MBS were thousands of loans cobbled together and sold as bonds, collateralized debt obligations ( CDO ) are bundles of MBS, or bundles of tranches of MBS.
A CDO might sound like a loan secured by some type of collateral, it's nothing of the kind. When the housing market could no longer keep up with Wall Street's demand for MBS to sell, the investment banks began to combine tranches from multiple MBS, usually the most risky, and sell them as bonds. The marketing angle was that while each piece of the CDO was high risk, together they would be less precarious as it would be considerably less likely that they would all default. What this did for the banks, beyond giving them new product, was to enable them to take the lower-rated tranches of MBS and raise their rating. This made them easier to sell ( or so the argument goes, more on this later ) as most bond buyers avoid junk. These CDO followed they same tranching and redemption regime as the underlying MBS.
So mortgage lenders made loans; sold them to Wall Street; the banks bundled them into bonds and then when demand outstripped supply, they repacked and repackaged and repackaged. Inevitably, some market speculators began to understand that these mortgages upon which the new business model was based were not likely to be repaid, and they began to look for ways to short the housing market. Fortunately for them, our friends at Goldman Sachs had invented something called a credit default swap. A CDS is an insurance policy on an investment. Even more fortuitously for prospective shorters, JPMC went one step further and invented the naked credit default swaps. This enabled speculators to buy insurance policies on investments in which they played no part. With a NCDS, one didn't have to own the CDO to buy a policy against it. A NCDS is simply a wager on whether a bond will be redeemable.
Like any insurance policy, one has to pay a premium. For speculators shorting CDO, timing is especially important. Say you have a ten million dollar policy against a CDO default, and that your premiums are a million per year, if the CDO busts a year after you bought the policy, then you made nine million: you paid one year's premium ( a million dollars ) and got ten million from your policy. If the collapse happens two years after you took out the insurance policy, then you make only eight million, and so on.
So why were the investment banks so keen to get repackage tranches of MBS to get better ratings? There are bond buyers who prefer non-AAA rated bonds due to their higher yield, it should make no difference to the bank as they make as much from the sale of BBB as AAA bonds. What I can allege and what I can prove are different things, but here are some factors to consider:
While riskier bonds have better yields and may be a better investment, if one is considering shorting them then lower-rated bonds present the thorny problem of higher premiums. Obviously, it costs more to insure high-risk bonds. Using the example above, if the ten million dollar policy has yearly premiums of 1.5 or two million dollars, the NCDS is considerably less profitable.
So why would the banks care about this as they only sell the CDS and are not themselves at risk? The answer is proprietary trading. Banks not only buy and sell financial instruments on behalf of their clients, they can, within certain ( woefully insufficient ) limits, engage in investing themselves. Indeed, the larger banks and the hedge funds they own are the largest investors of all. JPMC and Goldman Sachs, the largest brokers of MBS and CDO, were also the largest buyers of CDS. While they were selling CDO and MBS to their clients who would later sue them, they were buying insurance policies against those same investments. Yes, that's right, they were creating financial vehicles, telling their customers that they were worthwhile investments, and then placing bets with their own money that these bonds would go belly-up.
If you are starting to think, dear reader, that the investment banks pimped these CDO to their unsuspecting customers, then bought CDS at just the right time, and got out of the CDO business just before the crash occurred so that they had no liability on their books, well, guess what? You are right. That's precisely what did happen.
One of the greatest myths of the collapse was that JPMC and GS were in danger of collapse. Their only problem was that the insurers who had written CDS for them would not be able to pay. Fortunately for them Hank Paulson did his infamous face-boite and decided not to buy up toxic assets, which is what Congress allocated the requested bail-out money for, but instead to use the embezzled funds to pay off GS's CDS on behalf of AIG and other insurers. How good for them.
One last thing to ponder: It was suggested that the bail-out money be used to pay off the mortgages that were in default, and make other arrangements for defaulters to pay the government. After all the argument was that the credit crunch which brought the economy to a standstill was due to the lack of money coming from mortgage borrowers. The government lending money directly to home-owners would resolve this problem. Again, what I can say and what I can prove are different, but if this had been done the banks would not have been able to keep the bankrupted properties, and the insurers who issued CDS would not have had to pay on these policies. These would be great outcomes for the public, but less lucrative for those who held CDS policies.
