Interest Rate Risk and Credit Spread Curves
There's always interest rate risk, and for investment-grade bonds, higher interest rates mean lower prices. No surprise there. For high-yield bonds, credit risk is a much bigger concern than it is for investment-grade bonds. No surprise there, either.
But then consider which portfolio, investment grade or high yield, would be more sensitive to interest rates. The answer may surprise you. To get there, start with something beyond the simple theory of a coupon bond: credit spreads. When times are good, interest rates can be high. What do you think would happen to credit spreads?
No.
That would reflect more default risk, which doesn't really rise when conditions are good.
Absolutely.
Good economic conditions reduce default worries, and so credit spreads. When times are bad, policymakers lower rates, and spreads widen to reflect greater uncertainty. So there's a negative correlation between interest rates and spreads.
No, they are.
Think about what a credit spread represents and how that would change when times are good.
Consider what this means for the price change of a high-yield bond, given an interest rate change in some direction. What do you think the credit spreads would do to the bond's price?
Exactly!
No.
Just the opposite. The price would move less.
Since higher interest rates correlate with lower spreads, and lower interest rates correlate with higher spreads, those riskier bonds should actually see some "protection" from changes, in a way.
Credit spread changes generally follow the credit cycle. Credit spread levels are often stable in recoveries, fall in late economic expansions, rise during economic peaks, and then peak during economic contractions. So the credit cycle tends to look like a shadow of the business cycle, with each comparable phase coming later for credit.
Now formalize this idea. A risky 10-year bond with a modified duration of 9 should see its price change a certain amount. But __empirical duration__ is the actual, observed price effect of this bond regressed on changes in some risk-free, 10-year rate over time.
How should the risky bond's empirical duration compare with its modified duration?
Yes!
And here are the effective durations and empirical durations of bonds categorized by Moody's rating spectrum:
Voilà.
No.
That would contradict the discussion to this point. Recall the negative correlation between interest rates and credit spreads.
No.
They are related, just in the way mentioned previously. Recall the negative correlation between interest rates and credit spreads.
You'll see that each empirical duration is lower than the related effective duration, but that last one is really something. It's negative! Consider again the forces that cause the lower empirical durations. How would you justify a negative empirical duration for these junk bonds?
No.
Interest rates will always have an effect on the price of a typical bond.
Precisely.
The idea of effective duration is basically "assume interest rates drive prices, and ignore spreads." If spreads were constant, these two duration measures would be the same. But the more credit risk matters, the more that negative correlation between rates and spreads takes effect. This is a general relationship that lower credit spreads are related to higher empirical durations. The relationship between option-adjusted spread (OAS) and empirical duration is also negative.
So if you ever calculate the estimated price change of a high-yield bond using its modified duration, you just may be surprised at the error.
No.
Credit spreads are very important here. In fact, if you think about credit spreads being constant (which they aren't), then that empirical duration would jump up to where the effective duration is.
To summarize:
[[summary]]
They should widen
They should narrow
They should be unaffected
They would make the price move less
They would make the price move even more
Empirical duration should be lower
Empirical duration should be higher
Empirical duration wouldn't be qualitatively related
Interest rates have no effect on their price
The credit spread changes dominate the interest rate changes
Credit spreads aren't very important here, since bonds are already closer to default
Continue
Continue
Continue
