"Crude at ninety dollars" sounds like information and mostly is not. It is a level. It says what a barrel costs on one day, and says nothing about the question a macro reader needs answered: is the market short of barrels right now, or long of them?
That question has a published answer, and it is not the price. It is the shape of the futures curve.
Two Curves, One Price
Crude trades as a strip of dated contracts, and the relationship between them has a name. Backwardation is a downward-sloping curve — the front month above the deferred months, someone paying more for a barrel now than for the promise of one half a year out. Contango is the reverse: the market pays you to wait.
The level is free to be anything in either case. Ninety dollars in steep backwardation and ninety dollars in deep contango are not the same world. The first is a market bidding for immediate physical supply; the second has barrels it does not currently need.
The standard account is about storage. Holding a barrel costs tankage, insurance and financing, and that cost pushes deferred prices above spot — contango is the ordinary shape of a market with room in its tanks. Pulling the other way is the option value of actually holding the barrel: a refinery that can run tomorrow, a trader who can meet a physical obligation on short notice. Ample inventories make that option close to worthless; tight ones push it past the carrying cost, and the curve inverts.
So the sign of the curve reads present physical tightness. It is not a forecast — a backwardated curve is not predicting higher prices, since its deferred contracts are pricing lower ones.
Roll Yield: The Shape With a Price Tag
Any futures position held past a contract's life must be rolled: the expiring month sold, the next bought. In backwardation that roll is mechanically favourable: the position sells the richer front contract and buys the cheaper deferred one. In contango it runs the other way and compounds every cycle regardless of what spot does.
Annualized, that is the roll yield: steepness converted into a rate. A few dollars between two contracts means little on its own, and means something quite different if those contracts are one month apart rather than six. Annualizing collapses both facts into one comparable number.
Where This Reaches Currencies
Transmission runs through terms of trade. A country that sells more energy than it buys sees export earnings move with crude; one that buys more sees the opposite. Canada sits on one side of that line, the euro area and Japan on the other — which is why ENERGY is one of the framework's four driver words, and why USD/CAD's named driver is energy. The import side has its own long-form treatment in the EUR/USD energy article.
What the curve adds is a reason to distinguish two rallies that look identical on a price chart. Crude rising while the curve steepens into backwardation is a supply story: barrels are harder to source and the terms-of-trade shift lands quickly. Crude rising while the curve stays in contango is closer to an inventory or expectations story, and can persist without much happening to anyone's balance of payments.
The Arrow Points the Other Way
Here is the finding that should discipline all of it. Chen, Rogoff and Rossi (2010) examined whether commodity-currency exchange rates and global commodity prices forecast each other, and found the relationship strongly asymmetric: exchange rates predict commodity prices with what they describe as remarkably robust power, in-sample and out-of-sample; the reverse direction, commodity prices forecasting exchange rates, they find notably less robust.
That inverts the everyday narration — oil moves, therefore the Canadian dollar moves. The currency is a forward-looking asset priced continuously by people with a view on the whole export complex, and it tends to move first. Which is why the energy channel is read as one structural condition among several rather than as a signal standing alone: turning a curve state into a currency call needs the causal arrow pointing the direction the evidence finds weakest.
What the Monitor Publishes, and How to Read It
The FX Monitor's energy channel is published rather than described, so it can be checked. It carries a channel status — Dormant, Active or Dominant, set by how far crude has moved over the last week and month — a WTI price with both changes, and a label naming which upstream fed it. Beside that sits a term-structure object: state, conviction, annualized roll yield, a plain-sentence narrative, and the date the read reflects. The two are kept deliberately separate, so that losing the futures source does not take the price channel down with it.
On 3 September 2026 that object read steep backwardation, the front contract roughly fourteen dollars above one six months out, and the spot channel read Dominant. That snapshot has a date on it and will be stale by the time most people read this sentence; the monitor carries the live one.
Three habits keep curve-reading from becoming curve-storytelling. The shape is a statement about now — reading it as a price forecast imports a claim the curve never made. The flip is the information: a move from contango into backwardation says something changed in physical supply, while months of steady backwardation say only that it is still there. And the number depends on the contracts — a steep front-month dislocation and a gentle six-month slope can annualize alike and mean different things, so read the spacing with the rate.
None of that produces an instruction. It produces a condition: the physical crude market is tight, or it is not, and currencies with energy in their terms of trade are exposed either way. What follows — whether it matters for a given pair, at a given horizon, against everything else the transmission chain is saying — is the reader's call.
References
- Chen, Yu-chin, Kenneth Rogoff, and Barbara Rossi. "Can Exchange Rates Forecast Commodity Prices?" The Quarterly Journal of Economics 125, no. 3 (2010): 1145–1194. Oxford Academic · NBER Working Paper 13901