As of the end of February, the spread on the investment grade corporate bond index was ~170 bps and the high yield index ~650 bps (up from the 1994-2010 average of 140 bps and 510 bps respectively). Below is a chart of the variance between the two since 1994 (as far back as I was able to pull data).
Below is the relationship between this variance and the subsequent 12 month out/under performance of high yield relative to investment grade corporate bonds. As can be seen, there appears to be a weak relationship between spread variance and performance when spreads are "tight" (in this case less than 600 bps [in blue], where correlation is -0.31), but when spreads were north of 600 bps [in red], correlation spiked to 0.75 (though it is important to note that spreads have only been this wide in two periods and 19 months since 1994 [i.e. small sample set]).
Below is a chart of the above data in a different form broken out by spread "bucket" rather than as a scatter plot. Again, except when spreads were at "world is ending" levels (and the world didn't end), high yield has tended to underperform.
The important point I will make is that any investment in high yield is by definition... risky. Especially at relative "tight" levels coming out of the worst credit crisis since the Great Depression.
Source: Barclays Capital
Very nice analysis!
ReplyDeleteI would add one caveat to any historical analysis of spreads, namely that junk is junkier now than in the past, that is, the proportion of lowest-rated bonds is now much higher than the historical average. So the spreads are not completely comparable over time.
This just strengthens the conclusion that junk is not a particularly promising bet right now.
Here is a post of mine that covers the ratings distribution issue.
Jim Fickett
ClearOnMoney.com