The 20-Year US Treasury Yield Curve Kink: Stripping Economics & Butterfly Arbitrage

Plot the par US Treasury yield curve across long maturities and a striking structural anomaly jumps off the screen: the 20-year bond consistently yields 4.74%—42 bps above the 10-Year Benchmark Note (4.32%) and 13 bps higher than the longer-duration 30-Year Long Bond (4.61%). Reintroduced in May 2020 after a 34-year hiatus, the 20-year sector suffers an acute "habitat vacuum" that leaves the 10s20s30s par butterfly (2*Y_20Y - Y_10Y - Y_30Y) persistently positive at +55 bps despite quarterly auction cuts and $4B Treasury Buybacks.

Preferred-Habitat Segmentation and the CME Futures Orphan Problem#

Why do institutional buyers crowd into 10-year and 30-year points while demanding a concession to hold 20-year bonds? Three microstructural frictions reinforce the kink:

1. No Dedicated Liquid CME Futures Contract: CME launched a 20-Year T-Bond future (TWE) in 2022, yet open interest remains a fraction of Classic T-Bond (ZB, 15Y-25Y basket where 15Y-16Y bonds are CTD) and Ultra T-Bond (UB, 25Y-30Y basket). Fresh 20-year new issues sit stranded in the middle—not Cheapest-to-Deliver into ZB for another four years.
2. LDI Convexity Efficiency: Pension LDI managers prefer 30-year Principal STRIPS (29.3 years duration, 886 convexity) over 20-year bonds (12.95 years duration, 206 convexity), creating rich reconstitution demand at 30Y.
3. Index & ETF Concentration: Multi-asset macro hedgers benchmark duration at 10Y while long-duration Treasury ETFs (such as 20+ year funds regulated by the SEC) rebalance mechanically across the whole 20Y-30Y bucket.

Why Stripping Arbitrage Caps the 10s20s30s Fly Near +65 Bps#

If 20-year par bonds cheapen too far above the interpolated 10Y-30Y zero-coupon curve, primary dealers such as Cantor Fitzgerald execute coupon-strip vs principal-strip arbitrage: buying cheap cash 20-year bonds on FINRA TRACE, stripping them via Fedwire into 40 semiannual coupon strips plus a 20-year Principal STRIP, and selling the generic 1Y-20Y coupon strips into rich 30-year bond reconstitution packages.

However, because the 20-year Principal STRIP itself cannot substitute for a 30-year Principal STRIP, dealers must warehouse the residual 20-year PO on balance sheet—imposing a no-arbitrage bound around +60 to +72 bps on the 10s20s30s par butterfly even after Treasury scaled 20-year new issues down to $16B ($13B reopenings).

DV01-Neutral 50:50 Butterfly Weighting & Convexity Adjustment Math#

Constructing a market-neutral long-belly 10s20s30s butterfly requires matching both total portfolio DV01 to zero and isolating pure curvature roll-down from parallel slope shifts.

Long-End Benchmark Comparison: 10Y vs 20Y vs 30Y Microstructure#

Comparing auction sizes, futures hedging liquidity (CFTC), and buyback support shows why real-money income portfolios at firms like Alpine Woods harvest highest sovereign par yield at the 20-year point.

Harvesting Roll-Down from 20Y into the Classic ZB CTD Window#

Seasoned 20-year bonds issued in 2021-2022 represent one of fixed income's cleanest multi-quarter roll trades: as a 20-year bond seasons past 4.5 years and approaches 15 years and 3 months remaining to maturity, it suddenly becomes Cheapest-to-Deliver into the ultra-liquid CME Classic T-Bond (ZB) contract—compressing its spline cheapness by 8 to 14 bps as leveraged basis funds accumulate deliverable float.

Historical calibration series link NY Fed ACM 5-Factor vs Kim-Wright Term Premia, Holston-Laubach-Williams vs Lubik-Matthes Neutral Rate (r*), and FOMC Policy Path, SOMA Runoff & US Treasury Term Premia directly to FOMC Policy Path and verification registry.