Systematic Yield Curve Steepener

Thesis (edge)

Normally, lending money for ten years pays more than lending for two. You are tying up capital for longer and taking more risk, so you demand more. When that relationship compresses, or flips so that the two-year pays more than the ten-year, something unusual is happening. Usually it means the central bank has pushed short rates high to fight inflation, while the bond market is already looking past that and pricing in the slowdown and the cuts that follow.

Inversions do not last. Eventually the central bank stops hiking, then starts cutting, short yields fall, and the curve steepens back out. The steepener trade is a bet on that resolution: buy the short-dated bond, sell the long-dated one, and profit as the gap between them widens.

The economics are reasonably solid. What makes the trade hard is not the direction but the timing, and the fact that while you wait for the curve to steepen, the position usually costs you money every single day.

Where it works (regimes)

The trade works best in the window after a hiking cycle has visibly ended and before or during the first cuts. That is when short yields fall fast, the curve steepens sharply, and the position pays off in a matter of months rather than years. The steepenings that followed the 2001 and 2008 easing cycles, and the sharp re-steepening in 2024, are the archetype.

It works terribly during the hiking cycle itself. While the central bank is still raising rates, the curve keeps flattening, and a steepener put on because the curve "looks extreme" simply gets more extreme. Anyone who put on a steepener in early 2022 because the curve was already flat spent the next year losing money while being told, correctly, that the recession was coming.

There is also a bear steepening variety, where the long end sells off because of fiscal or inflation concerns rather than the short end rallying. It produces the same profit on the spread, but with a very different risk profile, and a model that cannot distinguish the two will misjudge its exposure.

Signals

• Slope: the yield on the ten-year minus the yield on the two-year. Compare it to its own history over several decades. A slope in the bottom few percent of its historical range is the setup.
• Policy stage: this is the filter that separates a good steepener from a painful one. Is the central bank still hiking, on hold, or cutting? Extremity alone is not enough. An inverted curve with a central bank still hiking is a trap. An inverted curve with a central bank clearly finished is an opportunity.
• Carry of the position: crucially, a steepener when the curve is inverted has negative carry. You are long the low-yielding instrument and short the high-yielding one, which means you pay the difference to hold the position. Know this number precisely, because it is the clock you are trading against.
• Confirmation: require that the curve has stopped flattening, for example that the spread is above its own recent short-term average, before entering. This one rule prevents a large share of the historical losses.

Portfolio construction

Build the trade duration-neutral rather than notional-neutral. A two-year note is far less sensitive to yield changes than a ten-year note, so an equal number of contracts on each side leaves you with an enormous unintended bet on the direction of rates. Scale the front-end leg up so that a parallel move in all yields produces roughly no profit or loss. What remains is a clean bet on the shape of the curve.

Even then, the neutrality is approximate. Curves rarely move in parallel, and in a genuine risk-off event the long end can move in ways the model did not anticipate. Treat duration neutrality as a good first approximation, not a guarantee.

Size the position by the volatility of the spread itself, not the volatility of the underlying bonds. The spread is far less volatile than either leg, which means a given risk budget allows a large notional position, which in turn means the position is highly leveraged in nominal terms. This is exactly how curve trades blow up. Respect the leverage even when the volatility looks small.

Risk model

The dominant risk is the bleed. Negative carry on an inverted curve steepener can run to a meaningful percentage of the position per year. You can be completely right about the direction and still lose money if the resolution takes long enough. This is not a hypothetical: the 2022 to 2024 inversion lasted well over two years, far longer than most participants expected, and it destroyed a great deal of capital held by people whose only mistake was being early.

The second risk is the deceptive smallness of the volatility. Because the spread moves less than the outright bonds, risk models permit large positions. When the curve moves 40 basis points in a week, as it does around policy surprises, the loss on a heavily leveraged spread position is not small at all.

The third risk is the mechanical one: the two legs are different contracts with different liquidity, different roll dates and, in the case of Treasury futures, a deliverable basket whose composition can shift. The hedge ratio you calculated last month may not be the right one today.

Set a hard stop, a maximum holding period, and a rule that reduces size as the carry bleed accumulates.

Costs & implementation

Both legs trade in deep, liquid futures markets, and direct transaction costs are low. Quarterly rolls on both legs add up, and because you are rolling two contracts rather than one, the roll cost is roughly double that of an outright position.

The real cost is the carry, and it belongs in the cost analysis rather than being treated as a market risk. Compute it daily. A steepener with a 60 basis point annual bleed needs the curve to steepen by 60 basis points a year just to break even, which reframes the trade from "the curve is extreme" to "the curve must steepen by this much, this fast, or I lose".

Failure modes

• Entering purely on extremity. The curve being inverted tells you the trade will eventually work. It tells you nothing about when, and when is the entire problem.
• Ignoring the carry. Many backtests of this strategy quietly omit the daily bleed and produce results that are simply not achievable.
• Notional-neutral rather than duration-neutral construction, which turns a curve trade into an accidental directional bet.
• Underestimating the leverage implied by a low-volatility spread.
• No exit rule, which converts a trading position into a slow, silent, permanent loss.
• Assuming every steepening looks the same. A bull steepening driven by cuts and a bear steepening driven by fiscal panic have very different implications for the rest of your portfolio.

Our Notes & Suggestions

The most useful improvement to a naive version of this strategy is the policy filter. Do not put on a steepener because the curve is flat. Put it on when the curve is flat and the central bank has stopped tightening. That single condition transforms the historical distribution of returns, because it moves you from fighting the flattening force to standing behind it.

The second most useful improvement is a carry-aware sizing rule. Start small when the bleed is high and the resolution looks distant, and add as the policy pivot becomes visible. The temptation is to do the opposite, because the trade looks most attractive on a valuation screen precisely when it is most expensive to hold.

Treat this as a patient, macro-aware position rather than a signal you can systematize entirely. It is one of the few strategies where a mechanical rule with no regard for the policy context reliably loses money over long stretches, even though the underlying economic logic is sound.