The mysteries of ABX index pricing became a whole lot more understandable to me after I read Freddie Mac’s (FRE) February 28 report discussing its subprime exposure. In a nutshell (and despite what most people think) the ABX wasn’t designed to represent all AAA-rated subprime ABS that have been issued recently, but rather just the longest-duration, highest-risk ones. Combine that fact with the recent spread widening that’s occurred throughout the financial markets lately (which of course hurts the longest-duration bonds the most) and uncertainty surrounding the ultimate credit losses, and you can see why the 07-1 AAA ABX is lately trading at around 62.
It’s worth taking a look at the summary of the Freddie report directly. It includes a section, starting on page 4, on how the enhancements in deals works (an area too often ignored) and, on page 6, why Freddie’s subprime bond exposure (which is essentially exposure to all subprime AAAs, in aggregate) is much different than the ABX’s exposure.
What is in the AAA ABX indices
First off, the simple fact is that not all subprime AAAs are created equal. The way deals are set up, some AAA-rated classes of MBS have an earlier call on loans’ cash flows than other AAA-rated classes do. The table below, which shows two bonds I randomly picked from the ABX 07-1, provides an example. It shows that the longest-duration, lowest-rated AAAs are the ones that have been included in the AAA ABX indices.
click all images to enlarge
What this means for losses
The ABX’s huge skew toward long-duration classes that are at the back of the cash flow line has a huge impact on the potential losses imbedded within the indices. The trusts are structured to pay down by class, which means that the lowest-rated AAA bond, will be the last to get principal payments generated by the loans in that particular MBS. Thus a small loss for the AAAs in the aggregate equates to a large loss for the bonds included in the ABX index.
For example, if you assume 50% default and 50% severity, the AAA bonds in aggregate might see a 5% loss, Freddie assumes. However, under that same scenario the AAA bonds in the ABX (the longest-duration, lowest-rated ones) would see a 33% loss.
One caveat here is that Freddie’s analysis does not appear to include triggers that would re-direct cash flows from the subordinate bonds and excess spread to the AAA bonds. This would add roughly 10% subordination to the AAAs (see section “Excess interest within the trust” portion of the “How Subprime Credit Enhancement Works” subsection on page 5 of the Freddie report.) Accordingly, the AAA bonds can handle roughly 60% defaults at a 50% severity (21% enhancement plus the 10% from the excess spread).
What this means for AAA ABX prices
The fact that the longest-duration bonds are included in the ABX has two major impacts on the ABX pricing. First, the relatively long duration causes spread-widening to have a greater impact on the price of the ABX than on subprime AAAs in aggregate, since the ABX bonds will simply have to eat uneconomic returns for a longer time. Second, the bonds’ position at the bottom of the AAA pile exposes them to higher potential principal loss, relative to higher classes of bonds, if ultimate default rates hit very high levels.
Accordingly, the bonds in the AAA index are the most sensitive AAAs in the entire stack to spread widening and loss estimates. Both of those factors have clearly caused the AAA indices to trade to seemingly crazy levels. However, given the huge amount of losses the AAAs can handle, it’s likely that it is spread widening more than credit fears driving the current prices.
Disclosure: The author is an employee of bankstocks.com and an investment management firm. His fund often buys and sells securities that are the subject of his articles, both before and after the articles are posted.