BGM Calibration to Caplets and Swaptions
Caplet quotes constrain volatility of individual forward rates. Swaption quotes also constrain how several forward rates move together. Co-terminal swaptions share an end date and provide a particular set of calibration instruments. Review a fit by examining price errors, the validity of the implied correlation matrix, and the stability of parameters when quotes or instrument weights change.
How it works
This calibration estimates forward-rate volatilities and dependence from caplet and swaption prices. Caplets mainly constrain individual forward variances. A swaption depends on a weighted combination of forwards, so its quote adds information about co-movements. Co-terminal swaptions share a swap end date and give one practical calibration strip.
Worked example
Suppose two forwards with the same 20% volatility have equal weights in a simplified swap-rate approximation. The volatility of their average is about 14.14% at zero correlation and 17.32% at correlation 0.5. The caplet volatilities are unchanged in this calculation. Market calibration uses time-dependent weights and each actual payoff.
Validation and limitations
Check positive semidefiniteness and the units of every quote. Report residuals per instrument and re-fit after removing selected swaption quotes or changing weights. Held-out prices and hedges distinguish calibrations that have similar in-sample errors. See correlation matrix repair.
Further reading
- The Market Model of Interest Rate Dynamics
Alan Brace, Dariusz Gątarek and Marek Musiela · Journal record
- Computing the Nearest Correlation Matrix
Nicholas Higham · PDF