
Markets read that as a September hike becoming live, the dollar firmed, yields rose, and gold, which pays no coupon, did what a zero-yield asset does when the discount rate jumps. That is the textbook mechanism and it worked exactly as advertised for one session.
The textbook stops working when you zoom out. Over the twelve months in which the Fed went from cutting to openly discussing hikes, and in which a self-described inflation hawk took the chair, gold rose 29.8%. If rising real rates reliably sank gold, that number would be negative. It is not. Something is bidding for gold that does not read the dot plot, and the World Gold Council data says plainly what it is.
Who was buying while the price was falling
The second quarter is the cleanest natural experiment gold has offered in years, because the price fell and the two big buyer groups moved in opposite directions.
Central banks bought 289 tonnes, a 62% increase on Q2 2025 and the strongest second quarter in the World Gold Council’s series. The National Bank of Poland was the largest single buyer at 51 tonnes, taking its reserves to 632 tonnes by the end of June. The People’s Bank of China added 33 tonnes, its biggest quarterly addition since the fourth quarter of 2023. None of these institutions bought because they expected a rate cut. They bought because they are diversifying reserves away from assets that another government can freeze, and a falling price makes that cheaper.
Exchange-traded funds did the opposite, shedding 45 tonnes over the quarter in response to weaker prices and upward revisions to both inflation and interest rate expectations. The regional split is the tell: across the first half, North American funds lost 61 tonnes while Asian funds added 70 tonnes and European funds added 8 tonnes. The marginal seller of gold is Western and rate-driven. The marginal buyer is Eastern, official, and policy-driven.
This is a pattern anyone who has covered commodity markets will recognise from oil’s strategic petroleum reserve dynamics, where a price-insensitive sovereign buyer sets a soft floor that private speculators repeatedly underestimate. It is also why our recent analysis of central bank gold demand concluded that forecasters cannot agree on where gold ends up: they are modelling a private-sector asset that has acquired a public-sector buyer of last resort.
The silver test, and what it proves about gold
If you want to know whether gold’s decline is a monetary story or a positioning story, compare it with silver. The two metals peaked within a day of each other in late January. Since then silver has fallen 43.2% while gold has fallen 17.5%. The gold-silver ratio, which compressed to about 46:1 at the twin peaks, now sits at 67:1, having been 88:1 a year ago.
That divergence is diagnostic. Silver’s January move was a speculative and industrial mania, and manias unwind violently: silver lost 28.5% in the single session of 30 January and another 15.8% a week later. Gold’s worst day in the same episode was 10.3%, painful but a third of silver’s. Both metals lost their speculative bid at the same moment. Only gold retained a second, separate bid underneath it, and that second bid is the official sector. Readers weighing the two should also see our silver price prediction, which treats the industrial demand side that gold simply does not have.
The practical conclusion for a long-term holder is that gold and silver are not the same trade and should not carry the same position sizing. Silver offers more upside per unit of the same macro thesis and considerably more drawdown. At 67:1 the ratio sits close to the middle of its twelve-month range, which means neither metal currently screens as obviously cheap against the other. That is an unglamorous finding, and it is more useful than a target.
Gold spot to 28 August 2026 against the $6,200 bull and $3,800 bear cases, with the two Warsh events marked, and Q2 2026 central bank buying against gold ETF flows. Sources: World Gold Council Gold Demand Trends Q2 2026; prices derived from SPDR Gold Shares closes calibrated to spot.
The long-term case: what has to be true by 2028
The bull case, and how we get to $6,200. It requires the official bid to persist and the Western bid to return. The first is already happening and the World Gold Council expects another strong year, if below 2025. The second is a rates question: North American ETFs shed 61 tonnes in the first half precisely because rate expectations rose, which means that flow reverses mechanically once the market stops pricing hikes. Add a returning Western bid on top of an official bid that never left, and gold clears the January high of $5,405. Extending roughly 15% beyond that prior peak, which is a modest overshoot by the standard of previous gold cycles, gives approximately $6,200. That sits above UBS’s $5,200 June 2027 path and just below J.P. Morgan’s $6,300 end-2027 figure, so it is a bull case, not a fantasy.
The bear case, and how we get to $3,800. Warsh hikes in September and again before year-end, real yields rise by something like a percentage point, and gold’s historical sensitivity to real rates of roughly 8% to 10% per 100 basis points drags it toward $4,010. Then add a genuine liquidation quarter: Q2’s 45 tonnes of ETF outflow was orderly, and a disorderly one in a hiking scare could be three times that. The overshoot takes gold to about $3,800. Note what this bear case still concedes. Even at $3,800 gold sits 11% above its twelve-month low of roughly $3,433, because the official bid raises the floor even when it cannot hold the ceiling.
The structural change nobody is pricing. Buried in the same speech, Warsh announced that he intends to dismantle forward guidance, saying the practice “has overstayed its welcome” and that he prefers his commitment to be “to a discipline, not to a decision.” For equities that is a communication story. For gold it is a valuation input. Forward guidance existed to suppress volatility in the expected path of rates; removing it raises the variance of that path even if the average level is higher. Gold is, among other things, a long-dated option on monetary disorder, and options are worth more when variance rises. A hawkish chair who also removes the market’s forward visibility is not straightforwardly bearish for gold, and this is the piece of Friday’s speech that the one-day 3.2% selloff did not price at all.
The political risk that cuts both ways
There is a second-order risk here that is unusual to see priced explicitly. The Wall Street Journal has reported that President Trump has called Warsh repeatedly since he became Fed Chair, and Warsh’s own framing at Jackson Hole, placing responsibility for 65 months of inflation squarely on the central bank, reads as a man defending institutional prerogative rather than accommodating the White House. Prediction markets currently put roughly a 10% probability on Warsh being out as Fed Chair.
That 10% is a genuinely two-sided number for gold, and it is worth being precise about why. If Warsh stays and hikes, the near-term path is the bear case. If Warsh is removed for refusing to cut, the message to every reserve manager on earth is that US monetary policy is subject to political override, which is the single most powerful argument for holding a reserve asset that no government issues. In that scenario the official-sector buying that already runs at 289 tonnes a quarter accelerates rather than fades. Gold is therefore short the Fed’s hawkishness and long the Fed’s fragility at the same time, which is precisely why it can fall 3.2% on a hawkish speech and still be up 29.8% on the year. The same tension is visible in rate positioning, as our coverage of traders buying both a cut and a hike into September showed.
What happens next: three predictions
One: the September meeting matters less than the November PCE prints. Warsh committed to “a discipline, not to a decision,” and with the six-month PCE rate at 4.1% against 3.7% over twelve months, the data decides. We expect no hike in September and a live one by year-end if the six-month rate stays above the twelve-month rate. Gold’s floor holds through the meeting itself; the risk is in the data that follows.
Two: central bank buying prints above 250 tonnes again in Q3. The Q2 figure of 289 tonnes came with the price falling, Poland adding 51 tonnes and China adding 33. Nothing in the third quarter changed the reserve-diversification logic, and a lower price improves the entry. If Q3 comes in below 200 tonnes, the bull case weakens materially and the $3,800 bear case becomes the working assumption rather than the tail.
Three: the ETF flow, not the price, is the signal to watch. North American funds have been the marginal seller all year. The first quarter in which North American gold ETF flows turn positive while central bank buying holds is the moment both bids are pulling in the same direction, and that is historically when gold makes new highs rather than merely defending old ones.
The honest summary is that gold got cheaper on Friday without getting weaker. The price fell because one man raised the discount rate applied to a non-yielding asset. The demand that has actually driven this bull market, 289 tonnes a quarter from institutions that do not trade on Fed speeches, was unaffected by anything said in Wyoming. For a long-term holder, the useful question is not where gold trades in September. It is whether the official sector is still buying in 2028, and on the evidence of the last four years, the burden of proof sits with whoever argues that it stops. For the commodity complex more broadly, see our oil price prediction and our work on strategic materials.
Frequently asked questions
Why did gold fall on 28 August 2026?
Fed Chairman Kevin Warsh used his first Jackson Hole keynote to say that underlying inflation had not meaningfully improved and that the Fed may have “work to do,” which markets read as making a September rate hike live. The dollar firmed and Treasury yields rose, and gold, which pays no yield, fell about 3.2% to roughly $4,456 an ounce.
What is a realistic gold price prediction for 2027?
The published bank range runs from UBS at $5,200 by June 2027 to J.P. Morgan at $6,300 by end-2027, with Goldman Sachs at $5,400 to $5,600 and Bank of America at $6,000 on a twelve-month view. Every one of those sits above the current $4,456 spot price. Our own bull case is $6,200 and our bear case is $3,800, and the bear case exists because no major bank currently publishes one.
Are central banks still buying gold in 2026?
Yes, and at an accelerating rate. The World Gold Council recorded 289 tonnes of central bank purchases in Q2 2026, up 62% year on year and the strongest second quarter in its data series, with Poland adding 51 tonnes and China 33 tonnes. Crucially, that buying occurred while the gold price was falling, which is what distinguishes official demand from investor demand.
Why is gold down 17.5% but silver down 43.2%?
Both metals peaked in late January 2026 and lost their speculative bid at the same time. Silver’s demand is roughly half industrial and its January move was a mania, so it unwound violently, losing 28.5% in a single session on 30 January. Gold retained a second, non-speculative bid from central banks, which silver does not have. The gold-silver ratio has widened from 46:1 to 67:1 as a result.
Does a hawkish Fed automatically mean lower gold?
Not reliably. Over the past twelve months the Fed moved from cutting to openly discussing hikes and a noted inflation hawk took the chair, yet gold rose 29.8%. Rate expectations dominate day-to-day moves, but reserve diversification dominates the multi-year trend. Warsh’s plan to retire forward guidance also raises the variance of the expected rate path, which increases rather than decreases the hedging value of holding gold.
What would invalidate the bullish case for gold?
A quarter in which central bank purchases fall below roughly 200 tonnes while ETF outflows continue would remove both legs of support and make the $3,800 bear case the base case. Sustained six-month PCE readings below the twelve-month rate would do similar damage, because that would signal the Fed is winning and the monetary hedge is less necessary.
This article is for information only and is not investment advice. The $6,200 bull case and $3,800 bear case are FinanceFeeds estimates derived from published demand data and stated central bank policy, not price targets. Prices are as of the 28 August 2026 close. Primary sources: “In Our Time”, keynote remarks by Chairman Kevin Warsh at the 2026 Jackson Hole Economic Policy Symposium; World Gold Council, Gold Demand Trends Q2 2026; CNN on the 30 January 2026 nomination; Trading Economics gold price data.