Is a 47% haircut big enough? That’s my question of the day. And to be quite frank, my answer is: “I’m not sure.” I’m inclined to say no.
Word’s out that Ireland’s NAMA – National Asset Management Agency bought €16bn in property loans (1,200 actual loans) for €8.5bn. A 47% haircut.
But did the agency buy the loans cheap enough? Because the surest way to lose money – and violate Warren Buffet’s first rule: “never lose money” – is to overpay for something. The answer to that question can only be found in the loans themselves. How were they underwritten? How was the collateral valued? When were the loans underwritten? Because ultimately, it’s the recovery of the cash flows at the time of underwriting that matter in this ordeal.
A paper by Bruche and Gonzalez-Aguado (hat tip: DefaultRisk.com) goes right to the heart of this issue. Recoveries and default rates are negatively correlated. So while we’ve been seeing default rates rise, the recoverability of those loans and bonds suffered. In a perfect world, we’d have some well-grounded expectations in what we could recover from a distressed loan or bond (as well as the timing), which we could then discount via NPV to decide if we want to purchase the asset or not at the time of default.
Problem is, we don’t have a lot to go on. The data Bruche and Gonzalez-Aguado used was supplied by Moody’s, which is used by researchers all over. So while there’s an abundance of models that can be developed, your still limited by the confines posed by the data available. Data has been very scarce in this area and even with well-constructed models and great tools, lack of data makes purchasing distressed loans like taking Potions class at Hogwart’s without a Book of Potions. You’re flying blind and hoping your intution/gut instinct is going to bail you out.
And why the lack of data? Simple: surviorship bias. It’s not like we have a lot of data on banks that have failed through the years, especially since the biggest crises prior to this one pre-dated the computing/processing power/data storage capacity we have now. Hard to do collateral/recovery analysis on a loan-by-loan basis for if the surviving institutions didn’t have a lot of bad loans to begin with and had no appetite for keeping reams of data on assets they purchased from failing shops.
So basically, this is a taxpayer-funded research project. As NAMA purchases the loans, they’ll have to turn around and try to sell off the assets without taking a loss. If they get the same amount they purchased for them, they break even. No blood, no foul for the taxpayer. If there’s a loss, then the taxpayers are essentially falling on the grenade because NAMA is rushing in where angels fear to tread. Yes, I realize they could make a gain, but let’s face it: the probabilities for gain, break-even, and loss are not equitable for the trinomial tree. In all likelihood, they’re skewed towards break-even or loss. But in any event, they’ll need to be ticking and tying data all along the way.
I just hope NAMA is keeping good records of all of this. There’s going to be a lot of interested parties. Moody’s, Fitch, S&P, the BIS, other central banks even. Just think of the research grants that could be written to study this stuff!
Come to think of it, NAMA should go into the data services business. Maybe the taxpayers will get their money back through data subscriptions…