in the money for AA- and A-rated bonds is substantial, with that of AA-rated bonds being
significantly larger. In Panel C of Figure 1 we see that the yield spread of BBB-rated bonds
is always significantly higher than the average doomsday spread plus two standard
deviations. Thus, based on our sample, we conclude that the probability of the doomsday call
being in the money for a BBB-rated bond is virtually zero, and thus BBB-rated bonds are
economically noncallable.
***Insert Figure 1 here***
Mann and Powers (2003) study U.S. bonds with make-whole call provisions. They
analyze 318 bonds carrying this provision, issued between October 1995 and September
1999. Similar to our findings for Canadian doomsday bonds, Mann and Powers show that at
issuance the make-whole call option for U.S. bonds is out of the money. Although this is true
for all ratings, lower-rated bonds are deeper out of the money.
Note that the history of the make-whole call provision in the U.S. is significantly shorter
relative to that of the doomsday call in Canada (first U.S. issue is from 1995 while first
Canadian issue is from 1987). If the make-whole or doomsday call option is initially issued
out of the money, the bond yield spread has to shrink over time for the option to move in the
money. In other words, the bond credit rating has to ameliorate or the overall market
conditions have to change significantly. Given the longer history of this option in our
Canadian data, we indeed find that the doomsday option for our higher-rated Canadian bonds
is sometimes in the money.
The presence of the doomsday call has important implications for studying the yield
spread - riskless rate relation. For a standard call provision carried by most U.S. corporate
bonds, lower riskless rates imply a higher probability of the issuer calling the bond. For the
doomsday call of BBB-rated bonds, the effect of lower riskless rates is defused by the call
price floating upwards. Thus, the existence of the doomsday call provides a useful
11