Hi,Let's consider of a portfolio of various fixed income instruments. Is there a straight relationship / formula between the portfolio convexity and the individual convexities of its underlying instruments?Thx.

If you are speaking of Macaulay convexity, it is linearly additive, so just weight by market value to get the portfolio convexity.

Hi,Never heared about Macaulay convexity, I only know of Macaulay duration. I was referring to Convexity as Gamma (second derivative of the portfolio price function). Frankly, I intuitively doubt that the Portfolio Gamma would be the weighted average Gamma of its underlyings. My case portfolio is made of underlyings having optional features (callable bonds, CBs, etc.)Any idea?Cheers,

I am currently working on portfolio convexity and I have the same problem : convexity can't be linear regarding the callable features of the convertible bonds.What I think is that you should take three points (i.e. three values for the underlying basket), compute them and calculate the portfolio convexity thanks to the three values obtained.A question rises then : how to take the three points ?A first approach would be to take a volatility spread up and down.A second one would be to divide the space (in n-dimensions) into equal-probabilistic spaces, but for that you need to inverse the gaussian cumulative density.Or you can do as everybody does in the market, i.e, you take [-sigma%,+sigma%] on each stock to compute the three points.If you have any other idea, plesae tell me, I am really interested in this problem.