What is the reason behind using basis point value instead of duration in multiple liabilities immunisation. Also we talk about less dispersion in single liability whereas in multiple liabilities it is convexity of assets greater than convexity of liabilities. Why…
You can use BPV or money duration. I prefer using money duration (it’s how I was brung up), but if they give me BPV in the vignette I’ll use that instead of adding extra steps in the calculation.
As dispersion increases, convexity increases. A single (bullet) liability has the lowest convexity for a given duration (dispersion is zero), so they’re really saying the same thing: you want the convexity of the assets to be as low as possible, but not lower than that of the liabilities.
Can u please elaborate it.
I could probably write a chapter on it.
It would be better if you asked specific questions to which I can give specific answers.
Haha – funny.
I understand for multiple liabilities we would like convexity of A > convexity of L but convexity of assets being greater than convexity of L’s by a minimal amount.
But, what if there is LARGE parallel shift (not just 20 bps)? I would think a safe bet would be to have C of A > C of L again but not by a minimal amount. I would figure having greater dispersion (high convexity – not just minimally greater than convexity of L). of cash flows for the asset portfolio would help in this situation.
Or am I thinking just nonsense.
A little convexity goes a long way; the greater the parallel shift, the greater the benefit from any amount of additional convexity.
The problem with too much additional convexity is that it’s expensive. Think of convexity as insurance: if you’re insuring a $50,000 car, you’d like $50,000 of insurance. If your only choices are $60,000 and $100,000, you’ll buy $60,000; buying $100,000 costs too much. Remember, you’re trying to immunize (insure) the liabilities, not make a killing off of them.