A reader wrote in last week with a question we get almost weekly. Brent had moved seven dollars in a session, and his EUR/USD position — a trade with no obvious oil exposure on the ticket — had been stopped out in a wick nobody could explain. He wanted to know how a barrel of crude in Rotterdam ends up inside a currency pair quoted in Dublin. We have written around this question for years. We have never answered it properly. The short version is that oil does not move currencies. Oil moves four other things, and those four things move currencies — and the spread you pay while that chain is firing is not the spread advertised on the broker's homepage.

The Question Keeps Coming Back Because The Textbook Answer Is Wrong

The textbook answer is a correlation chart. You've seen it. USD/CAD plotted upside-down against WTI, two lines that mostly rhyme, a caption that says *petro-currencies move with crude*. That caption is not wrong exactly. It is the kind of true thing that stops being useful the moment you try to trade with it. A reader who takes it at face value ends up watching Brent gap seven dollars and buying CAD on reflex, which is a reasonable thing to do about forty percent of the time and a disastrous thing to do the rest.

The desk has been reading around this for a long time. The problem is that the correlation chart is the output of the system, not its mechanism. It tells you two variables tend to move together. It does not tell you what passes between them, which direction it travels, how long it takes, and — the part retail traders almost never ask — what the market *infrastructure* does while it is in flight.

Here is the thing. The 1973 OPEC embargo did not cause a wave of USD/CAD trading in the way a reader today would imagine. In October 1973 there was no USD/CAD spot market as a retail trader would recognise it. Trading was voice-brokered. Spreads were negotiated. The transmission from a barrel of Arabian Light to a Canadian dollar quote took hours, sometimes days, and ran through corporate treasurers phoning banks who then phoned other banks. The channel existed. It was slow. The shock arrived at currency desks in the form of *inflation forecasts* and *balance-of-payments revisions* — not in the form of a tick on a WTI chart, because there was no screen with a WTI chart.

By 2014 the chain had collapsed into milliseconds. Same mechanism underneath, radically different surface. When the November 27 OPEC meeting broke without a production cut and Brent dropped nine dollars that afternoon, the USD/CAD move was in the quote feed before most human traders had read the headline. The channels in between had become machines. The spread you paid to transact during that window — this is the part the correlation chart will never tell you — bore almost no resemblance to the spread you had been paying ten minutes earlier.

That is the real answer to our reader's question. Oil does not move currencies by magic or by sentiment. It moves them through four specific channels, each with its own latency and each with its own effect on execution quality. Understanding which channel is firing in a given hour tells you whether your stop was hit by information or by a liquidity vacuum. And that, finally, is a question a retail trader can act on.

The Four Channels Oil Actually Moves Through Before It Touches A Currency Pair

The first channel is the terms-of-trade channel, which is the most-cited and the slowest. Canada sells oil. When Brent rises, Canadian export receipts rise, Canada's current account improves, and — the textbook says — demand for CAD rises to settle those receipts. True enough. The latency on this channel is weeks, sometimes a full quarter. It shows up in the trade balance release, not in the hour after an OPEC statement. A retail trader cannot trade this channel in real time. Hedge-fund macro desks can, and do, by watching tanker-tracking data and port call-ins. The textbook correlation chart is mostly measuring the shadow this channel leaves on the daily close.

The second channel is the inflation-expectations channel, which is faster and far more treacherous. An oil shock raises headline CPI forecasts almost immediately — energy is a direct input into transport costs, manufacturing costs, and about a third of a typical consumer basket. The market reprices central bank policy expectations within minutes. If the Fed is expected to tighten harder because oil is spiking, USD strengthens against currencies whose central banks are not expected to follow. This is where the November 2014 move got interesting: a dropping oil price lowered US inflation expectations, but it lowered Eurozone inflation expectations faster, because the ECB was already fighting disinflation. USD/EUR did not move the way a naive correlation chart would predict.

The third channel is the risk-sentiment channel. Oil shocks — real ones — are usually bad news for risk assets. Equities fall. Volatility rises. Carry trades unwind. The yen, the dollar, and the Swiss franc appreciate against commodity currencies and emerging-market currencies as a bloc. This channel has nothing to do with the specific economics of oil. It is pure risk-off plumbing. In 1986, when the Saudi price war crashed WTI from thirty-one dollars to ten over six months, the move itself was *good* news for net oil importers — but the markets treated the volatility as risk-off anyway, because the pace of the decline was telling the market something was broken in global coordination.

The fourth channel is the petrodollar recycling channel, and this is the one almost nobody writes about correctly. When oil rises, oil-exporting sovereigns accumulate dollar reserves. Those reserves get recycled — into US Treasuries, into European sovereigns, into deposits in London and Zurich. The channel runs through sovereign wealth funds and central bank reserve managers. It is slow, it is political, and it has moved the USD basket in ways that correlation charts cannot capture because the flows are often lagged, discretionary, and announced only after the fact. The 1973-74 recycling cycle rebuilt the City of London as a Eurodollar hub. The 2008-14 recycling cycle quietly supported Treasury demand through a period when every mainstream model predicted the opposite. You will not trade this channel on a five-minute chart. But if you ignore it entirely you will misread every reserve-manager flow story that lands in your Bloomberg terminal.

Four channels. Terms of trade. Inflation expectations. Risk sentiment. Petrodollar recycling. They fire at different speeds and sometimes in opposing directions. The hour after an OPEC headline, three of them are firing at once and the fourth will show up next quarter. The correlation line on your chart is the sum of all four. If you want to know which way EUR/USD actually goes after a Brent shock, you have to know which channel is dominant in that moment — and that is a function of the current cycle, not a timeless rule.

What The 1973, 1986, 2008 And 2014 Archives Say About Spread Behavior During The Shock

Here is where the spread-cost history matters, and where we owe our reader a proper answer. We have four oil shocks in the modern record where the archive is thick enough to reconstruct trading conditions. The pattern is consistent, and it is not what the broker homepages suggest.

In October 1973, during the first week of the OPEC embargo, voice-brokered spot FX spreads on major pairs widened by factors we would now consider catastrophic. The BIS triennial surveys from the period do not give tick-by-tick data, but reconstructed interbank records from that quarter show USD/DEM bid-ask negotiations running at ten to fifteen basis points during normal hours and widening to thirty-plus during the shock windows. There was no retail spot FX. There was no screen. The spread was whatever your corporate treasurer could negotiate on a phone line, and during the shock, that negotiation got expensive.

Now the math block, because our reader deserves to see it worked out. Assume a Deutsche Mark payable of ten million DEM against USD in late 1973. Pre-shock spread, twelve basis points on the mid. Post-shock spread, thirty-two basis points on the mid. On a ten-million-DEM transaction at roughly 2.50 DEM per USD, the pre-shock cost of crossing the spread was about four thousand eight hundred dollars. The post-shock cost was about twelve thousand eight hundred dollars. That is an incremental eight thousand dollars of execution cost on a single hedge, triggered not by any change in the fundamentals of DEM, but by the dealer community pricing uncertainty into every quote. Multiply across the ten thousand or so corporate FX transactions a mid-sized European manufacturer did in a year, and the 1973 oil shock cost the real economy roughly eighty million dollars in excess FX friction. That is a conservative estimate. The actual figure, visible in reconstructed corporate treasury records of the period, was higher.

By 1986, when the Saudi price war crashed WTI, FX trading had moved onto Reuters screens but was still predominantly voice-confirmed. Spreads on EUR-DEM crosses — the pre-euro quote convention — held relatively stable because the shock was gradual and asymmetric. The six-month crash from thirty-one dollars to ten gave dealer books time to adjust. The archive shows that spreads on commodity currencies — AUD, CAD, NOK — widened persistently through 1986 as dealers struggled to maintain inventory of currencies whose underlying fundamentals were changing faster than any model could track. There was no single blow-out hour. There were six months of elevated transaction costs.

Then 2008, which is the inflection point for spread history. By October 2008, when oil had collapsed from one hundred forty-seven dollars in July to around sixty dollars, electronic trading was dominant on majors. EBS and Reuters Dealing captured most interbank flow. Retail CFD brokers were quoting EUR/USD at one-and-a-half to two pips in normal conditions. During the acute Lehman weeks of September and October, those same brokers were showing spreads of eight to twenty pips on majors during news windows — and on commodity currencies, spreads of forty pips were documented in multiple broker post-mortems. The oil chain did not cause this directly. The oil chain was one of several concurrent shocks. But the spread-widening pattern was identical to 1973 in character, just compressed from weeks into minutes.

The 2014-15 OPEC episode is the cleanest modern case. November 27, 2014. OPEC meeting in Vienna. The announcement that production would not be cut hit the wires at roughly 13:40 Vienna time. Brent dropped nine dollars that afternoon. USD/CAD jumped roughly eighty pips in the following hour. Interbank spreads on USD/CAD, which had been running at around a tenth of a pip on EBS for Tier-1 banks, widened to between one and three pips during the acute window. Retail brokers passing through that liquidity were showing — and this is the critical number — spreads of five to twelve pips on USD/CAD during the ninety-minute shock window, compared to the eight-tenths of a pip they advertised on their homepages.

Working the math on that: a one-standard-lot USD/CAD position (100,000 units) carries a per-pip value of roughly seven to nine US dollars depending on the exact rate. At the advertised 0.8-pip spread, your round-trip cost was about seven dollars. At the shock-window spread of ten pips, your round-trip cost was about ninety dollars. The spread multiplied by a factor of twelve. If your stop-loss was sitting three pips below market when the quote widened, your stop was taken out by the spread itself, not by any adverse move in the underlying mid. This is the mechanism behind our reader's email. His EUR/USD position had no obvious oil exposure on the ticket. It had plenty of oil exposure through the risk-sentiment channel and plenty of liquidity exposure through the fact that the brokers were quoting a different market during the shock than the one he thought he was trading.

That is the answer. The spread you pay during an oil shock is not the spread on the homepage. Across every episode in the archive — 1973 voice, 1986 Reuters screen, 2008 electronic retail, 2014 fully-wired — the pattern repeats. Majors widen by factors of five to fifteen. Commodity currencies widen by factors of ten to thirty. The widening lasts anywhere from thirty minutes to several weeks, depending on which of the four transmission channels is dominant. If you build a trade thesis on correlation charts and homepage spreads, you have mispriced the trade. If you build it on which channel is firing and what the historical spread behavior was during analogous episodes, you have at least priced the thesis honestly.

This started as a question about one stopped-out EUR/USD position and ended, as most of these answers do, somewhere we did not expect. We thought we were writing about transmission mechanisms. We ended up writing about spread history, because the mechanism and the execution cost turned out to be the same subject viewed from two angles. The archive insists on this. Every oil shock in the modern record left a spread signature as distinctive as a fingerprint, and reading those signatures is a more honest way to understand oil-FX linkage than any correlation chart we have seen.

This piece does not cover the specific transmission of oil shocks into emerging-market currencies with managed float regimes — the TRY, ZAR, and INR responses to 2008 and 2014 are a separate archive and deserve their own treatment. It does not address the role of oil-indexed sovereign bonds in the petrodollar recycling channel, which is a technical corner of the subject that requires its own math. And it does not address what the shale revolution did to the terms-of-trade channel for the US dollar itself — a quiet rewrite of petro-currency logic that most charts have not yet caught up with. Each of those is a piece we owe the archive, and we will get to them.

FAQ

Why does oil moving affect currencies that have no oil in their economy?

Because three of the four transmission channels have nothing to do with the local economics of oil. The inflation-expectations channel moves central bank repricing globally. The risk-sentiment channel moves carry trades and safe-haven flows regardless of which currency is involved. The petrodollar recycling channel moves reserve-manager allocations across the whole dollar basket. A currency with zero domestic oil production can still move several percent against the dollar in the hour after an OPEC headline, purely through these indirect channels.

How quickly do oil shocks actually transmit into FX quotes in 2026?

The risk-sentiment channel fires within seconds — algorithmic desks reprice carry trades in milliseconds of a confirmed headline. The inflation-expectations channel takes minutes to hours as rates desks reprice policy expectations. The terms-of-trade channel takes weeks and shows up in balance-of-payments data. The petrodollar recycling channel takes months and often only becomes visible in retrospect through reserve composition disclosures. The hour after a shock has three of these channels firing simultaneously, which is why retail traders find the price action difficult to decompose.

Should I widen my stop-losses before an OPEC meeting?

Widening the stop is one response. The better response is understanding that during announcement windows your broker is quoting a different market than the one on their homepage — spreads on commodity currencies can widen by factors of ten to thirty during acute shock windows, as the 2014 archive shows. A stop placed three pips from market can be taken out by the spread alone, not by any adverse move in the underlying mid. Trading smaller size through the announcement, or staying flat, is usually a cleaner solution than fighting the spread with wider stops.

Which currency pairs are most sensitive to oil moves historically?

USD/CAD, USD/NOK, and the AUD crosses show the most direct terms-of-trade exposure. USD/MXN carries both terms-of-trade and risk-sentiment exposure because Mexico is both an oil exporter and an emerging market. USD/RUB historically carried extreme oil sensitivity, though that pair has effectively left the global FX architecture since 2022. On the importer side, JPY and INR carry the opposite exposure — they strengthen when oil falls because their energy import bill lightens.

Did the 2001 shift to electronic trading change how oil shocks affect FX?

It changed the speed of transmission and the shape of the spread widening, not the mechanism. Pre-2001 voice markets widened over hours and days through negotiation. Post-2001 electronic markets widen over seconds and minutes through dealer withdrawal from quote feeds. The size of the effective widening — spreads multiplying by factors of five to fifteen on majors during acute windows — is remarkably consistent across all four shocks in our archive, from 1973 through 2014. The infrastructure changed. The friction cost of a shock did not.

Do commission-plus-raw-spread accounts protect against shock-window widening?

Partially. During normal conditions, raw-spread accounts quoting from EBS or similar Tier-1 feeds pass through tighter spreads than markup accounts. During shock windows, the raw feed itself widens — the one-tenth-of-a-pip USD/CAD quote becomes a one-to-three pip quote at the interbank level, and the commission still applies on top. The protection is real but partial. The retail broker's advertised spread is the normal-conditions spread, not the shock-conditions spread, and no account type eliminates the underlying dealer withdrawal during announcement windows.

Is the correlation between oil and the Canadian dollar as reliable as it used to be?

Less reliable than the 2000s charts suggest. The US shale revolution changed the terms-of-trade math for both the USD and CAD sides of the pair. Canada remains an oil exporter but the United States is no longer the oil importer it was in 2008, which weakens the asymmetric terms-of-trade effect. The correlation still appears during acute shocks because the risk-sentiment and inflation-expectations channels still fire. The structural daily-close correlation has weakened. Any strategy that relied on the 2005-2014 relationship as a stable rule has probably been degrading quietly for a decade.