Yield curve spread trades—flatteners and steepeners—let you profit from how different parts of the Treasury curve move relative to each other. They're precise, hedged, and absolutely brutal to size correctly in Excel. Here's how Sourcetable AI handles the duration math in seconds.
Andrew Grosser
February 24, 2026 • 14 min read
October 2023: The 2-year Treasury yields 5.08%, the 10-year yields 4.98%. The 2s10s spread sits at -10 basis points—a mildly inverted curve. You believe the Fed is done hiking, that rate cuts are 6 months away, and that short rates will fall faster than long rates when the pivot comes. The trade is obvious: a 2s10s steepener. Go long the 2-year, short the 10-year in duration-neutral proportions. But calculating those proportions requires knowing the modified duration of each leg, solving for notional amounts that equalize the DV01, and stress-testing the position across a dozen rate scenarios. In Excel, that's an hour of fixed income math—DURATION() functions, nested formulas, scenario tables—before you've placed a single order.
Excel breaks on yield curve trades because the math compounds. Duration-neutral sizing means the dollar gain on one leg must precisely offset the dollar loss on the other for any parallel shift. Calculating DV01 (dollar value of a basis point) requires duration, price, and notional working together across both legs. Scenario analysis requires rebuilding the whole model for each rate assumption. Change your notional and every number needs recalculating. Most traders skip the rigor and use rough sizing—which turns a spread trade into an unintentional directional bet. sign up free.
Both strategies are simultaneous long/short positions in Treasuries at different maturities. They profit from changes in the spread between yields—not from the direction of rates overall. A well-structured trade is duration-neutral: if rates move in parallel (all yields up or down equally), the position neither gains nor loses. It only profits when the spread changes.
The two core structures:
Here's a concrete example. You want a 2s10s steepener with $10 million notional in the 2-year leg. The 2-year Treasury has modified duration of approximately 1.9 years; the 10-year has modified duration of approximately 8.5 years. To be duration-neutral, the notional on the 10-year leg must be scaled so that its DV01 matches the 2-year's DV01.
The math: DV01 = (Modified Duration × Price × Notional) / 10,000. For the 2-year at par with $10M notional and duration 1.9: DV01 = (1.9 × 100 × 10,000,000) / 10,000 = $19,000 per basis point. To match that DV01 with the 10-year (duration 8.5 at par): $10,000 / (8.5 × 100) × 10,000 = $2,235,294 notional on the 10-year leg. So you need: long $10M 2-year, short $2.24M 10-year.
Now do all of that while also running:
That's four separate analytical workflows requiring fixed income math expertise, historical data lookups, and scenario modeling. Miss the duration neutrality calculation and your spread trade becomes an unhedged rate bet. Sourcetable eliminates the gap between knowing what to do and being able to do it.
Upload your Treasury yield data—from Bloomberg, FRED, or your broker—and the AI speaks fixed income fluently. You interact with the trade the way you'd talk to a desk analyst who already understands DV01, duration neutrality, and basis point values.
In Excel, you'd write DURATION() and MDURATION() formulas for each leg, calculate DV01 for the anchor notional, then back-solve for the offsetting leg's notional using Goal Seek. In Sourcetable, just ask: "Size a duration-neutral 2s10s steepener with $10 million in the 2-year."
The AI returns the complete structure: Long $10,000,000 2Y at 5.08% (Duration 1.91, DV01 $19,100). Short $2,247,059 10Y at 4.98% (Duration 8.49, DV01 $19,072). Note: DV01 values are matched—the position profits purely from spread changes, not parallel shifts. Follow up: "What if I anchor to $5M in the 10-year instead?" → instant recalculation, new notionals, matched DV01.
Knowing the spread is -10bp isn't enough—you need to know whether that's extreme or normal. Ask Sourcetable: "Show the 2s10s spread over the past 20 years and tell me where today's level ranks historically."
The AI calculates the historical spread time series, marks current levels, and reports: "Today's -10bp sits in the 18th percentile—significantly inverted but not at the extremes seen in 2022–2023 (-90bp). Over the past 20 years, spreads below -10bp have steepened by an average of 65bp within 12 months in 7 of 9 instances." That's the context that turns a trade idea into a data-driven thesis.
Ask: "What's my P&L if the 2-year drops 75bp and the 10-year drops 25bp?" The AI calculates: 2-year leg gains (75bp × $19,100/bp = $143,250). 10-year leg loses (25bp × $19,072/bp = $47,680 loss on short position, but short profits as rates fall... wait, you're short the 10-year, so 10-year rally hurts you). Net: $143,250 - $47,680 = $95,570 profit. The 2s10s spread steepened 50bp (75bp - 25bp), exactly what you wanted.
Request a full scenario grid: "Show P&L for all combinations of 2-year moves (-25 to -100bp) and 10-year moves (-10 to -75bp)." Sourcetable generates the heat map showing your profit zones in green and risk zones in red—the visualization that makes position sizing obvious.
Curve trades are primarily macro bets on Fed policy and growth expectations. The setup that favors each direction:
Fed tightening cycle (early-to-mid phase): Short rates rise faster as the Fed hikes, while long rates are anchored by growth/inflation skepticism. Classic 2004–2006 and 2022 setup.
Strong growth with contained inflation: Long rates stay low (low term premium) while short rates rise with Fed normalization. Spread compresses as shorts catch up.
Risk-off flight to quality: Long Treasuries rally harder than short in crisis, compressing the spread. Though this is often a carry-negative steepener setup in disguise—be careful.
Spread at historically wide levels: When 2s10s exceeds +150bp (steep normal curve), flatteners offer mean-reversion opportunity with history on their side.
Fed cutting cycle (early phase): Short rates fall faster as the Fed eases, while long rates are sticky due to inflation uncertainty. The classic post-inversion steepening dynamic—2001, 2008-09, 2019-20.
Deep inversion: When 2s10s is below -50bp, historical mean reversion strongly favors steepeners. The deeper the inversion, the more asymmetric the upside.
Rising inflation expectations: If inflation breakevens widen, long rates rise faster than short rates anchored by Fed credibility. Steepener profits from the long-end selloff.
Fiscal expansion concerns: Large Treasury issuance in long maturities pushes long yields up, steepening the curve. A risk for any flattener holder.
The key decision variable is carry. Steepeners are typically positive carry (you earn more on the long 2-year than you pay to finance the short 10-year) when the curve is inverted. Flatteners carry negatively in that environment. Sourcetable calculates your carry automatically: ask "What's the daily carry on this steepener if I hold it for 3 months?" and the AI factors in coupon accruals on both legs.
Yield curve spread trades (flatteners and steepeners) profit from changes in the yield differential between two maturities—not from the direction of rates overall. Duration neutrality ensures the position only profits from spread changes.
Duration-neutral sizing requires calculating DV01 for each leg and solving for the notional that equalizes them. In Excel, that means DURATION() formulas, Goal Seek, and manual scenario tables. In Sourcetable, it's a single question.
Historical spread context is essential: knowing whether today's -10bp 2s10s is at the 18th or 80th historical percentile determines whether the trade has asymmetric upside or is entering crowded territory.
Steepeners are the natural trade after deep curve inversions and ahead of Fed cutting cycles. Carry is typically positive when the curve is inverted, making the position earn while you wait for the spread to normalize.
Sourcetable handles sizing, historical context, scenario modeling, and carry calculation in plain English—replacing the four separate Excel workflows that make spread trading inaccessible without dedicated fixed income systems.
If your question is not covered here, you can contact our team.
Contact Us