I didn't actually post my first entry of the day at 10:20 this morning, honest: Movable Type seems to have eaten my real first post, about the Abu Dhabi investment in Citigroup. My blog entry compared the 11% coupon on that deal to the 7.5% coupon that Bank of America is getting on its Countrywide investment. It was cheap and snarky, but hey, what do you expect first thing in the morning.

In any case, I didn't really need to go there: everybody else, it turns out, had the same idea. Most journalists, however, alighted on 9% junk-bond yields as the proper comparison.

Alphaville has a blog entry, headlined "Junk Citi", which quotes the WSJ:

Citi is paying a higher interest rate than companies that borrow on the high-yield, or junk-bond, market; currently they pay roughly 9% for straight bonds. Typically, convertible bonds pay lower interest rates than straight bonds, although a particular bond’s structure could affect the interest rate paid.

Meanwhile, David Wighton in the FT talks about "the high cost of the new funds", and Dana Cimilluca says it's "stunning" that the coupon "is nearly 2 percentage points more than the average U.S. junk bond yield". Oh, and in case you hadn't got the message yet, a Bloomberg headline this morning said "Citigroup Pays Junk Rate to Keep Dividend After Mortgage Losses", and although the word "junk" has now been dropped from the headline, it's still there in the first sentence, and there's even some attempt at analysis:

The 11 percent interest rate on $7.5 billion of convertible shares that Citigroup sold to the Abu Dhabi Investment Authority is almost double the rate it offers bond investors. Countrywide Financial Corp. paid 7.25 percent to Bank of America Corp., the second-biggest U.S. bank by assets, for bailout financing three months ago. Citigroup's common stock pays a dividend equivalent to a 7.1 percent yield.

So give me nul points for originality: maybe Movable Type has some kind of bullshit filter I should be thankful for. Because if you actually bother to look a little deeper into the deal, it starts making rather more sense.

Andrew Clavell has done a bunch of legwork on this one and come to a less newsworthy conclusion:

Citi has raised tax deductible, upper tier capital funds for 4 years at a cost equivalent to another financing source of Libor+150. Smart business.

Clavell's post can be hard to follow, however, for people who aren't comfortable with the mechanics of callspreads. (Er, me.) So let me try to explain in English what's going on.

Abu Dhabi is buying a $7.5 billion stake in Citigroup. But it's not buying the stake at the current Citi share price. For the time being, and indeed until March 2010, Abu Dhabi will own no Citi equity at all as part of this deal. Instead, it gets debt instruments paying that 11% coupon. Then, from March 2010 until September 2011, those debt instruments automagically become Citigroup shares in a process known as a "mandatory convert". But here's the rub: the higher the Citi share price, the fewer shares that Abu Dhabi will end up with.

Abu Dhabi will receive a maximum of 235 million shares, if the share price at the time of conversion is less than $31.83. It will receive a minimum of 201 million shares, if the share price at the time of conversion is greater than $37.24. But as a result, Abu Dhabi essentially loses out on a very large part of that 17% share price appreciation from $31.83 up to $37.24.

Now remember too that if Abu Dhabi had bought 235 million shares outright at $31.83, it would be receiving a dividend yield on those shares of 7.4%. So the coupon on the bonds has to be at least 7.4% to make up for lost dividends. Then, on top of that, there needs to be a bit of extra coupon to make up for the fact that Abu Dhabi is not going to participate in most of the first 17% of any price appreciation in the stock.

All this can be modelled using puts and calls, which is what Clavell has done, and he's come to the conclusion that the value to Citigroup of all those implicit puts and calls means that the 11% coupon is really rather reasonable. I'm not nearly sophisticated enough to double-check his math, but it seems like the man knows what he's talking about.

Of course, Abu Dhabi is guaranteed its 11% coupon for the next few years, while Citi's shareholders are by no means guaranteed their dividend. On the other hand, any cut in the dividend might conceivably result in a rise in Citi's share price, if shareholders are convinced it would put the bank on a much more sustainable footing going forwards. No one really knows. What does seem clear, however, is that a simplistic comparison of the 11% coupon to junk-bond yields of 9% is not really kosher.