Excess Spread Return for Spread-Based Bonds

$$\displaystyle \text{Excess Spread}_0 \approx \text{Spread}_0 - ( \text{EffSpreadDur} \times \Delta \text{Spread}) $$ How would you interpret the negative coefficient in front of the term with the change in spread?

Interest rate movements and credit events are two different things. Bond portfolio managers know this and manage these two things separately (as they should).

No, sorry. That wouldn't make sense. Widening spreads have to reduce returns.

If the change is positive, the spread has widened, and that's a negative impact on return. If spreads narrow, then the bond must be increasing in value, increasing return. That's shown here by the double negative of "minus the negative change" in spreads. So it's consistent with logic.

Where it's not consistent with logic is in the lack of any default variable. It assumes zero defaults, and that assumption won't allow you to be a credit portfolio manager for long. So change excess spread return to _expected_ excess spread return, and add in the probability of default (POD) and loss given default (LGD): $$\displaystyle E[\text{ExcessSpreadReturn}] \approx \text{Spread}_0 - ( \text{EffSpreadDur} \times \Delta \text{Spread}) - (\text{POD} \times \text{LGD}) $$. What do you think is the goal of active credit management in terms of this value?

No, actually it's to maximize this expected excess spread.

You got it!

Now you're looking at yield net of default losses, so you'll definitely be looking for as much of this as possible.

Yes!

For fixed-rate, option-free bonds, the spread duration is very similar to modified duration. But for floaters, it isn't. Floating-rate bonds adjust to interest rate changes; that's a macro thing. But it doesn't adjust for spread changes from changes in default probabilities. Think about how each would affect the duration measures. How would you expect a floater's duration measures to differ?

So if a bond manager is focused on these components of return and decides to hedge interest rate risk, what excess return is still available?

Not really. If the interest rate hedge was thorough, this will be gone as well.

No, sorry. Yield curve changes have been hedged.

Exactly. This is what *excess return* means in the realm of credit securities.

What does that mean for the credit spread of a security in terms of how it may relate to these excess returns?

No, that isn't it. Credit spreads mean higher yields due to credit risk. Consider how that relates to excess returns.

No, actually. A higher credit spread couldn't provide a lower return for investors; no one would buy them.

That's right. The excess spread depends on the initial yield spread, the effective spread duration, and the change in the spread:

Yes!

That's not it. Modified duration would have to be smaller.

Since the coupon adjusts to interest rate changes, floaters have a very small modified duration. But their price will react to credit issues just like any other, so the spread duration remains larger and similar to that of a fixed-rate bond. Finally, you can calculate effective rate duration and effective spread duration for an FRN in a very similar way to other effective duration calculations. Since there's a periodic reset of market reference rates (MRRs) in both the top and bottom of the FRN, the rate duration is about zero for many floaters, but using the MRR and the Z-DM (it's shown as just DM here) you can use both of these calculations: $$\displaystyle \text{EffRateDur}_{\text{FRN}} = \frac{(PV_-) - (PV_+)}{2 \times (\Delta MRR)(PV_0)} $$ $$\displaystyle \text{EffSpreadDur}_{\text{FRN}} = \frac{(PV_-) - (PV_+)}{2 \times (\Delta DM)(PV_0)} $$

To summarize: [[summary]]

Widening spreads reduce returns

Narrowing spreads reduce returns

To maximize it

To minimize it

Convexity

Declined yield curve

A premium for credit risk

Credit spreads are unrelated to excess returns

Credit spreads are inversely related to excess returns

Credit spreads are positively related to excess returns

Continue

Modified duration would be smaller than spread duration

Spread duration would be smaller than modified duration

Continue