How Unfixing a Floating Stub Changes Swap DV01
Summary
The document considers two ways to calculate DV01 for an interest rate swap when its first floating payment has already fixed. One method holds that stub payment constant under a curve shock; the other lets it move as part of the recalculation. The question concerns a long-dated swap with a shorter remaining term and separate discounting and projection curves.
An illustrative bucketed sensitivity shows that including the near-term stub can shift total DV01 relative to the sensitivity of the longer swap exposure alone. The answer suggests that allowing the stub to move may be more useful when its volatility is low, or when comparing swap sensitivities with bond sensitivities, since bonds have no analogous floating stub. The response explicitly presents this reasoning as a guess, not a definitive market convention or general rule; the appropriate treatment depends on the risk measure being sought.
Key ideas
- Swap DV01 can be calculated with the first floating stub held fixed or allowed to move under the curve shock.
- Including the stub contributes a short-dated sensitivity that changes total DV01.
- A sensitivity excluding the stub may be more comparable to bond delta.
- The proposed rationale is tentative and does not establish a universal convention.
Tags
Full text
# Answer by dm63 (score 1, accepted) # For a DV01 calculation on swaps, why "unfix" the fixed stubs on the libor instruments while bootstrapping the libor curve? When calculating DV01 for a (portfolio of) swap(s) (of various maturities), one can either keep the first fixing period constant while shocking the curve, or one can let it move along. I don't quite understand why one would not keep the first floating payments constant, but apparently the market consensus it to use the non-constant method (at least for LDI managers in the Netherlands), probably. (One other explanation for what I'm seeing in the DV01s as calculated by various parties could also be that they are using curves for the DV01 calculation built with tenors starting at 1Y) Could somebody shed some light on this matter? Most online information regarding DV01 calculations is super basic. And the books that I've read didn't mention the non-constant method, as far as I can remember. Note that I'm looking at a 20y swap (of which 17y are left) that pays libor 6m floating, and it's discounted on the ester. curve. ## Answer by dm63 (score 1, accepted) https://quant.stackexchange.com/a/83997 This is just a guess but consider the following : suppose you have a 100mm 10yr swap with a 3 month stub which has just fixed. Then the bucketed dv01 of the trade is something like: 10yr: -70,000; 3 month : +2500; Total delta : -67,500 If you unfix the stub when you calculate dv01 you will get -70,000. If you fix it you will get -67,500. So it depends what you want. Arguably, the -70,000 may be more relevant because the stub may have low volatility if central bank is on hold. Also, if you are combining swap deltas with bond deltas , then the 70,000 is more like the bond delta because bonds have no stub. As I said, just a guess.
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.