Friday, September 19, 2008

The Great Collapse Part 4: Tranche Warfare

So yes. The 80's brought the computing revolution and in few places was the computer more revolutionary than in the world of finance. The level of analysis that computing unlocked truly revolutionized every aspect of finance. This is why there was a relative explosion of new techniques and financial products in the 80's. But for our purposes here, we're primarily concerned with one: mortgage securitization. (By the by, everything I relate here is from memory of the book Liar's Poker by Michael Lewis. It's good, as is most of his writing.)

As we learned in the last installment, Fannie and Freddie with their quasi-government status were kind of the be-all, end-all of what the suits call "the secondary mortgage market". (The primary market would be somebody getting a mortgage. When the bank sells it off so it can be bundled and sold to someone else, that's the secondary market. Something else that might come up in our wonderful tour of The Great Collapse is the tertiary market. This would be where groups of those bundles of mortgages are put together into their own little investment vehicles. Oh yes, never underestimate how many times you can sell the same product if you're clever about it.)

Well, in the 80's the relatively staid bond-trading business was about to really hit its stride when it applied the new computing power at its finger-tips. See, different bonds have different interest rates. The riskier something is, the higher the rate, as I mentioned previously.

But there's a bit more to it than that. Obviously, if whoever gave you the bond doesn't go out of business, then you just get the higher rate, even if they are very risky. The reason it gets the higher rate is because some of them go out of business and then you get no money. Safer bonds are less likely to do this, so you're more likely to actually collect your lower interest payments through the life of the bond.

Fine and dandy. But there's another wrinkle. Not everybody is interested in risk. Some pretty big players, pension funds and the like, are -- in theory at least -- looking for safe, predictable returns. They are not interested in the higher interest rates if they come with higher risk of no payments as well, which they do.

The problem with mortgages is that you can't tell which ones are the risky ones. Oh sure, you can kind of tell, some people get higher rates because they seem less financially responsible, but from an investor's point of view buying a big pool of, say, 1000 mortgages, the risk is that you can't tell which ones are going to refinance and pay off early. You can't tell which ones are going to go into default. So the solution had traditionally been to have Fannie, with it's huge resources and quasi-governmental nature, take care of it.

The genius idea that these fellows at Solomon had, was that if you organized the mortgages correctly, you could get a much better idea of the risk involved in the pool. And with the rise of computers, tracking the level of detail necessary was suddenly possible.

Basically it works like this. Say you have 1000 mortgages that you're going to sell. You can look at history and know that, say, ultimately 3% of those will go into default, 10% of them will get paid off early, that kind of thing. The problem is that you don't know which will. But what if you slice it up specifically by when they pay-off.

So say you slice into three pieces (tranches, they're called, which is where my very clever title for this post came from), and as mortgages get paid off or go into default, you put them in the 3rd piece until that piece is full of bad mortgages then -- and only then! -- does the second piece see its first bad mortgage and they all go there until that piece is full of bad mortgages and finally they'll start going in the first piece.

What this means is now that first piece is looking pretty good. The only way it gets hit with pre-paid or bad mortgages is if the total rate of them rises above 66.7%, which historically is just about impossible. The second tranche is less safe, but still very safe, since it only gets hit if the mortgages that drop out rise above 33.3%, still a pretty high number. That last tranche, on the other hand, is basically a junk bond: very risky. It's likely to see a significant portion of its constituent loans go bad.

So why is this helpful? Because now you can market those safe pieces to investors who are only interested in very safe investments. Like, say, massive pension funds and the like. It unlocks a whole new class of potential investor and, as we know from our basic economics, an increase in demand means an increase in price. So there's now a premium to be made in selling these things.

What does this mean over all? Well, for one thing, it means that Fannie and Freddie are no longer the only game in town for the secondary mortgage market. It means that there are even more outlets for banks to get mortgages off their books, which means an even more expanded ability to write mortgages.

And it means that investors who are looking for relatively safe investments are now also players in the mortgage market.

But this, I'm sure you're saying, doesn't sound like a disaster at all! It sounds like unmitigated goodness: the kind of sweetness and light that we all wish peppered our days eternally.

Perhaps, but just wait until we put it together with the housing bubble and give everybody involved some serious incentives to keep pumping the bubble up...

No comments: