Two commenters at Victor Niederhofer's Daily Speculations had some thoughts on the August 2 post "Somebody Tries to Predict, from Alston Mabry":
...Russ Sears writes:
High yield bonds returns are not normal. This is because there are periods of very little volatility and then periods of high outliers. There were 2 days with in the financial crisis with more than 10% losses in the ETF JNK. That is about -12 standard deviations away from the mean (the std dev is 0.88% mean of 0.010% per day) And one day in the recovery or 15.6% returnsNeither are returns "random" as there are average spreads, over corporate bonds are predictive.
What does this have to do with the article. Well this suggest that junk bonds by definition are companies struggling to find financing. Hence in each crisis they are not diversified but will tend to behave like the struggling sector. Each recession causes and recovery are different. Last recession the struggling sector got bailed out.
But care and very long term studies with many recessions are needed to reach definitive conclusions. But yes, high spreads with widening spreads probably mean hard times and low spreads and lowering spreads are probably signs of a bull market.
I would add the real question is are junk bond "spreads" a leading indicator of recession (better than stock indexes) concurrent or lagging indicator. The financial crisis seems to imply it's a leading indicator… however, bond guys generally are insiders of financial institutions. Bonds are not as liquid as stocks and hence unless one is using an ETF (not developed till 2007), one is comparing a theoretical class backwards put together returns or a sub sector of the class from a managed fund. In short because of the cyclic/non-random nature and because of the very fat tail/non-normal distribution, one must be highly versed in statistic or perhaps economics to use junk bonds in any way to allocate assets.
anonymous writes:
I haven't read this paper, but my guess is it captures a psychological/structural phenomena which may not be systematically useful:
Making a gross generalization, I believe that most bond investors are pessimists and most stock investors are optimists. The most a bond investor ever earns is his coupon plus par, whilst there is no limit to what a stock investor can earn. Because of this tail asymmetry (to which Russ sort-of alludes), it is plausible that bond investors may be faster to pull the rip cord on a problem investment. But that doesn't mean that the bond guys are right. It just means their expected value shifts in a different way. And once a bond price has fallen away from par and into the range of distress, it trades like a stock, not a bond. So this is a dynamic process too.
SpecList has had many posts about companies whose bonds are priced for default/impairment yet the stock prices still reflect positive equity /going-concern value… certain mining stocks (that subsequently filed for bankrupty) and Valeant are just two recent examples. But which came first, the chicken or the egg???
