The arrangement in plain terms
The post war monetary system agreed at Bretton Woods in 1944 fixed exchange rates against a dollar that was itself convertible into gold at an official rate. That convertibility was suspended in 1971, which left the world with a reserve currency backed by nothing but the depth of the markets and institutions behind it. The obvious question at the time was what would anchor demand for it.
Part of the answer arrived through energy. Through the 1970s, oil exporters accumulated very large surpluses, and those surpluses were substantially recycled into dollar denominated assets, particularly government debt. Oil continued to be invoiced in dollars. Out of that came the shorthand term petrodollar, and out of the shorthand came a theory far stronger than the facts underneath it. Notice what was never written down anywhere in it: an obligation on anybody to keep holding the currency once a cargo had been paid for.
Invoicing is not holding
Here is the breaking point. The currency written on an invoice is a convention. It reduces friction, because everyone quotes the same benchmark and nobody has to renegotiate a currency on every cargo. It does not oblige anybody to hold that currency for any length of time.
A buyer with any major currency can convert into dollars in a market deep enough to absorb the transaction in seconds, pay the invoice, and hold nothing. A seller receiving dollars can convert immediately into whatever they prefer. The invoicing currency therefore creates a flow through demand that lasts as long as a settlement cycle. What creates lasting demand is the separate decision about what to do with the surplus once it exists. That is a savings decision, and it is made for entirely different reasons.
What actually sustains reserve currency status
If invoicing is not the mechanism, what is? Mostly the quality of the asset market a surplus holder can put money into. A reserve currency needs an enormous pool of safe, liquid, easily transferable instruments that can be bought and sold in size without moving the price. It needs predictable law and enforceable contracts. It needs a payments and trade finance network that counterparties already use, because network effects in finance are extremely sticky.
It also needs the issuer to be willing to supply the asset, which in practice means running external deficits so the rest of the world can accumulate claims. That combination is unusual and not quickly replicated. Any serious challenger has to offer somewhere comparably deep to park a surplus, with comparable certainty that the money can be retrieved. That is a far higher bar than agreeing to settle a trade in a different currency.
The channel that is genuinely real
Strip away the exaggeration and a real mechanism remains. Reserve managers have to decide what to hold, and that decision has been shifting. A reserve held as a claim on another government is only as reliable as the willingness of that government to honour it. Once reserves have been frozen in any dispute anywhere, every reserve manager in the world updates their assessment of that risk, whatever their own politics.
An asset with no issuer and no counterparty answers that specific concern. It cannot be frozen if it is held domestically, it carries no credit risk and it does not depend on anyone's permission to be sold. This is the honest core of the story, and it is visible in official sector behaviour rather than in oil invoicing, a subject covered in central bank gold buying. It is also slow, lumpy and reported with a long lag.
Why settlement headlines are not regime change
Announcements that two countries will settle some bilateral trade in their own currencies arrive regularly and are routinely presented as the end of an era. They are usually solving a payments problem, not a savings problem. Settling in local currency can reduce transaction costs, avoid a sanctioned payments channel or support a domestic policy goal.
The question that follows is the one that matters: what does the surplus holder do with the local currency they have now accumulated? If the counterparty currency is not freely convertible and there is no deep domestic bond market to invest in, the surplus holder is left holding an asset they cannot easily deploy. That is precisely the problem that pushes reserve managers towards assets with no issuer in the first place. So these agreements can be mildly supportive of the metal, through a channel almost the opposite of the one usually claimed.
What this does and does not do to the chart
It does almost nothing intraday. Reserve allocation shifts happen in quarterly reports, not in a London session. A trader who takes a position because of a settlement headline is trading sentiment about a decade long process, which is a poor match for any stop that fits a normal risk framework.
What the channel does change is the composition of demand underneath the market. A persistent, price insensitive buyer that is not comparing the metal with a bond yield alters how the market behaves on dips, and it helps explain stretches when the usual real yield framework fits badly. That is useful context for interpreting a chart, not a reason to enter a trade. Day to day, the metal still answers mostly to real returns and to the currency, as covered in gold and the US dollar.
Holding the idea honestly
The reasonable version of this thesis has three features. It is structural rather than cyclical. It has no timing and no identifiable catalyst. And it is partly observable already, through the behaviour of official buyers, which means some of it is reflected in the market rather than waiting in the future.
The unreasonable version attaches a date, a mechanism it has not established, and a price objective. Be particularly careful with the step that goes from a diversification trend to an inevitable collapse in the reserve currency. Diversification can proceed for a very long time while the dollar remains the dominant reserve asset, because the alternatives have to be built, not merely wished for. Keeping the idea in the portfolio frame it belongs in, as discussed in is gold a good investment, is more useful than trying to express it through the live chart on a short horizon.
FAQ
Did the petrodollar arrangement end?
The informal arrangement of the 1970s was never a single treaty with an expiry date, so there is nothing clean to point at. Oil is still mostly invoiced in dollars and surpluses are still substantially held in dollar assets, while the share of alternatives has been rising slowly. It is a gradient, not an event.
Would oil priced in another currency crash the dollar?
Unlikely on its own. Invoicing creates only a brief, pass through demand because conversion in deep currency markets is cheap and instant. The lasting demand comes from where surpluses are saved. A change in invoicing without a change in saving behaviour is mostly an administrative detail.
Then why do central banks buy gold?
Mainly to hold a reserve asset with no issuer and no credit risk, which answers a specific worry about reserves held as claims on another government. Diversification away from concentration in any single currency is the other reason. Neither motive has much to do with the currency on an oil invoice.
Can I trade this story?
Not on any horizon a stop can cover. It describes a slow change in the composition of demand, with no catalyst and no schedule. It belongs in a decision about how much exposure to hold over years, and it should not be used to justify an entry on a single session.
Is sanctions risk already priced into gold?
Partly, and nobody can say how much. Reserve managers have been visibly adjusting for a while, and markets generally discount a trend once it is widely discussed. What cannot be priced is the next escalation, which is exactly why some holders treat the asset as insurance rather than as a trade.
ⓘ See these ideas on real price: open the free XAUUSD live chart.