They're made on all kinds of loans. So to say they are "5 times as much as the mortgages they're made on" is inaccurate.
Here's a piece from Gillian Tett of the Financial Times (who the BBC has been using a lot recently to explain what the hell is going on), from back in May, giving the history about how these things got started:
...
This was where the story of men such as Reoch began. In the mid-1990s, soon after he joined JPMorgan, about 80 other bankers who worked in the company's derivatives business were summoned to a plush hotel in Boca Raton, Florida, for a brainstorming session. The man who ran that team was a British banker called Peter Hancock. Hancock had taken over the derivatives team in the late 1980s, and seen the business become commoditised. For a few days, the young JPMorgan bankers brainstormed ways of overcoming this. Eventually, the group alighted on a potentially fertile new frontier for derivatives: credit. Until that point, banks that made large loans didn't have a way of protecting themselves against the chance of a borrower defaulting. Similarly, investors who held bonds did not have any mechanism to insure against an issuer refusing to pay out. What would happen, the JPMorgan bankers asked, if somebody created a contract that mimicked that credit risk, and then sold that risk to another investor, for a fee?
...
Most important of all, JPMorgan's top management had a compelling reason to innovate. At the time, the bank had so many loans on its books that it was finding it expensive to keep doing business: it needed large "rainy day" reserves to protect against the chance of the loans turning sour. Hancock's team believed that if they found a way to sell this "default risk" to somebody else by repackaging the loans into derivatives, then they could persuade the regulators that they did not need to post such big reserves. "They say necessity is the mother of invention," recalls Andrew Feldstein, a former lawyer who worked with Hancock. "In this case, JPMorgan had a good reason to look at how it handled credit."
In the late 1990s, men such as Feldstein were trying to develop financial techniques that would turn loans into derivatives they could sell on. They started off doing this on an ad-hoc basis but soon discovered that if they created bundles of derivatives contracts linked to loans, then it was easier to sell these instruments to investors - in the same way that it is easier for banks to sell an investor a stake in a mutual fund than shares in an individual company.
...
As the new century dawned, the teams that had traditionally handled "subprime mortgage" finance - or loans extended to borrowers with a poor credit history - started talking to the derivatives groups. "One of the crucial points happened in late 1997 - around then credit derivatives structurers started to meet with securitisation structurers," recalls Reoch, who by this time was working at the Bank of America. "The bingo moment was in the coffee queue of our Chicago office when the two groups met by chance and realised they needed to talk to each other."
http://us.ft.com/ftgateway/superpage.ft?news_id=fto053020081431572421&page=2There's more there than can be explained in 4 paragraphs, of course. Tett has a PhD in Social Anthropology from Cambridge, so she's pretty good at explaining the way the groups of bankers were thinking. She knows what she's talking about - many have pointed out she was one of the first in mainstream commenting to point out there was a problem. He she is, on August 23rd 2007:
One is the fact that nobody quite knows exactly where the subprime losses truly lie, since these credits have been sliced into millions – if not billions – of securities and scattered between all these modern investment vehicles. Hence the Bank's baffling diagram.
But the second problem is that nobody knows the real value of these instruments either. For many have never been traded, but simply stuffed into these vehicles and left there, seemingly unnoticed – until now, when investors are panicking about potential losses.
Common sense would suggest the best way to deal with these two problems would be to take two steps: namely inject more transparency into the system, by encouraging institutions to reveal their exposures – and then encourage financial institutions to create a proper market to trade the assets, and thus determine a price.
...
So, for my money, it is worth watching closely for any hint that this restructuring effort is getting under way on a large scale. And if it does – and it remains an "if" – that will be a thoroughly good thing, not just for the banks themselves, but for financial confidence at large.
http://us.ft.com/ftgateway/superpage.ft?news_id=fto082320071223400214&page=2 But they didn't try to restructure the vehicles, and it all came crashing down around everyone's ears.