Why Citi’s Accounting Trick Makes Even Less Sense Than You Think

No one appears to have pointed out the following yet. (The reasoning strikes me as sound and would destroy any last vestige of credibility for the accounting weirdness that allowed Citigroup to book a $2.5 billion gain because the company edged ever closer to bankruptcy, which drove down the value of its debt.)

First the quick recap (and it’s still not completely clear to me what Citigroup did, but I think I’ve got the basic accounting principle down at least). Citi’s earnings were buoyed by a rule that, according to Bloomberg, allows companies “to record any declines in their debt as an unrealized gain.”

The Business Insider went on to explain the rationale thusly:

In other words, because Citi's debt is trading at distressed levels, the company could, theoretically, book some balance sheet gains by buying it back -- reducing liabilities by more than it costs to retire that debt. Of course, they haven't done that yet. But they could!

But could they? Of course the accounting rule implies this, leading us to a real noggin scratcher of a paradox: run your company into the toilet and watch your earnings soar! Shouldn’t this alone tell us that something here might be mathematically suspect? That maybe these accountants have overlooked something rather, um, important?

Let’s back the truck up and play with some real-world numbers, simplifying everything. Let’s say Winkie’s Widgets sells bonds (debt) that would pay $110 million in 12 months. Let’s say average investors want a 10 percent yearly return on a well-run company of Winkie’s size and characteristics. So they pony up $100 million for the offering.

First day of trading in the bonds: oops, a big bombshell. Winkie’s has been making widgets out of old rebar from nuclear reactor cooling towers; they’re secretly up to their necks in lawsuits; they’re staring at bankruptcy at close range. Immediately the price of the bonds plummets to $55 million. Investors now are saying: "Hey, you’re pretty risky after all, so we want a 100 percent yearly return."

Now the CFO of Winkie’s should be glum, but he’s no slouch. He realizes he's looking at a big chunk of gold: the company can book a $55 million gain on earnings. The reasoning: Winkie’s could go onto the market and buy back their bonds for $55 million, retire the existing $110 million of bonds, and pocket the difference.

Sounds great. But what's missing? The backend of this transaction. Winkie’s has to come up with $55 million. Because the widget maker is judged to be a high bankruptcy risk, it will pay a very steep premium for funds ... basically investors will want a 100 percent return. That’s what the bond buyers want now, remember?

So Winkie’s borrows $55 million at 100 percent, buys up its old bonds and books a profit of exactly ... zero. There is no profit. There’s simply the illusion of profit. This is why I don’t understand how accounting types can defend this sort of bookkeeping. It ignores what's going on at the heart of the bond valuation.

Am I missing something? Dissenters welcome.