Most traders learn the four yield curve shifts in the wrong order — and never recover from it. Hear us out. The textbook sequence teaches bull steepening first because it feels intuitive: rates fall, the front end falls faster, the curve steepens. But the FX desks that read curve moves as a leading indicator for currency flows do not start there. They start with bear flattening, because bear flattening is the shape that shows up first when a central bank surprises the market — and the surprise, not the direction, is what moves the pair. This piece reconstructs the four shifts from that vantage point.
The Pattern Nobody Names Correctly on the First Try
There is a pattern this desk keeps seeing in reader mail and in the way junior FX analysts describe curve moves in their first year at a bank. They see the two-year yield rise by twelve basis points and the ten-year rise by four, and they call it "flattening." Technically correct. But when asked *which* flattening — bear or bull — a majority hesitate. The vocabulary is not intuitive. Bull and bear here do not refer to direction of price. They refer to direction of yield, and the direction of price on the underlying bond is the inverse. A bear move in bonds is a rising-yield move. A bull move is a falling-yield move. The flattening or steepening refers to the spread between the short end and the long end.
Four combinations. Bear flattening: yields rise across the curve, short end rises faster. Bull flattening: yields fall across the curve, long end falls faster. Bear steepening: yields rise across the curve, long end rises faster. Bull steepening: yields fall across the curve, short end falls faster. That is the mechanical definition. It fits on an index card. It also tells you almost nothing about why the shape matters.
What matters is what each shape *implies* about who is doing the buying and the selling, and about what the market believes will happen next. A bear flattening does not just describe a curve. It describes a specific belief structure — that the central bank is about to tighten more aggressively than the long-end holders had assumed, and that the tightening will slow the economy enough that ten-year yields do not need to rise proportionally. The four shifts are four different theories about the near-future path of monetary policy, translated into a slope.
The reason FX desks care is the transmission. A currency pair is priced off interest rate differentials at multiple points along both countries' curves. A bear flattening in the front end of the dollar curve while the euro curve does nothing is a tightening of the dollar-euro two-year differential. Front-end differentials drive carry flows in the fast money segment. Long-end differentials drive real-money asset allocation. When they diverge — when one end of the curve is telling you one story and the other end another — the currency reaction depends on which segment of positioning is heavier at the moment. That is why the desk starts with bear flattening. It is the shift that surprises the market, and the surprise trips the fast money first.
Bear Flattening: When the Front End Punishes the Long End
The pattern: a central bank meeting, or a data print, or a speech, delivers a hawkish surprise. Front-end yields — the two-year, the three-year — jump. Long-end yields rise as well, but by less. The curve flattens in a rising-yield environment. That is bear flattening.
The mechanics are worth slowing down for. Front-end yields move because the market re-prices the near-term policy path. If the market had been pricing three hikes over the next twelve months and a hawkish surprise implies four, the two-year yield — which is essentially an average of expected short rates over that horizon plus a term premium — rises. The long end contains that same expectation but also contains a view on where the neutral rate sits, on where inflation lands in five to seven years, and on the term premium demanded to hold duration risk. A single hawkish surprise does not move those distant anchors much. The two-year absorbs the shock. The ten-year barely blinks. The spread narrows.
For FX, this is the classic dollar-supportive setup when it happens to the US curve. Dollar carry becomes more expensive to short. Positioning in short-dollar carry trades — funded by cheap dollars against higher-yielding EM — comes under pressure. The pattern this desk observes in reader questions is that traders often notice the dollar rally but attribute it to "risk-off" or to a "flight to quality" narrative, when in fact the mechanism is more prosaic. The front-end differential widened. Carry positioning had to be trimmed. Dollars were bought to close shorts. Risk-off narratives layer on top, but the curve move is the leading indicator.
There is a secondary reading that matters more. A bear flattening that persists — that does not reverse over the following weeks — is the market telling the central bank that its hawkishness is credible on the near-term path but excessive on the medium-term outlook. In extreme cases the ten-year yield falls even as the two-year rises. That is not bear flattening anymore. That is an inversion in motion, and it is the market pricing recession. FX pairs read that transition differently again. The dollar can strengthen on the initial bear flattening and weaken on the inversion because the long end is signaling that the tightening will require reversal.
Bull Flattening: When the Long End Refuses to Believe the Front
The pattern: yields fall across the curve. The long end falls faster than the front end. The curve flattens in a falling-yield environment. That is bull flattening.
This shape is the market's way of telling the central bank that the growth outlook has deteriorated but the central bank has not yet responded. The front end cannot fall much because the policy rate is still where it was at the last meeting. The long end is free to reprice on any horizon it wants. If ten-year yields drop thirty basis points on a weak payrolls print while two-year yields drop only ten, the market is saying: we now expect the terminal rate to be lower, we expect cuts sooner or deeper than we thought yesterday, and we are willing to lock in duration at these levels because we expect yields to be lower still in six months.
The FX transmission here runs through the real-money channel more than the fast-money channel. Long-end yields are what pension funds, insurance companies, and sovereign wealth funds care about. When the long end of the dollar curve rallies (yields fall) faster than the long end of the euro curve, the dollar's attractiveness to a Japanese pension fund with a thirty-year liability profile diminishes. The hedged yield pickup shrinks. Rebalancing flows out of dollar assets follow, though on a lagged timeline — quarterly rebalancing cycles, not intraday reactions.
The primary documents this desk reads for bull flattening episodes are BIS quarterly reviews from the periods immediately preceding recession calls. The BIS staff repeatedly identify bull flattenings that preceded formal recession dating by two to four quarters, and they identify the specific dynamic where the ten-year yield led the two-year lower — the market pricing the response function before the central bank delivered it. In FX terms, the currency reaction is muted at first because the differential move is dominated by the long-end drop that most fast money desks are not positioned in. The dollar drifts rather than breaks. Then the central bank capitulates, the two-year catches up, and the pair moves in the direction the long end had been signaling for weeks.
Every curve shift is a disagreement between what the central bank has already said and what a specific segment of the bond market believes will happen anyway.
Bear Steepening and Bull Steepening: The Two Shapes That Get Confused
The pattern this desk sees confused most often is bear steepening. Not because the definition is hard — long end rises faster than the front end in a rising-yield environment — but because two very different macro stories can produce the same shape.
Story one: reflation. The economy is accelerating, inflation expectations are rising, but the central bank is either behind the curve or explicitly tolerating an overshoot. The long end rises to price higher inflation over the coming decade. The front end rises less because the central bank has committed to slow policy adjustment. This is the "good" bear steepening. It shows up during recoveries, during coordinated fiscal-monetary reflation, and — in the historical record cited by BIS working papers — during the post-1971 dollar depreciation cycle where long-end US yields ran ahead of Fed action for years.
Story two: fiscal deterioration. The long end is rising because the market is demanding more term premium to hold sovereign duration, not because growth is accelerating. This is the "bad" bear steepening. The front end is anchored by an unchanged policy rate. The long end reprices on supply concerns, on debt sustainability worries, on a downgrade risk that has entered the conversation. The 2022 UK gilt episode was this exact shape in extreme form — the long end of the gilt curve moved faster and further than the front end in the days after the mini-budget, and the currency reaction was violent because the FX market read it as story two, not story one. A steepening driven by supply fear is currency-negative. A steepening driven by growth acceleration is currency-positive. Same shape. Opposite trades.
Bull steepening is less frequently confused but no less important. Yields fall across the curve, front end falls faster. This is the shape that appears when a central bank pivots — when the market has priced in the pivot and the front end drops sharply on the announcement or the anticipation. The long end also falls, because lower short rates over the immediate horizon mechanically pull down the average that anchors the ten-year, but it falls less because the medium-term inflation and neutral-rate anchors are unchanged. Bull steepening episodes tend to be currency-negative for the currency doing the pivoting, particularly against currencies whose central banks have not yet capitulated. This is the shape that dominated the dollar's behavior during the successive Fed pivots between 2019 and the pandemic response, and the shape that fast-money desks watched for again during every subsequent tightening cycle.
The distinction between these four shapes matters because the same absolute yield change in the ten-year can mean four different things depending on what the two-year did. A twenty basis point drop in the ten-year is a rally. Whether it is bull flattening or bull steepening depends entirely on the front end. And the FX trade — buy the dollar, sell the dollar, wait — is different for each.
The Contradiction Between Central Bank Language and Curve Reaction
There is a recurring pattern in the primary record this desk finds most instructive. A central bank publishes a statement. The statement is parsed by analysts as hawkish. The curve reacts in a way that does not match. The contradiction is where the interesting analysis lives.
Consider the shape of the disagreement. Bank of England minutes, Federal Reserve press conference transcripts, and ECB accounts frequently contain forward guidance language calibrated to produce a specific curve shape — a bear flattening, typically, because central banks prefer front-end tightening without long-end distress. The market then trades the curve. When the shape that emerges does not match what the language was calibrated to produce, one of two things is true: either the market disbelieves the guidance, or the guidance was misread.
The BIS annual reports track this specific mismatch. Their staff routinely identify episodes where central bank communication was interpreted as hawkish by sell-side analysts while the curve produced bull flattening — the long end rallied more than the front end fell, which is the market saying "you are hawkish now but you will be dovish soon." The FX pair reads the curve, not the transcript. Sell-side commentary that quotes the transcript to justify a dollar-bullish call while the curve is bull flattening is fighting the tape. This desk has watched that mismatch play out repeatedly across the 2018-2019 cycle, the 2022-2023 cycle, and the current environment.
The primary-document cross-reference this desk finds most useful: read the central bank's own quarterly monetary policy report or inflation report, then read the BIS quarterly review commentary on how the market received it. The two documents will often disagree in tone even when they discuss the same event. The central bank report describes the intended policy stance. The BIS commentary describes the priced-in market response. The gap between the two is the trade — the curve shape tells you which side of the gap the market has taken, and the FX pair follows the curve, not the central bank.
So What Do You Actually Do
Start with the front end. Every curve analysis begins with the two-year yield of the country whose currency you are trading, because the front end is where surprises show up first and where central bank pricing lives. If the two-year has moved more than five basis points on any given day, the curve has repriced and the FX pair has almost certainly followed. Check the two-year before you check the pair. This inverts the habit most traders develop of watching the pair first and rationalizing after.
Then classify the shape. Rising yields with faster short-end move: bear flattening, likely dollar-supportive if the move is in the US curve, likely a signal of imminent hawkish action. Rising yields with faster long-end move: bear steepening, and the critical follow-up question is whether the driver is reflation or fiscal deterioration — read the auction data, read the ten-year term premium estimates that the Fed and the ECB publish, and if the term premium is doing the work, this is a supply-driven move and the currency is at risk. Falling yields with faster long-end move: bull flattening, the market is ahead of the central bank on recession pricing, the currency reaction will be muted until the central bank catches up. Falling yields with faster short-end move: bull steepening, the central bank has pivoted or is about to, currency-negative in the near term.
Finally, do the work of reading the primary sources. This desk has argued repeatedly that sell-side summaries of central bank statements are compressed to the point of losing the signal. The full text of the Fed's Summary of Economic Projections, the ECB's monetary policy account, the BOE's Monetary Policy Report — these documents contain the language the curve is actually reading. The BIS quarterly review contains the interpretation from the desk that gets it right most often. If you are trading FX on curve moves, these are the documents that pay the rent. Everything else is derivative commentary.
This piece did not cover the mechanics of forward-rate calculation or the specific math of term-premium decomposition — both are separate exercises that require their own space and their own primary sources. It did not cover the credit-curve interactions with the sovereign curve, which matter enormously for EM currencies but require a different framework than the developed-market rates work above. And it did not name specific historical episodes in the depth this desk usually reserves for reconstruction pieces — the Plaza Accord and its Louvre follow-up, the 1994 bond massacre, the 2013 taper tantrum, the 2022 gilt crisis, each of which is a case study in one of the four shifts and each of which deserves its own reconstruction.
FAQ
Why do FX desks watch the two-year yield before they watch the currency pair itself?
The two-year yield reprices the near-term policy path faster and with more information than the currency pair, because the rates market has deeper liquidity in the front end and reacts to central bank data or communication within milliseconds. The FX pair follows the differential. A trader watching the pair first is reading the second derivative of the signal — the two-year move already contains what the pair is about to do.
What separates a "good" bear steepening from a "bad" one in practical terms?
The driver is the entire distinction. A bear steepening driven by rising real-growth expectations and rising break-even inflation is currency-supportive because it signals economic strength that the central bank has not yet fully priced. A bear steepening driven by rising term premium — visible in the Fed and ECB's own term-premium decomposition models — is currency-negative because it reflects supply concerns or fiscal deterioration. Read the term-premium data before you name the move.
Can the curve invert without any of the four shifts occurring first?
Inversion is not a fifth shift — it is a state that the curve reaches after enough bear flattening or bull flattening in the same direction. A curve that has been bear flattening for months eventually inverts when the front-end rise overtakes the long-end rise. Bull flattening episodes reach inversion when the long end has fallen far enough below the anchored front end. Inversion is the outcome, not the mechanism.
How does the BIS quarterly review differ from a central bank's own communication as a source?
The BIS review is written by staff whose job is to describe how markets received central bank actions, not to justify the actions themselves. Central bank publications describe intended stance. BIS publications describe realized market response. When the two disagree — and they often do — the curve is siding with one of them, and the FX pair follows the curve. Reading both is how you locate the disagreement.
Why does bull flattening produce a muted currency reaction at first?
Bull flattening is dominated by long-end buying, and the long end is held by real-money accounts — pension funds, insurers, sovereign wealth funds — that rebalance on quarterly cycles rather than intraday. Fast-money FX desks are not positioned in ten-year duration, so the initial curve move does not force any positioning trades. The currency drifts until the central bank confirms the pricing with an actual cut, at which point the front end moves and the pair breaks.
Does the same shape mean the same thing across every developed-market curve?
The mechanics are the same but the interpretation shifts. A bear flattening in the US curve is dollar-supportive because the Fed sets the global funding rate. The same shape in the JGB curve is complicated by Bank of Japan yield-curve control and by the historical anchoring of the front end. A bear flattening in the gilt curve since 2022 is read partly through the fiscal-supply lens. The four shifts are universal categories, not universal trades.
What documents should a trader read to interpret the curve, in priority order?
The central bank's most recent full monetary policy report — not the press release, the full document. The BIS quarterly review chapter on the currency area in question. The Fed's Summary of Economic Projections or its overseas equivalents. Auction data from the sovereign debt management office for supply pressure diagnosis. The term-premium decomposition series that major central banks publish. These five inputs cover more than sell-side commentary aggregated across a whole quarter.
How often does the shape flip from one shift to another within a single week?
More often than most traders assume. A bear flattening produced by a hawkish central bank meeting on a Wednesday can transition to a bull flattening by Friday if a weak data print reprices the medium-term outlook. The two-year yield stays elevated because the meeting already anchored it; the ten-year drops on the data. The shape flips. Currency reaction follows the second flip, not the first. Watch for the transitions — the shape at the end of the week is the one the pair traded on.