Bond Futures Part 2: Hedging, Rolls, Squeezes, and Cash Settlement
Fixed Income Trading and Risk Management by Alexander Düring (ISBN 9781119756354)
Previous: Bond Futures Part 1 | Next: Swaps
Part 1 set up basis and conversion factors. Part 2 is where bond futures get weird: hedging ratios that change when CTD switches, rolls that are not one-for-one, and squeezes that show up in repo before they show up in delivery.
Hedging and negative convexity
Futures are the go-to hedge because they are liquid. Match bond PVBP to futures PVBP with opposite sign. Simple in theory.
With a clear CTD, the future behaves like a forward on that bond. Market practice equates futures PVBP to CTD PVBP divided by the CTD conversion factor. Hedge ratio for the CTD itself? Just the conversion factor.
When CTD status is uncertain, things break. Lower yields favor shorter bonds as CTD; higher yields favor longer ones. The futures contract can show negative convexity: PVBP rises when yields fall, opposite of a normal bond. A perfectly hedged long bond / short futures package bleeds money as you re-adjust the hedge through yield moves. That drag is the cost of the quality option, linked to Black-Scholes intuition (option price = expected hedge cost).
Most desks skip the full simulation today because yields sit so far below notional coupons that CTD switches are rare. They treat the future as a CTD forward and move on.
A “fictitious bond” approach (treat Bund as 10Y 6% forward) gives a yield on the future but misstates PVBP versus the actual shorter CTD.
At notification, risk jumps: before notification you run converted CTD DV01; after, you hold actual bond DV01 at full notional. When conversion factors are far from 1, that is a sizeable step change.
Futures rolls
Only a small single-digit percent of contracts go to delivery. Everyone rolls: sell front, buy back, usually as a calendar spread (Roll = Front minus Back).
Bond futures usually trade in backwardation (positive roll) when carry is positive. Contango can appear when front and back CTDs differ.
Roll fair value depends on quality option time decay (more time on the back contract), notional coupon vs market yields in conversion factors, and CTD basket changes (shortest bond dropping out of the back basket).
You rarely roll one-for-one. PVBP-neutral roll ratio = front PVBP / back PVBP, often below 1 because the back CTD tends to be longer. Why does open interest still grow over time? New bond issuance adds duration to hedge.
Near CTD switch levels, roll ratios get messy because front and back PVBP curves kink at different yield levels.
Delivery windows and futures vs cash
Eurex and JPX: single delivery day (10th of month). US and UK: any business day in the delivery month. The short picks timing; late delivery earns carry in a positively sloped curve unless quality option or repo fail risk pushes earlier delivery.
Futures drive price discovery in Europe and Japan. Two decouplings matter:
- Future rich vs CTD (squeeze, negative net basis)
- CTD rich vs surrounding bonds (stress proxy; common in crises)
Japan is special: global index investors proxy JGB exposure via futures and park cash elsewhere. That structurally rich future subsidized non-CTD carry for domestic hedgers. BoJ QE disrupted the pattern but did not erase it.
Squeezes
Open interest often exceeds CTD supply by multiples. Squeezes usually work through repo: buy specials, constrain shorts’ ability to borrow for delivery. Sign: negative CTD net basis and IRR near central bank or unsecured funding rates.
Regulators hate it now. Exchanges use position limits. Free-rider problems make collusion unstable. Mark Twain’s railway short squeeze is the ancestral template.
Cash-settled futures and EFPs
Australian-style cash settlement: PVBP = 1 always, so hedge ratios to the basket move. Fair value still comes from the deliverable basket weighted by yields. Convexity adjustment from daily margining would be tiny at short horizons.
Exchange-for-physical (EFP) in Australia: swap futures position against equal-notional swap, low residual rate risk, EFP spread = swap curve vs futures implied yield.
New issues in the basket
Expected new bonds before delivery matter. If the new issue is obvious CTD, infer its fair price from the future. If not, model the new bond and see if the market is already pricing a CTD switch.
Bond futures look simple from the outside. Düring’s message in Chapter 28 is that delivery options, repo, rolls, and cross-market flows turn them into one of the richest topics in the book. Part 2 is where that richness shows up in risk management, not just pricing formulas.