J.P. Morgan's ARS Book: A lack of vision

Results of J.P. Morgan's ARS auctions Thursday were fascinating. 50 out of 93 auctions failed, with the "failure" interest rate ranging from 3.41% to 15%. The average among the "failed" bonds was 5.26%. Remember that if an auction fails, there is some mandatory "failure" rate, which varies greatly. My understanding is that a lot of student loan issues have relatively low failure rates, and that's borne out in the J.P. Morgan list. J.P. Morgan shows 12 failed student loan issues, one of which reset at 12%, but the rest reset below 6% despite the failures.

The failures are only half the story, though. Among the "successful" auctions, the average rate was 8.69%, with 14 "successful" issues resetting above 10%. That tells me there are buyers out there who are willing to buy these bonds, and are bidding just below the mandatory failure rate.

There was absolutely no rhyme or reason to what failed vs. succeeded and what rates resulted. The City of New York (rated Aa3/AA) had three issues auctioned, all three of which were fully tax-exempt. The rates were 5%, 5.25%, and 6%. Energy Northwest, a municipal power provider in Washington State, which is rated Aaa/AA-, had their ARS fail, and got a rate of 6.23%. Meanwhile, at one healthcare institution with a Baa3 rating and Radian insurance had their auction succeed with a 4.67% yield.

Several dealers, including UBS and Merrill Lynch are publicly stating they will not necessarily support ARS by buying bonds at auction. As has been discussed, in the past it was common for dealers to step up and buy the ARS to prevent an auction failure. But right now the volumes of failing bonds are just too great, and capital is too constrained. I'm certain the dealers are worried about lawsuits, with investors claiming they were told the bonds could be sold at each auction date without question, which is of course not the case. I don't know how anyone is going to sue when no one has actually lost any money so far, but we'll see.

Meanwhile, I've received many e-mails from people interested in how one can play these ARS failures. One interesting idea is closed-end funds. Most closed-end bond funds are leveraged, usually in the 1.2-1.5x area, and they commonly use auction-rate preferred to create this leverage. Obviously if the ARS reset at 10%, and the fund is buying municipal securities which yield 4%, that's not a good situation. But let's think about closed-end ARS for a second. The closed-end fund has pledged its portfolio of bonds to preferred shareholders. Let's say its a $150 million portfolio of munis, $50 million of which was bought with proceeds from the auction-rate preferred (i.e., 1.5x leverage). But really that means the preferred holders are covered 3 times over: $50 million in debt covered by $150 million in assets. Really that's safer than any single muni issuer.

And yet closed-end muni funds are getting hammered, with several falling more than 5% today. I looked at the worst performers on the day, all of which had leverage created with auction-rate preferreds with 7-day auctions. The highest reset number I found was 3.3%. Excuse me if I don't panic.