The slogan is easy. Asia's gold producers have lost faith in the dollar, so they are hoarding the metal, and the price must rise. The tonnes are harder, and they are the only part an investor can use. A few governments are changing the rules at the mine gate. Those rules are real. They are not, by themselves, a shortage. The bid that has carried this bull market is still the official buyer. The force that knocked the price off its highs is still the interest rate. Mix those three facts into one cheer and you will buy the wrong week.
That is the whole idea. A producer that keeps gold at home is a political fact. It becomes a market fact only when the ounces that leave the world market are large next to the ounces central banks and Chinese buyers are already taking. On the numbers in front of us in the first week of October 2026, the politics are broader than the shortage.
Nothing in this piece is a recommendation to buy or sell gold, a gold fund, or a mining stock. Forecasts cited below are other people's forecasts. They are not promises.
What the producers actually did
Start with the small country, because the headlines do. Laos mined roughly 12 tonnes of gold in 2025. The World Gold Council and Metals Focus put that output sixth in Asia. The country talks about resources of 500 to 1,000 tonnes. Treat that range as a local estimate, not as a reserve a mine builder could take to a bank. Until recently most of the gold left as ore, through official channels and unofficial ones. In 2024 the government set up the Lao Bullion Bank. The aim is to refine gold at home, raise gold's share of the nation's foreign-exchange reserves, and give citizens a place to store savings that is not a suitcase.
The prime minister, Sonexay Siphandone, has called gold development a priority for the economic foundation. A guest from the Japan Bullion Market Association, at the launch, called the speed astonishing. Speed is not scale. Twelve tonnes is a meaningful mine for Laos. It is a rounding error in a world that mines on the order of three and a half thousand tonnes a year. If every ounce Laos produces stayed home forever, the global balance would barely feel it. The story is the direction of policy, not the weight of the bar.
Indonesia is large enough to matter more, and the policy is more precise than the slogan. The country is often described as the world's tenth-largest producer, at more than 100 tonnes a year. It is also a place where local investment demand has been running ahead of what miners want to leave in the country. Jakarta's answer was not a ban. It was a tax. The rule that took effect for this cycle, under a 2025 finance regulation, charges more when the reference price sits above $3,200 an ounce. In the first half of October 2026 that reference price was still far above the line, near $4,336 an ounce. The duty then depends on the product. Doré, the impure bar from a mine, is taxed at 15 percent. Non-doré granules are taxed at 12.5 percent. Cast bars and minted bars are taxed at 10 percent. The trade ministry also cut the export benchmark price for October 1 to 14, by about 3.5 percent from the prior two weeks. The rates did not change. The base of the tax did.
Read that again before you call it hoarding. Gold can still leave Indonesia. It costs more to leave, and it costs most when it leaves in the least processed form. A 15 percent duty on doré, at a gold price a little above $4,100, is on the order of $600 an ounce. That is a real wedge. Smelters and local buyers are supposed to win the wedge. Foreign refiners are supposed to lose it. Whether miners actually ship less will show up in export data, not in the decree. A tax is a price. A price can be paid.
China is the producer that already lived this way. It mines a little over 380 tonnes a year, about a tenth of world output, and it is also the world's great importer. As a rule, taking gold out of China is restricted. That line, given to the Nikkei by market analyst Jeff Toshima, is the old policy, not a new one. Beijing does not need a 2026 tax to keep its own mine output at home. It has been doing a version of that for years. The new information is not that Chinese gold stays in China. The new information is how much gold China is pulling in from everywhere else.
The number that breaks the slogan
If producers were simply sitting on their metal, China's import line would be quiet. It is the opposite. Heraeus, writing at the end of September, said China imported 142 tonnes in August and 1,141 tonnes in the first eight months of 2026. That was 72 percent more than the 663 tonnes imported in the same months of 2025, and already more than the 940 tonnes imported in all of 2025. If the pace held, the year would land near 1,700 tonnes, the high for this decade. Heraeus was careful about the label. These are non-monetary imports. They are separate from the People's Bank of China, which added about 20 tonnes to official reserves in August.
Put Laos next to that. Twelve tonnes of annual mine output against 1,141 tonnes of Chinese imports in eight months. The country that makes the speeches about a new refinery is not the country that is clearing the market. The country that clears the market is the one that mines a lot and still buys far more. China is not hoarding instead of importing. China is keeping its own gold and importing other people's gold on top. That is a tighter story, and a bigger one.
Indonesia's tax points the same way, if you listen to the reason Jakarta gave. Domestic supply could not keep up with local investment demand. The tax is there to bias metal toward home buyers, not to bury it. Some of those home buyers are investors. Some may be the state. The export form that gets the highest tax is the least refined. That is industrial policy as much as monetary policy. A Western refiner who used to receive Indonesian doré has a problem. The London price does not move one-for-one with that problem. The gold can be refined somewhere else. The tax can be absorbed. The metal can leak. The honest claim is narrower. The cheap, quiet pipeline from tropical mines to Swiss and British refiners is being fenced. Metals Focus and the traders quoted in the Nikkei are right that the fence will touch the refining trade. They have not shown a hole in world supply the size of the price drop investors just lived through.
Africa is in the same conversation, at a smaller scale. Madagascar's central bank has been buying domestically produced gold since the early 2020s under a purchase program its own staff have called a cornerstone of reserve diversification. Ghana has worked with the World Gold Council on keeping more of the benefit of its gold at home and on the illegal flows that skip the state. These are real programs. They are also easy to over-read. A purchase program is a buyer of last resort inside one country. It is not a vacuum under the world price.
The buyer that does move the price
Central banks are the scale that matches the bull market. Goldman Sachs, in the work Nikkei and others were citing this week, said the official sector was buying on a three-month seasonally adjusted pace of about 91 tonnes a month. That is more than five times the pre-2022 average of about 17 tonnes a month. China, in Goldman's July nowcast, accounted for about 35 tonnes, roughly double what Beijing reported. The bank's longer forecast is a bit cooler than the nowcast, and the difference matters. Goldman raised its working assumption for central-bank demand to 60 tonnes a month on average through 2026 and 2027. It had been using 50 tonnes for 2026 and 40 tonnes for 2027. Keep the two figures apart. Ninety-one is a recent pace. Sixty is the pace they are willing to underwrite in a model that still ends at $5,400 an ounce in December 2027.
Goldman was explicit about the split inside that forecast. Continued central-bank diversification, in their note, does nearly all of the expected gain of about 23 percent through the end of 2027. Exchange-traded funds and speculative futures are a small slice of the same chart. The bank cut its year-end 2026 fair value to $4,650 from $4,900, and kept the 2027 number. The path got slower. The destination, in their book, did not. A slower path is the sentence investors skip. It is the sentence that explains a market that can fall 12 percent from a late-summer high and still sit inside a bullish model.
Société Générale's cross-asset work, also circulating this week, put the dollar's share of official foreign-exchange reserves near 57 percent in 2025. That is down more than five points since 2022. In the 2026 survey they cited, 62 percent of reserve managers expected that share to keep falling moderately over five years. And 84 percent expected gold to be a larger share of reserves. The same shop's chart of China shows official gold holdings up about 20 percent since 2022. They are more than double the 2015 level. China's Treasury holdings are down about 41 percent since 2020. The precise tonne count moves with each monthly update. The direction has not. Gold up, Treasuries down, for years, at the one central bank with both a huge mine industry and a huge import book.
The private Chinese bid may be hiding more official buying inside it. Adam Gillard, a Goldman commodities voice, noted that non-monetary imports were running very hot. He put them on the order of a thousand tonnes from January through July, plus another 142 tonnes in August. A lot of the strength was in flows into Beijing and Guangdong. Those channels have been associated with official buying before. Maybe some of the "non-monetary" metal is monetary. Maybe it is households and funds. An investor does not need to solve that mystery to use it. Either way, the metal is not sitting in a Western vault waiting for a rate cut. It has a bid in China that does not turn off when New York has a bad Monday. Between March and July, on the flow comparison making the rounds, Chinese imports offset a drop in Indian imports. They also offset a swing to selling in gold funds outside China. One buyer's surge covered other buyers' retreat. That is demand concentration. Concentration is a gift on the way up and a risk on the way down. If that buyer pauses, the chart will notice before the essayists do.
The list of central banks adding gold has also stopped looking like a Western club. The recent top five, in the Société Générale table cited this week, ran through Poland, China, Kazakhstan, the Czech Republic, and Chile. Poland is not a revolt against NATO. It is a European central bank that wants more gold. The politics are not one story. The behavior is. After Russian reserves were frozen in 2022, a dollar asset looked safe only if your relationship with Washington was safe. Gold does not clear through that relationship. An ANZ analyst, quoted by Nikkei, put it in the polite form. Gold is gaining weight as an asset insulated from any one country's political and fiscal choices. That lesson is three years old. Producer-country rules are the new chapter. They are not a replacement chapter.
Why the price still fell
Structural does not mean imminent. Gold set a record above $5,500 an ounce in January, on the London spot measure used in this week's reporting. It was near $4,700 late in the summer, and about $4,110 on September 28. That is a drop of about 12 percent from the late-summer peak, and a much larger drop from January. By October 6, spot was back in the neighborhood of $4,170. Still a high price against any year before this one. Still a bull market that has already hurt anyone who bought the headline high.
The cause of the setback is not a secret, and it is not Laos. The Federal Reserve raised rates in September for the first time in more than three years. The market has been bracing for at least one more move before the year is out. Nikkei's line was plain. Downward pressure is likely to persist until the ultimate level of the policy rate is clear. Goldman's desk, watching a 3 percent drop on September 28 as length came off in Shanghai, said the same thing in trader language. Front-end real rates were back near two-year highs. When cash pays a large positive real return, the cost of holding a bar that pays nothing is no longer theoretical. China's physical buying, in that same note, was support on a selloff. It was not enough to sustain a rally. A later commodities-desk line was just as split. Rates are holding it back. The long term is still constructive. Short-term conviction on the next move was described as very low. A materials specialist, James McGeoch, came home from the road with the shortest version. Gold is the interesting asymmetry. A floor near $4,000. Pick a ceiling.
A floor is a hope until it is tested. Several desks have parked that hope near $4,000. The price has not lived there this month. It has lived a hundred and some dollars above it, after visiting $4,110. Support is a zone you only believe after it holds, not a number a strategist liked on a Thursday. Bank of America's technicians, back in July, warned that the top of this year could leave the second half vulnerable. They talked about a death cross, crowded positions, and the ghosts of 1980 and 2011, and they put $3,315 in play if 2026 turned out to be a major top. Jefferies' mining team argued in August that gold had recoupled with real rates. Both warnings deserve a hearing, because the price did fall. Neither warning included a world in which central banks buy on the order of 60 to 90 tonnes a month and large producers fence their exports. The old tops did not look like this. That does not mean a new top is impossible. It means a model copied from 2011 is missing the buyer who showed up after 2022.
There is a third piece, and it can cut both ways. Call-option interest on the big gold fund, GLD, has been running about three times a normal level, on Goldman's count. Some of that is a macro-policy hedge. People worried about deficits in rich countries buy upside so they do not have to hold the bars. If those calls stay bid while central banks keep buying, the dealers who sold the calls may have to buy gold as the price rises. That loop can push the price through a forecast. It can also unwind. A hedge book is not a central bank. It leaves when the fear leaves, or when the loss hurts. Goldman showed a path where a very hawkish Fed and a full unwind of those hedges sit well below the base case. The upside case sits well above it. The spread between those paths is the risk. It is not a target.
What to watch, and what to ignore
An investor who wants one dashboard, not a worldview, can keep three lines.
First, the official tonne. Each month the People's Bank reports a change. Each quarter the World Gold Council and the banks update the wider official bid. If the reported Chinese streak breaks, and if the estimated 60-to-90-tonne pace rolls over, the structural story is being revised in public. Until then, the story is intact and the week can still be ugly. Those are allowed to be true together.
Second, the Chinese import tonne, separate from the official tonne. Heraeus's 1,141 tonnes through August is the number that tells you whether the private and semi-official bid is still clearing metal. A holiday week, including the early-October break in China, will scramble a few days of flow. It will not scramble an eight-month total. If imports fade for a quarter, the offset that covered weak India and weak Western funds will have to be replaced by someone else. There is no law that says the someone else appears.
Third, shipped ounces, not announced taxes. Indonesia's schedule is public. The test is whether doré exports fall, whether local refining rises, and whether the 15 percent rate gets paid or avoided. Laos's test is whether the bullion bank actually refines local gold and whether the central bank's share of reserves in gold rises by a number you can read. A plaque on a building is not a reserve.
Ignore, for the price, any sentence that treats 12 tonnes in Laos as a shock to a 3,500-tonne industry. Ignore, also, any sentence that treats the dollar's reserve share as a cliff. A slide from roughly 70 percent toward the high 50s over a generation is a grind. Grinds move gold over years. They do not pay the margin clerk this afternoon. And ignore the empire metaphor. It is tempting, and it is useless. London built a refining trade on other people's ore. Some of those countries now want the refinery and the reserve. That is industrial policy plus a lesson from 2022. You do not need the British Empire to count a 15 percent duty or a 20-tonne monthly print from Beijing.
What this means for miners, and what it does not
A tax on doré is aimed at the least processed product. Integrated producers who already pour a bar, inside or outside Indonesia, are not the target in the same way a small mine shipping concentrate is. A Chinese export restriction does not stop a Canadian or Australian major from selling into China. It may be the reason China is such a large importer. The producer-country trend, if it lasts, is a mild support for the gold price and a possible headache for a specific mine in a specific country. It is not a blanket upgrade for every gold equity.
The equity still lives on the gap between the gold price and the cost of an ounce, and on whether the company produces the ounces it promised. September showed the gap shrinking. Gold futures fell 6.4 percent on the month in one widely cited ranking, and the large gold miners fell about twice that. A structural bid from central banks did not stop that month. A new tax in Jakarta will not stop the next one if real rates jump again. Miners are a claim on the gold price with a multiplier, plus a claim on rocks, governments, and fuel. The multiplier works down. Investors who wanted the reserve story and bought the miner got a different instrument.
There is a narrower mining point that is worth keeping. If more producer states copy Indonesia, the ore will be pushed toward domestic plants. Mines that depend on a foreign refiner, and that sit in a country writing a new export schedule, carry a policy risk the headline gold price does not capture. Mines in places that already refine at home, or that sell bars rather than doré, carry less of that particular risk. Read the fiscal terms. Do not read them as a view on the dollar.
The trade people are actually describing
The long version of this story, the one Nikkei closed on, is that producer-country behavior could become another prop under the price over the rest of the decade. That can be true without being tradable on a Thursday. Combine a central-bank pace of tens of tonnes a month with a slow fencing of mine supply and you have a reason the $4,000 area is talked about as a floor. You do not have a reason it cannot break, and you do not have a reason the next hundred dollars are up. Goldman's own book says the near-term path is slower, the 2026 fair value was cut, and the desk's short-term conviction is low. Anyone selling certainty against that book is selling something the bank did not write.
The hard trade is the one that needs several things at once. The dollar's share of reserves would have to climb again. China would have to stop importing. Indonesia would have to drop the tax. Laos would have to go back to shipping ore. And real rates would have to stay high. That combination is possible. It is not the combination the last three years have been building. The easy mistake is the mirror image. It is buying every dip because a small producer opened a refinery, and calling the dip a gift from history. History is not a bid. The bid is a tonne, a rate, and a flow. This week the rate is still in charge. Over a few years the tonne has been in charge. An investor who cannot say which horizon they mean will feel betrayed by a market that is doing both.
China's return from the early-October holiday is a test of the flow, not of the thesis. If buyers there step back in, the physical bid is still alive and the rate story still caps it. If they do not, the "Asia is hoarding" line will meet a quiet room, and the price will have to find another buyer at a level where cash pays well. Either print is information. Neither print retires the other.
The close
Faith in the dollar has been leaking, in the slow way reserve shares leak, since well before this rally. Producer countries adding refineries and export taxes is a new verse in an old song. Count it. Do not sing it. Laos at 12 tonnes is a policy. Indonesia at a 10 to 15 percent export schedule is a price wedge. The wedge is steepest on unrefined gold. China is the market. It imported more than a thousand tonnes in eight months. It added about 20 official tonnes in August. A Goldman nowcast says the true official buy is larger than that. Central banks at a recent pace near 91 tonnes a month are the scale. A forecast pace near 60 is the cautious version of the same bid. That is why a serious model can still point at $4,650 later this year and $5,400 at the end of 2027, after a nasty drawdown. The Federal Reserve is why the drawdown happened anyway.
Hold those apart and the position sizes itself. The decade can be constructive and the month can be hostile. The investor's edge is not a louder opinion about the dollar. It is a refusal. Do not let a 12-tonne story do the work of a 1,100-tonne story. Do not let a 1,100-tonne story do the work of a rate hike. The metal will reconcile them on its own clock. You do not have to reconcile them in a headline.
A note on sources and limits
The producer-country reporting draws on Nikkei coverage as summarized on October 6, 2026, including the Lao Bullion Bank, the roughly 12-tonne Laos output figure attributed to the World Gold Council and Metals Focus, Indonesia's place among producers, and China's long-standing limits on gold exports. Indonesia's actual October duty is more specific than "up to 15 percent." For October 1 to 14, 2026, the trade ministry set a reference price near $4,336 an ounce and an export benchmark near $139,409 a kilogram. With the reference price above the $3,200 threshold in the 2025 finance regulation, the rates in force were 15 percent for doré, 12.5 percent for non-doré granules, and 10 percent for non-doré bars. Those rates can be changed. Check the decree.
China's import figures are from Heraeus, via Kitco, on September 28, 2026. Imports were 142 tonnes in August and 1,141 tonnes in the first eight months, against 663 tonnes in the same span of 2025 and 940 tonnes in all of 2025. The 1,700-tonne full-year figure is a pace, not a result. Official People's Bank buying of about 20 tonnes in August is the reported number. Published tallies differ on whether that month was the 22nd or 23rd in a row, and on the exact reserve stock, which has been reported in the mid-2,300s of tonnes. Goldman's estimate that China bought about 35 tonnes in July, more than it reported, is an estimate. A larger "true reserve" claim of several thousand tonnes exists in outside research and is not treated here as fact.
Goldman figures are from the bank's precious-metals notes as quoted in the October 6 reporting. They include a seasonally adjusted central-bank pace near 91 tonnes a month. They include a raised forecast assumption of 60 tonnes a month for 2026 and 2027. They include a year-end 2026 fair value cut to $4,650 from $4,900, and a December 2027 figure of $5,400. Société Générale figures on the dollar's reserve share, the survey, and China's gold-versus-Treasury chart are from that firm's work as cited the same week. Desk comments from Gillard, McGeoch, and others are color from those notes, not a house view of this publication. Bank of America and Jefferies warnings are from July and August notes as cited in that same reporting.
The gold price path in this piece follows the October 6 reporting for the January record above $5,500, the late-summer area near $4,700, and the September 28 print near $4,110. Spot near $4,170 on October 6 is from contemporaneous market reports. Different data vendors print different highs. Use one series if you are measuring a drawdown. The September miner move, a 6.4 percent drop in New York gold futures and a larger drop in the big gold miners, is from mining.com's October ranking and is background, not the subject.
This is not investment advice, not an offer, and not a solicitation to buy or sell any security or commodity. Gold can fall hard while a long-term buyer is still active. Export rules change. Central-bank data are revised and sometimes incomplete. Readers should go to the primary notes and official decrees, and speak with a licensed adviser, before any decision.