A while back I built a spreadsheet to do math about AIG, and it took me a long time and led to basically one short post with what I still think was a rather lovely blobby picture, so I'm just going to shamelessly reuse that spreadsheet with slight updates and be all OOH LOOK AN IRR:
So yeah: as the AIG bailout saga comes to its sort-of conclusion, we can sort of conclude that the government made a 5.6% return on its money. Assumptions etc. in the original post; the accounting profit ties out reasonably well, if you squint, with the Treasury's official math.
Herewith some random observations and questions on AIG:1
1. Is that IRR good, bad, indifferent? The main answer to that is "shut up, this was done to save the world, not make a profit, so you should be happy that it ended up doing both, even though the profit was sort of uninspiring, math-wise." This seems about right.
One thing worth noting is that a lot of save-the-world-and-profit programs posted higher returns, but with some numerical sleight of hand. So for instance the PPIP has done very well - here's BlackRock's press release touting the 23.5% IRR it provided on Treasury's equity investment - but that's because in addition to its equity, the Treasury put in twice as much debt, at a considerably lower return. One could say similar things about the bank bailouts, where high TARP IRRs require a grain of salt because they ignore big slabs of secured lending, etc.
AIG's bailout, on the other hand, ended up wrapping pretty much all the money into one blob, although to be fair that blob then spent much of the last four years migrating down the capital structure. But while 5.6% returns for "four-year equity investment in AIG" sounds just horrible, 5.6% for "mostly just senior secured funding of a pool of what turned out to be mostly salvageable debt securities" sounds less horrible. The truth is somewhere in between.
2. If you are all "Neil Barofsky should run the SEC," what do you make of his insistence on comically wrong accounting for the AIG bailouts? I find it encouraging; I want an SEC chairman who can appreciate creative accounting, even if his sort of goes the wrong way.
3. This is really just my problem and I'm seeking help for it, but my favorite part of today's AIG stuff is this, from the prosupp:
Regardless of the outcome of this offering, Treasury will continue to hold the Series D Warrant (as defined below) to purchase up to 2,689,938 shares of our common stock and the Series F Warrant (as defined below) to purchase up to 150 shares of our common stock, each pursuant to the current terms of the Series D Warrant and Series F Warrant. ...
On November 25, 2008, we issued to Treasury a warrant (the “Series D Warrant”) to purchase up to 2,689,938 shares of our common stock .... The Series D Warrant is currently exercisable at an exercise price of $50 per share. ... The Series D Warrant has a term of 10 years and is exercisable at any time, in whole or in part.
On April 17, 2009, we issued to Treasury a warrant (the “Series F Warrant”) to purchase up to 150 shares of our common stock .... The Series F Warrant is currently exercisable at an exercise price of $0.00002 per share. ... The Series F Warrant has a term of 10 years and is exercisable at any time, in whole or in part.
You can sort of piece together from the Treasury press release today, and the documents initially announcing these warrants, that Treasury will probably sell them to AIG at fair value as soon as the "yay you're not a ward of the state" hoopla dies down. Lazy Bloomberging suggests that the Series Ds were worth about $5.37 per share, or about $14.4 million, at the $32.50 offering price this morning, or about $6.70 ($18 million) at this afternoon's $35.26 close.2 So the government made a cool $3.6 million on those warrants today, which is like, really really really really tiny in the scheme of this bailout but also kind of a lot of money, especially to me, these days.
You can see why they're holding on to those Series D warrants: they're out of the money, they have a lot of value, AIG will buy them back, and it does make sense to get a bit out of the news before negotiating that buyback. But the government also has those near-zero-strike Series F warrants. One hundred and fifty of them! With a value, according to Bloomberg, of roughly 26 hundredths of a cent (per share) more than the underlying stock. For a total time value of 39 cents. The government could have exercised those warrants at any time - yesterday, for instance - and paid, oh, $0.00002 times 150 is ... three-tenths of one cent ... and gotten 150 shares and sold them in this offering at $32.50 per share, leaving them with one less bit of AIG to remember to do something about. They didn't.
Why? I have no idea, but a silly guess is that it's non-optimal to exercise an option before maturity, Treasury has the right to get fair value for those warrants from AIG, and they are damn well going to get that fair value. Given the politics involved, Treasury doesn't want to leave money on the table in its dealings with AIG. Even if the money being kept off the table here is 39.3 cents.
1.Ooh here's one in a smaller font: No the Treasury does not owe capital gains taxes on its gains. (!?) One reason that some people sometimes give for capital gains tax rates that are lower than ordinary income rates is that long-term capital gains represent inflation as well as real return. I invite you to ponder that argument in connection with the government's AIG returns. Plugging in a 1.36% CPI inflation rate since September 2008 for the government's cost of capital, I get a $16.6bn real profit for the government from AIG. This statistic is, I suspect, entirely unhelpful, but there it is, in its small font.
2.You may have to take my word on some of this since OV runs over one screen: