I am constructing a zero curve for a super national obligor for the period 2007 to 2015.
I am using the universe of bonds available on Bloomberg to construct the zero curve from maturities 3m to 50 years.
I believe that the bonds are of the same quality. However, lack of liquidity is an issue for the bonds where no price is available on a COB date given it did not trade. In this case, if no price exists for that bond, the algo will use the next available bond that has a price for that bond.
The coupon rates and prices are varied in the universe of bonds i am using to construct the curves.
The output of the algo is that it creates 'jumps' or spikes for each tenor. For example, for the 20 year tenor time series, the zero rate will jump 100 basis points one day, then fall another 100 basis points on the next day and so on which clearly is not representative of yield changes in the market and is likely due to the algo/quality of bonds
My question is that could the Price/coupon of the bond be responsible for this jump? If the bond being used to bootstrap the 20year tenor has a coupon of 3% and price at 104 is missing a price for a business day, the algo would jump to the next closest bond which has a coupon of 10% and price of 180, would that result in the jump/spike?