- Joined
- 6/26/14
- Messages
- 1
- Points
- 11
Hi there everyone,
I have a question about the Nelson-Siegel model to model the term structure of interest rates. If I understand it correctly the NS model is used to model a smooth continuous curve through zero coupon yields available at only several maturities. This is done in Diebold and Li (2006) by first creating unsmoothed Fama and Bliss rates, subsequently taking the exponential decay variable lambda as a constant and perform an OLS regression. I don't really understand The Fama and Bliss part, what are they doing here exactly? I presume since zero coupon yields are only available for very long or very short maturities that the yields of coupon bearing bonds for maturities in between are converted to yields of zero coupon bonds to estimate the NS parameters, but I feel I still miss a part.
Another question is, since I don't have access to recent Fama and Bliss unsmoothed yields, if its possible to use other yields in my own NS-model. I have found this data http://www.federalreserve.gov/pubs/feds/2006/200628/200628abs.html which I would like to use. However, I do not know which and why I would need to use a specific zero yield. I miss a little bit of theory on zero yields here...
I have a question about the Nelson-Siegel model to model the term structure of interest rates. If I understand it correctly the NS model is used to model a smooth continuous curve through zero coupon yields available at only several maturities. This is done in Diebold and Li (2006) by first creating unsmoothed Fama and Bliss rates, subsequently taking the exponential decay variable lambda as a constant and perform an OLS regression. I don't really understand The Fama and Bliss part, what are they doing here exactly? I presume since zero coupon yields are only available for very long or very short maturities that the yields of coupon bearing bonds for maturities in between are converted to yields of zero coupon bonds to estimate the NS parameters, but I feel I still miss a part.
Another question is, since I don't have access to recent Fama and Bliss unsmoothed yields, if its possible to use other yields in my own NS-model. I have found this data http://www.federalreserve.gov/pubs/feds/2006/200628/200628abs.html which I would like to use. However, I do not know which and why I would need to use a specific zero yield. I miss a little bit of theory on zero yields here...