A friend pointed out this excellent summary of the awesome size of three deals that AIG/the Fed paid out billions on.
Just like nearly 20 years ago when Howie Rubin took the fall for Merrill Lynch Trust 13 as a “rogue trader” who put his trade ticket in his desk drawer, there is simply no way even the largest institutions in the world take on single liabilities of a billion dollars or more without everyone up the line knowing about it, and quite a few of them signing their names to approve the exposure.
A little history is in order.
Merrill Lynch Trust 13 (pretty unlucky) was an IO/PO (Interest Only / Principal Only) deal with GNMA 11’s as its underlying collateral.
It was executed in the CMO market because the direct issuance of Agency IO/PO deals had not begun at that point, and some investors in the marketplace felt they could hedge or speculate on the prepayment behavior of the underlying FHA/VA mortgages by buying just the interest payments or just the principal payments from a normal pass-through MBS.
On the bearish side of the trade, the IO piece, was Ernie Fleischer’s somewhat famous Kansas-based Franklin Savings. Franklin is worthy of its own story later, but suffice to say that Ernie was one very smart guy who took a two-branch S&L and turned it into a hedge fund with $10 billion in assets. He used access to Federal Home Loan Bank borrowing and Wall Street’s structured finance machine (CMO’s and swaps) to build his portfolio without going through the slow and expensive process of opening lots of branches and getting savers to bring in their money to buy CD’s.
On the bullish side of the trade, the PO piece, was Merrill.
It must be obvious, but still worth saying, that the IO and the PO together were worth the value of GNMA 11’s, which traded for around 107 at the time. Throw in some legal, printing, accounting and hedging costs, plus some profits, and the billion dollar deal cost about $1.075 billion to put together.
As you might guess from the high premium price, 11.5% (the rate paid by those GNMA 11 borrowers) was quite a lot higher than the going rate for new mortgages at the time the deal was priced. It was natural to believe that the mortgages were going to pay off quickly.
If memory serves, the rumored price for the IO piece was around 38 cents on the dollar, leaving Merrill holding the PO piece for 69.5 cents. Naturally, they were out marketing it around 70 cents.
Salomon Brothers (still remembering that Howie got his start toward stardom there) correctly believed that 70 cents was too high for the PO. They forced the issue by pricing a GNMA 11 IO/PO deal with a cheaper PO. Again turning to my memory of the rumored price at the time, I recall hearing that “the Brothers” went into the market offering their GNMA 11 PO for around 62-64 cents.
Needless to say, Merrill couldn’t sell any of its bonds, and having already agreed to a price on the IO half of the deal, they were stuck.
What happened next was legend, and ludicrous.
Merrill announced that it had a problem. They put out PR that the problem was caused by just one person, and that it might amount to over $200 million.
The rest of the market was dumbfounded. How could you lose 20 points on a PO? After all, a PO has guaranteed principal. If you buy it at a discount and hold it, it is impossible to lose money on it. Nonetheless, Howie left, they even increased the amount supposedly lost on the trade to $275 million, an amount that was plainly impossible to lose.
Howie took a paycheck and left quietly, partly because of the dinner he had reportedly had with Ernie at which they decided to double the deal from $500 million up to $1 billion.
That being said, anyone who was involved in the early days of these “private label” CMO’s knows that there was more documentation and more signatures, by far, than ordinary people see at a residential real estate closing and financing.
Legally, Merrill Lynch Trust 13 was a bond issue for a billion dollars by a company owned by Merrill, after all. It had a printed prospectus from a financial printer. It had at least one Ratings Letter, an accountants’ “comfort” letter, legal opinions, etc. And that private company had a management team that pretty much looked like the senior management team in the bond department at Merrill, with one or two top corporate types, as well.
In other words, everybody and his even more important boss had signed notarized documents to make this deal happen. To pretend that only Howie Rubin knew about it was the height of (ahem) misdirection.
No need to cry for Howie. He got an even better job at Bear Stearns running their CMO trading desk, and since The Bear was run by traders, his paycheck inevitably went up after being at a shop where the Sales Tribe ruled the roost.
Which brings us back to AIGFP and its multi-billion dollar CDS deals. The fiction that nobody on the insurance side of the shop or in the rarefied top executive offices knew what was happening is just plain impossible. That may have as much to do with the million-bucks-per employee retention contracts that AIG offered its Financial Products team as anything.
We should rewind the film a bit and say what the AIG management was told they were insuring, starting with some facts.
Fact One is that mortgage bonds, among all types of bonds, had the best credit history of all. Plenty of (initially) investment grade corporate, municipal or even sovereign bonds had ended up defaulting. But not the MBS. With that kind of record down in the mid-tier credits with BBB and single-A ratings, the AAA’s seemed even that much more solid. To top that off, AIG was typically insuring the credit on super-senior AAA bonds, so they literally felt that they were taking no risk.
From the buyers’ viewpoint, the reason for paying AIG for protection against a vanishingly small event was to save capital charges. The way the bank capital regulations work, if you had a AAA insurer standing behind a AAA bond you held, the capital charge would be at or near zero (varies by country of domicile of the bank, but international rules set the capital at zero).
Some banks even played this game at the ragged edge. Rather than paying coverage from AAA monster AIG, they would buy insurance from tiny outfits that absolutely couldn’t pay off if the worst occurred. I heard that one Swiss bank bought $2 billion in CDS coverage from a hedge fund that had less than $100 million. Who were they kidding?
For the hedge fund, charging maybe 10 basis points per annum ($1 million per year per billion of coverage) for insurance against an event that couldn’t happen, it was “free money.” For the bank, not holding capital against the super-senior AAA CDO gave their investment in that bond an infinite return on capital.
The crux of the negotiation would be how much cash collateral the hedge fund had to deposit with the bank, and what happened to the interest on that collateral while the bank held it. If the bank paid out a CD rate of a few percent to the hedge fund, and only required something like $5 million per billion in coverage, the hedge fund was making 20% + on their capital. Not as nice as the bank’s infinite return, but enough to make the hedge fund investors happy. That is, until the music stopped.
More on when the music stopped, later. But trust me, it was more than a couple of “rogue traders” who knew the dance was happening.