Estimating MBS Duration and Convexity for Swaption Hedging
Summary
The question concerns inputs for studying whether mortgage-backed securities hedging affects swap-rate volatility. It describes a proposed ordinary least squares regression with expected 10-year swap volatility as the dependent variable and mortgage-market convexity, duration, and refinancing activity as explanatory variables. The researcher can access a refinancing index but cannot locate the historical effective duration and convexity series for the mortgage index used in the cited paper.
The answer suggests checking whether the Bloomberg or Barclays mortgage index provides relevant data, while cautioning that duration may be available without convexity. It also proposes estimating empirical convexity directly for the hedging study. A further caveat is that the market regime has changed since Fannie Mae and Freddie Mac were major convexity hedgers, so historical relationships may not carry over. The exchange does not provide a data source, a calculation procedure, or evidence that the suggested index fields reproduce the paper’s inputs; these remain research tasks.
Key ideas
- The proposed regression relates implied swap-rate volatility to MBS convexity, duration, and refinancing activity.
- A mortgage index may offer duration data without a corresponding convexity series.
- Empirical convexity estimation is one possible route when index data are unavailable.
- Changes in major mortgage hedgers may limit the relevance of historical relationships.
- The answer does not establish a definitive source or method for matching the paper’s dataset.
Tags
Full text
# MBS Market Duration & Convexity # MBS Market Duration & Convexity Soft question...hopefully. I am working on a swaption hedging strategy. Part of this strategy calls for a forward looking indication of changes in implied volatility, using 1m10y implied as a proxy for movements across the rest of the surface. I came across the following paper which suits my scenario and approach well: Mortgage Hedging and Rate Volatility; Roberto Perli & Brian Sack Perli and Sack use three variables to test whether the hedging activities of MBS holders has a material impact on swap rate volatility. In summary, they specify a Ordinary Least Squares regression with expected (implied) volatility of the 10Yr swap rate as the dependent against three explanatory variables: - The convexity of the MBS Market - The duration of the MBS Market - The amount of Mortgage refinancing I have Terminal access so I can use the MBA US Refinancing Index for the Refinancing independent. Perli and Sack state that their data set includes weekly observations from "the effective convexity and duration of the Merril-Lynch master mortgage index" but I can't find references to this index anywhere. I've tried various other searches for a composite MBS Market Duration/Convexity index, to no avail. Any ideas as to where I could get data for the MBS Duration/Convexity inputs? Thanks! ## Answer by GoPackGo (score 1) https://quant.stackexchange.com/a/45425 I'd be curious if you can get the convexity for the Merrill index. Bloomberg/Barc MBS index only shows duration on the terminal. That index would be my first choice. Regardless, it might be most useful to calc the empirical convexity yourself to study a swaption hedging strategy. Additionally, keep in mind the regime shifted dramatically since Fannie and Freddie (once the largest convexity hedgers in swaptions by far) became smaller.
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.