It's hard to find a perfect analogy to explain MBS, but let's try by thinking about pulpy orange juice. If it's all shaken up, then the pulp with be throughout the juice. But if it's left undisturbed for some time, then that pulp will settle at the bottom. In fact, juice makers have methods to create virtually pulp-free versions of the juice. Those who hate pulp can buy this variety, but those who love pulp can buy a much pulpier version that the juice makers create with the left over pulp and juice. Imagine that pulp is risk.
An investor comes along and wants to buy a subprime MBS without much risk. Like the juice maker does with pulp, a bank can structure an MBS where it minimizes the risk for some bonds within that deal, while other bonds are much riskier. So that investor can buy one of the "senior" pieces of the security without taking on much risk -- just like you could buy orange juice without pulp.
Then went so horribly wrong during the housing bubble? Everybody underestimated how much pulp there was in the juice -- or how much risk there was in the subprime loans. Imagine that you thought that the volume of some orange juice was only 30% pulp, but it was really 40% pulp. If you try to drink more than 60% of that juice, then you'll hit some unanticipated pulp. In a similar way, investors suddenly began facing losses that they hadn't anticipated, because the loans were riskier than they thought.
Imagining a AAA-Rated MBS
I am not familiar with the particular mortgage-backed security that the Bloomberg article refers to. But it's easily the case that a subprime MBS could be created with AAA-rated bonds. Imagine that, even in the worst imaginable scenario, the loans it is created from would incur 50% losses. When you think about mortgages, such losses would be very severe.
Now say that you structure the MBS so that the senior bonds are the last to take on losses and that they make up just 35% of deal. The other 65% is made up of subordinate bonds, which would not be rated-AAA. Under this scenario, it is inconceivable that the senior bonds would ever take a loss. As a result, it seems perfectly acceptable to rate them AAA.
Think about what would have to occur for these senior AAA-rated bonds we just dreamed up to take a loss. The pool of loans -- on a whole, not just a few bad ones -- would have to sustain losses exceeding 65%. These loans would have to be beyond awful for that to occur. Remember, even if a mortgage fails and faces foreclosure, the bank will sell the property for something. For the senior bonds to sustain a loss, the bank would need to sell the underlying properties for less than 35% of the value of the mortgages, on average for the entire portfolio of loans.
What About the U.S.?
So it's definitely conceivable that there could be some mortgage bonds that deserve AAA ratings. As for the one that the Bloomberg article refers to, it may or may not deserve the grade. I was unable to find its prospectus when I searched for it. But even if I could evaluate it, my conclusion might differ from somebody else's assessment who has different assumptions about the future. For that reason, the risk that these bonds pose should be determined by individual investors who intend to purchase them. Remember, S&P's rating is merely an opinion.