3/31/2008: Dick Bove on Wachovia and ABX.HE

Excerpt from Being Watchful of Wachovia
By Dick Bove

The indexes that are used as comparables for valuing securities and loans on financial company balance sheets are plunging once again. Some have declined by 22% in March alone. This suggests that banks, in general, and Wachovia, in specific, are going to have lower earnings than we thought just two weeks ago.

One must stop and think as to whether this adjustment makes sense. Let's take a few of the ABX indices for example. The ABX-HE-AAA 07-2 (home equity loans) is now valued at 51.93. It is just about six months old, so it is down 48.1% from Sept. 1, 2007. The ABX-HE-AA 07-2 index is now 21.32. This index has fallen 78.68%.

What else are the indices indicating? On Sept. 1, the loans in the securities of the ABX-HE-AAA 07-2 were being valued at 95%. Presumably, this meant that the index reflected a potential loss of 5% on these loans. Since Sept. 1, interest rates have declined. In theory, then, the value of the index should have risen. Any decline in the value of index given this event would reflect expected losses on the underlying loans.

What is that assumption? Well, for the AAA index it is now expected that approximately 48% of the loans will go bad. For the AA index, the assumption is that about 79% of the loans will go bad. Is this plausible?

Each quarter, the Federal Deposit Insurance Corporation publishes the actual loan quality experience of the nation's 8,223 banks and thrifts. In the fourth quarter of 2007 (ancient history by today's standards) the percent of home equity loans in these banks' portfolios that were 30 to 89 days past due was 1.14%. The percent of the loans that were noncurrent was 0.86%. The actual charge-offs in this quarter for banks over $1 billion in size (only data available) was 0.88%. The annualized rate would be 3.52%.

Make no mistake; this is a very bad number. It is the highest since the data were first collected in 1991. It is up more than fourfold from the year-earlier results and it is likely to get worse. However, this number would have to jump 12-fold to reach the assumed write-off in the AAA index and it would have to jump more than 22-fold to reach the assumption in the AA index.
Does this make sense? (No.) Has it ever happened before? (No.)