estimation of parameters two factor CIR

  • Thread starter Thread starter Mitor
  • Start date Start date
Mitor

Mitor

Joined
10/20/13
Messages
17
Points
11
Hello
I'm currently implementing two factor short rate models. I have implemented a one factor Cir choosing the short rate as the 3 month rate (for data reasons) and calibrating the model with the historical average of the yield and volatility using least squares. Since the two factor Cir (r= x+y) is essentialy the sum of two uncorreleted cir processes how should i calibrate them i.e. what is x and y? short, long rate? short and volatility?
Thanks
 
Normally these are the short rate and a longer rate which are uncorrelated.

Brigo and Mercurio discuss the model.
 
I already read the corresponding chapters in Brigo. They sadly do not discuss too much about the details of the implementation. At most, how long can the longer rate be? any criteria? any interesting papers using CIR 2?
 
There is a paper on the internet on the Heston Model: A Practical Approach. It discusses calibration of two factor models, and includes matlab code. Can probably even be altered to fit your specific circumstances.

Not sure if it discusses the pros and cons of various possible market data inputs.
 
Back
Top Bottom