Description

American Express 6.8 % Fixed-to-Floating Rate Subordinate Debentures

This write up is going to be relatively short because it will focus simply on the notes rather than American Express's business as a whole.  I point the reader to the various writeups on American Express common equity over the years done by Danarb, Spence and Neo.  To those, I have very little to add.

American Express is a top rate company that did a bunch of dumb lending in 2006 and 2007 that they are clearly on the path to recover from.  Their total risk book is going down and their credit metrics are stabilizing. 

In the last couple of weeks the stock has approximately doubled and 5 year CDS has come in about 200 basis points.

During this same time, the piece of paper I am writing about, the 2066 AXP fixed-to-floating subordinate notes, have hardly moved in price.  As of Friday, they were  48 bid by 53 offered and I have been able to buy them anywhere from 48 to 53 depending on the day.  The official close was 53.5, which is what I am going to use for the write up.  I think they can be had cheaper, but 53.5 is still a very good price.

I believe these notes trade inefficiently because they are a relatively small issue ($750 million).  They are the only issue of its kind in the American Express structure (they trade a ratings notch below all other AXP debt) and the notes themselves are rather complicated.  In a lot of ways they look like trust preferred securities but as I will describe, in my opinion they have much better protection than the typical trust preferred security and will likely be called by the company in 2016. 

If they are called in 2016, these notes provide approximately an 18 percent yield to maturity. If they are not called, they are paying 12.7 percent for the next 8 years and then about 1.9* libor+ 416 thereafter.  Either way, it is an awfully good return for the risk taken.

Features of the note:

1. In many ways the notes look like trust preferred securities.  American Express has the ability to defer payments on the instruments for up to 10 years (upon closer examination it is really five years but it would be unpleasant in either case).  If it does defer interest, then it cannot pay a dividend on its common shares. Dividend payments if they are suspended continue to accrue with interest and dividends cannot be paid on the common stock.
2. The notes pay 6.8% interest until September 1, 2016. At that point, American Express may either redeem them at face value or continue them until 2066 at three-month libor+ 222.75 bps.
3. The notes have two "mandatory trigger events."  These events require American Express to sell shares on the open market to satisfy interest payments to the notes.  These trigger events are tangible equity falling below 4% or consolidated net income being negative for 2 quarters and tangible common equity falling 10 percent during that time.  If either of these triggers are hit, American Express has to sell stock on the open market to pay the interest on the notes.

The total interest on these notes is $51 million a year, which is rounding error in the scope of American Express.  However, if the triggers are hit, the company is a forced seller of its own stock on the market, which is a relatively embarrassing outcome.  In short, the triggers are such that any time that American Express might be tempted to suspend dividends, it probably won't. 

Moreover, it is highly likely that these notes have totally outlived their usefulness for American Express.  On page S-16 of the prospectus, American Express notes that the notes were "preliminarily classified as common equity" by the Securities Valuation Office of the National Asociation of Commissioners.  As I understand it (and I could be wrong), this allowed American Express to count the notes as Tier 2 equity.  Playing this close to the line on equity levels is an exercise that American Express has almost certainly abandoned given the market's current concern with capital adequacy.

Also, so far as I can tell, the covenants in these notes are different and in their own way more restrictive than the covenants American Express's other notes (which do not appear to have mandatory interest payment clauses).  American Express would be buying itself several degrees of freedom by retiring these notes (which are trading way out of line with the rest of the capital structure).  While such retirement may not be forthcoming as long as American Express is borrowing from the federal government, it would seem to be a clear priority to repurchase of stock both from a value standpoint and from simplifying the restrictions on the business.

I believe that the notes trade the way that they do because they are a small issue and because understanding them is complex.  I also assume they get some discount because they are rated one notch lower than everything else in the American Express capital structure (which by the way is one more reason for American Express to want them gone) and could get cut to junk while every other instrument remains investment grade.  Their current rating is A-, so that is still a distance away.

I expect that as the market rationalizes and people examine these notes, that they will quickly rise to a level more consistent with the rest of the American Express debt structure.  If they do not do so, they should provide for an extremely satisfying return across a longer period of time.

Catalyst

Notes trading in line with other American Express instruments.  Eventual cal of notes by the company in 2016