bertilak wrote:The duration is intended to represent what maturity (if issued today) the bond in question would need to make it equivalent to other newly-issued bonds with the same coupon.
No. It is mainly used as a risk measure. Duration and maturity are the same only for zero coupon bonds and it is next to impossible to find replacement bonds (same maturity and coupons). It is really more usefully when comparing bonds with different maturities and different coupons and figuring out which of the bonds is riskier. If bonds have different maturities and different coupons but have the same duration then these different bonds probably have the same risk.
Consider this for a second: A 10 year floating coupon bond which resets every year will have a duration of about .5. Barbecue 1 = 1 year and we assume we are 1/2 through the year. Long maturity but almost no risk.
bertilak wrote:The calculation to reach that number can be made in several ways, with varying assumptions getting various answers.
meh. The formulas are pretty much standardized. However different formuals are used to figure out different things. What are you trying to figure out? That will more or less drive which formula you use. Heck, even what assumptions you use won't affect the end result that much.
bertilak wrote:However calculated, it can be used to compare bonds. You might want to do this to see which bond is the better deal -- (risk-adjusted?) payout vs. price. Some risks (default?, call?, inflation?) are left as an exercise to the buyer.
The general formula for yield = Risk Free Rate (Treasury) + Credit Spread + Option Adjusted
Modified Duration measures the risk to parallel changes in the yield curve. Mainly this is from the change in the risk free rate. It would also capture if the credit spread changed. It would not capture if the bond moved from one ratings to another, say from AA to A.