-
When trying to immunizate the liabilitie with assets, we need to match the effective duration or the modified duration?
-
And why the below statement is incorrect?
“immunization protects the portfolio from the reshaping of yield curve.”
Thanks!
Based on what I’ve studied so far, I am going to answer your questions - 1. We need to match the effective duration of liabilities with the assets. Almost all the definitions in the CFAI text mention effective duration and effective duration is a better indicator of interest rate risk than modified duration. 2. The statement is incorrect because classical single period immunisation assumes parallel yield curve shifts. i.e. the interest rates across all maturities change by the same amount. Reshaping of the yield curve usually means , interest rates change by different amounts, (so the shape of the yield curve changes) Hope that clears it. 
If the bonds in your portfolio have cash flows that don’t change when their YTM changes, then effective duration is the same as modified duration; the only circumstance in which they’re different from each other is if the cash flows can change with changes in yield (e.g., callable bonds, putable bonds, floating-rate bonds). As a practical matter, it’s probably very unlikely that you would use such bonds for an immunization portfolio.
What’s irritating is that the curriculum uses the time to maturity of a liability as that liability’s duration. That’s Macaulay duration, not modified or effective duration. You cannot mix them together; Macaulay duration is not a measure of interest rate risk.
If you have a portfolio of liabilities or a portfolio of bonds, modified or effective duration assumes the same change in YTM at all maturities.
If you wanted to account for reshaping of the yield curve, you could match key-rate durations instead of matching modified or effective durations.