Fannie Mae's Problem: Is There A Mathematician In The House?

I’m still trying to decide what would be more fun: Repeating my morning of trying to understand the accounting changes involved in Fannie Mae’s most recent 10Q or sticking pins in my eyes. I’m starting to think the latter.

You see, on Tuesday, an article in Fortune Magazine seemed to open up old wounds for Fannie. The company is still recovering from a massive $6.3 billion accounting scandal in 2004 that resulted in government sanctions and a real hit to the Fannie reputation. The new article claims that Fannie fudged its credit loss ratio in order to make things seem a little better than they are in mortgage land. Now let’s start from the beginning.

The “credit loss ratio” is how much money Fannie has lost from bad loans as a percentage of all its loans. I guess you would use this to gauge how well the company’s portfolio is doing. Now in the most recent quarter (Q3), Fannie bean counters stripped out current charges from the ratio, making it look a whole lot better than it would if those charges were included.

The conference call today by Fannie may as well have been in "Chinese," so we called Fannie back for clarification, and they explained it pretty well. First, they said the Fortune article was inaccurate, and they were in fact trying to be more transparent, not the other way around. Secondly, they break it down like this: When loans reach a certain stage of delinquency, Fannie is allowed to buy them back and attempt to modify them so that they will get out of delinquency.

The trouble is, there are accounting rules that require that Fannie report the loans--which are now worth less than when they originally purchased them--as charges when they take them back. They reported $670 million in these charges in the third quarter. But did they under report?

So here’s Fannie’s argument: They’re going to modify these loans, and in turn, the loans will “cure." “Cure” means the borrower makes restitution for loan arrearages by repaying missing installments, refinancing the loan or paying off the mortgage by selling the property. So if the bulk of these loans "cure," they won’t end up as charges, and therefore they don’t want them in their credit loss ratio--that’s only for the loans that are never repaid.

That makes sense. The only trouble is that Fannie refused to tell anyone on the call what their "cure" rate is, so it was hard to know if there was any truth to what they were claiming. How many of these loans really "cure?" Well, Howard Shapiro, of Fox-Pitt, Kelton, says, “what we’ve been able to do from secondary market sources, is to look at cure rates, and what we do find is that generally somewhere between mid 80’s to 90%+ of these delinquent loans do tend to cure, so it does sound like they’re being fair in their assessment.”

Does any of this make sense? It’s their outlook on their ability to modify the loans, and frankly they are doing just what they say. The trouble is that they’re getting more and more loans that they have to modify, and I have to wonder how much that’s going to hurt going forward.

Questions? Comments? [e-mail]