Effective Duration for Bonds With Embedded Options

Ordinary duration assumes a bond's future cash flows are fixed, no matter what happens to interest rates — a plain bond promises the same coupons and the same principal at maturity regardless of where yields go. A callable bond breaks that assumption: the issuer has the right, not the obligation, to redeem it early, and issuers exercise that right precisely when it hurts the bondholder most — when rates have fallen and the issuer can refinance more cheaply. Standard duration, computed off fixed cash flows, is simply the wrong tool for a bond whose cash flows themselves depend on the rate path. Effective duration is built to handle exactly this.

Why the standard formula breaks

A callable bond is, underneath, a plain bond that the bondholder has implicitly sold a call option on to the issuer — the issuer can "call away" the bond at a set price after a set date. When rates fall far enough, the issuer calls the bond and refinances at the new, lower rate, capping how much the bondholder's price can rise — the bond behaves less like a 20-year bond and more like whatever's left until the call date. When rates rise, the call is worthless to the issuer and the bond behaves like the full-maturity bond it was issued as. The bond's price sensitivity to rates changes depending on the level of rates, which a single fixed-cash-flow duration number cannot capture.

This produces negative convexity: over the range of yields where a call is likely, the bond's price gains are capped on the upside (as rates fall) while its price losses are not on the downside (as rates rise) — the opposite of the convex, gently-curving-both-ways shape of an option-free bond.

How effective duration is actually computed

Effective duration doesn't try to solve for the option analytically inside a formula. Instead it re-prices the whole bond, option included, twice: once assuming rates fall by a small amount, once assuming they rise by the same amount, using an option-pricing model (commonly a binomial interest-rate tree) that accounts for whether the issuer would call the bond at each shifted rate level. The formula is a straightforward finite-difference slope:

$$\text{Effective Duration} = \frac{P_- - P_+}{2 \times P_0 \times \Delta y}$$

In words: reprice the bond if yields fell (giving price $P_-$) and if yields rose (giving price $P_+$), take the difference, and scale by twice the starting price $P_0$ and the size of the rate shift $\Delta y$. Because $P_-$ and $P_+$ both come from a model that already knows to call the bond when it's optimal for the issuer, the resulting duration number automatically reflects the shrinking upside as rates fall — something a formula based on fixed cash flows could never produce.

A worked example

A callable bond currently trades at $P_0 = 101.20$. Using a binomial tree that models the issuer's call decision, a bond analytics system reprices the bond for a 50 basis point ($\Delta y = 0.0050$) parallel shift in each direction. If rates fall 50bp, the model shows the call becoming likely, and the bond price only rises to $P_- = 102.40$ — far less than an option-free bond would move for the same shift. If rates rise 50bp, the call is irrelevant, and the price falls to $P_+ = 98.70$, a larger move in the other direction. Plugging into the formula:

$$\text{Effective Duration} = \frac{102.40 - 98.70}{2 \times 101.20 \times 0.0050} \approx 3.66$$

Compare that to what the same bond's modified duration (ignoring the call entirely, treating all cash flows as fixed to the stated maturity) might show — commonly 6 or 7 for a bond of similar maturity and coupon. The gap between 3.66 and 6–7 is the option; effective duration is telling you this bond will move noticeably less than its stated maturity suggests once rates fall, precisely because the issuer would refinance it away.

• Effective duration requires a model, not just cash flows. Its accuracy is only as good as the interest-rate tree and the assumed call/refinancing behavior of the issuer feeding into it.
• Mortgage-backed securities face the same issue for a different reason — homeowners "call" (prepay) their mortgages by refinancing when rates fall, giving MBS pools the same negative-convexity shape as a callable corporate bond.
• Effective convexity, computed the same way with a second-order finite difference, is often negative for these bonds right where ordinary bonds are always positive — check the sign, not just the duration number, before assuming a hedge behaves symmetrically.