A curve move breaks into three pieces: level, slope and curvature. Slope is where the trade is, and 2s10s, the 10-year yield minus the 2-year yield, is the gauge everyone quotes. Two questions sort out the rest. Are yields falling or rising? Falling is bull, rising is bear. Is 2s10s getting wider or tighter? Wider is steepening, tighter is flattening. Cross the two and you get four regimes: bull steepener, bear steepener, bull flattener and bear flattener. Which one you are in tells you which end is leading, 2s at the front or 10s and 30s at the back.
The curve plots yield against maturity. Any move in it mixes three things. Level is the whole curve going up or down together. Slope is the front and long ends moving by different amounts. Curvature is the 5-year belly moving against the 2s and 10s wings. Traders quote slope as spreads. The 2s10s spread is the default on most desks. 3m10s is the Fed's preferred recession gauge, and 5s30s is the cleaner read on the long end and term premium.
An upward sloping curve, positive 2s10s, is the normal state, because investors want more yield to lock money up for longer. When 2s10s prints negative the curve is inverted. That says nothing about the level of rates, which can be high or low at the time. It is a statement about slope. The market is pricing the front end above the long end because policy is expected to fall from here. Read the spread in basis points, then look at where it is heading, which matters more than where it sits.
The 2-year is close to a pure bet on the expected path of the fed funds rate over the next couple of years. It lives and dies by the FOMC and the near-term data. The 10-year is roughly the average expected short rate over ten years, plus a term premium, which is the extra yield investors want for holding duration risk: supply, inflation uncertainty, fiscal risk. So 2s10s is a contest between where policy is headed near term and where the market thinks rates settle in the long run, term premium included.
That split explains most of what the curve does. When the Fed is hiking, the front end gets dragged up faster than the long end, the curve flattens, and eventually it inverts. When cuts come into view the front end rallies hardest and the curve re-steepens. A move led by the long end is usually a term premium story instead: a reflation scare, a heavy duration supply calendar, or a fiscal wobble lifting 10s and 30s for reasons that have nothing to do with the near-term policy path.
A steepener is a widening of 2s10s. A flattener is a narrowing. Desks put the trade on DV01-neutral, or duration-weighted, which means each leg is sized by what it makes or loses on a 1 basis point move, so the position is a bet on slope and not on level. A 2s10s steepener is long the 2-year and short the 10-year, sized so a parallel shift in yields nets out to roughly zero P&L. Get the weighting wrong and you have bought or sold duration by accident, on top of your curve view.
Carry and roll are part of the trade. On an inverted curve a steepener is negative carry, because you are long the higher-yielding front end and short the lower-yielding long end. You bleed a few basis points a month while you wait to be right, so timing the catalyst matters. Flatteners on a steep curve have the opposite problem. Know whether the market is paying you or charging you to hold the slope on.
Bull means yields down and prices up. Bear means yields up and prices down. That is standard bond convention. The catch is that a steepener or a flattener can happen in either one, and the label depends on which end of the curve is doing the work. A rally led by the front end, with 2s falling faster than 10s, is a bull steepener. A rally led by the long end, with 10s falling faster than 2s, is a bull flattener. Turn it into a selloff and the leader still sets the label: 10s rising faster than 2s is a bear steepener, and 2s rising faster than 10s is a bear flattener.
The read itself is mechanical and comes down to 3 checks. Are yields net rising or falling? That is bear or bull. Is 2s10s widening or tightening? That is steepen or flatten. Then whichever end moved more is the one driving. A move led by 2s is almost always a Fed story. A move led by the long end is almost always term premium, inflation or supply.
Yields are either falling or rising, and the gap between the 2-year and the 10-year is either widening or narrowing. Put those two together and you get four combinations. Traders have a name for each.
When yields fall and the gap widens, the front end is doing the work and 2s rally hardest. That is the market pricing Fed cuts, and you normally see it at the start of an easing cycle, while people still think a soft landing is possible. Traders call it a bull steepener.
When yields rise and the gap widens, the long end is doing the work instead, and 10s and 30s sell off hardest. Investors are demanding a bigger term premium to hold that duration, because they are worried about inflation, heavy supply or fiscal risk. That is a bear steepener, and it goes with reflation or a fiscal worry.
Yields falling while the gap narrows is the mirror image. The long end leads and 10s rally hardest. Usually that is money running into duration for safety in a growth scare, or long-run inflation expectations coming down. That is a bull flattener, and it turns up late in the cycle, when the bid for duration is a risk-off one.
Yields rising while the gap narrows is the front end taking the hit, with 2s selling off hardest, because the market is pricing hikes from a hawkish Fed. That is a bear flattener, and it is what the middle of a tightening cycle looks like.
The shortcut worth remembering is that the front end tracks the Fed and the long end tracks inflation and term premium. That is what separates the two steepeners: one is a rally in 2s because the Fed is turning dovish, the other is a selloff at the back on inflation and supply. The flatteners split the same way, a growth scare bidding up duration at the long end against a hawkish Fed knocking the front end down. Watch which end moves first and you know which of the four you are in. Nearly every headline-driven curve move drops cleanly into one of them.
A negative 2s10s, a yield-curve inversion, has preceded every modern US recession, which is why it gets so much airtime. The lead is long and it varies, often 12 to 24 months, so inversion is a warning and not a timing tool. The inversion itself is the product of a bear flattener and a bull flattener: the front end held up by a restrictive Fed while the long end prices eventual cuts and slower growth.
The un-inversion is what catches people out. The curve usually re-steepens through a bull steepener, with the front end collapsing as the Fed starts cutting into a slowdown, and that move has historically lined up with the recession arriving rather than with the all-clear. So a curve climbing back above zero is not a risk-off signal to fade. Check what is doing the steepening. A front end rallying on cuts is late cycle and a reason to be cautious. A long end selling off on reflation is early cycle and risk-on. The spread moves the same way in both. The regime is the opposite.
The curve moves on catalysts, and the skill is pinning the move on one end or the other. A hot inflation or jobs print that lifts the front end more than the long end is a bear flattener, with the market pulling hikes forward or pushing cuts out. One that lifts the long end more is a bear steepener, a repricing of term premium and inflation risk. Dovish FOMC guidance or soft Fed speak steepens the curve through the front end. Supply counts too. A weak 30-year bond auction is a long-end concession that bear-steepens the curve, the same read covered in how to read a Treasury auction.
Helious keeps the pieces in one place. The rates board shows 2s10s and the other spreads ticking in basis points, and the feed flags Fed speak and data catalysts the moment they hit. Release breakdowns put each surprise on a sigma z-score, so you can see if a print is a 2s story or a long end one, and the auction module scores every tenor's takedown seconds after the results. Watch whether 2s or the back of the curve is leading, and you have the regime before the desk chat has finished typing.
