Hiding the risk


Mortgages, MBSs, CDOs, CDSs, synthetic CDOs

Mortgages becoming MBSs becoming CDOs
Banks used to keep the mortgage until it was paid. With securitizing this changed. Securitizing, putting mortgages together in packages, is often advantageous compared to dealing in individual mortgages. Risks are spread and there is less paperwork.
Mortgage Backed Securities (MBSs) are packages typically put together from hundreds or thousands of home mortgage loans. Often, the MBSs are amassed and put together in a new sort of MBS, a Collateralized Debt Obligation (CDO); CDOs are even put together in new CDOs (CDOs squared). It is nothing very sophisticated, just shuffle and deal, reshuffle and deal again; still with all this intermixing, such securities are very difficult to analyze. Further, different types of securities are subject to different types of restrictions; a new type can easily fall outside the restrictions of the underlying securities. When it comes to CDOs, regulators like the Fed can't get the underlying documents because the regulators are not qualified investors.
With securitizing, things changed; the amassed mortgages were often sold to others. The banks did not care if they sold junk; they got their bonus, after that they had no interest in how the loans performed. Should the loans fail, well that's too bad but the loss is left to somebody else; to federally sponsored Fannie and Freddy or to investors like pension funds. Fannie and Freddie could keep them, leaving the risk to the taxpayers; they could also sell them to other investors.

A CDO is pretty much like an ordinary MBS but there is a difference, the revenue is split in another way. In a standard MBS the revenue is split proportionally to how much of the security you own; a CDO is split in packages (tranches), rated after risk. As the borrowers pay off on their loans, the money goes first to the top tranche, then to the next tranche and so on as long as there is money. The top trance is paid least because it is paid first and therefore is least risk; it is not necessarily the one with the highest rated securities.
The idea behind CDOs is not new; little risk pays less, high risk pays more. The CDO is more of a risk as you get either the agreed revenue or nothing, instead of a variable sum depending on how well your investment performs. Efficiently the CDO is an extra bundling layer, producing more fees for the mortgage servers and less for the investors. It also increases the risk by making the securities more difficult to analyze and regulate.

CDS for insurance
To keep within the rules that specified how much capital a bank has to have to cover its risks, Credit Default Swaps (CDSs) were introduced. These were financial instruments, basically insurance, freeing capital and transferring the risks of your CDOs for a regular fee; should your CDOs falter, you got money back from the CDS issuer.

CDSs and synthetic CDOs
CDOs got so popular the banks could not get enough of them. So they issued synthetic CDOs, CDOs that were not backed by assets but were linked to reference CDOs.
You could buy CDSs for synthetic CDOs just as you could buy them for ordinary CDOs. For standard CDOs, you are paid from mortgage payments. For synthetic CDOs, you are paid from CDS fees; you can buy CDSs and get the insurance value if the reference CDOs fail. CDSs for synthetic CDOs were popular because you could buy them without owing any CDOs. Buying synthetic CDSs was shorting, it was betting that the reference CDOs would fail, and your CDSs were your betting tickets. It's like anybody could take out insurance on your house; should it burn there will be many to pay for the insurer. Contributing to the risk was that although CDSs function as insurances, formally they are not; this means that those who issue CDSs are not requested to have the same backing in assets as those selling insurances.
With more and more financial institutions understanding that a financial collapse was inevitable, they wanted to profit from the coming crash; trading in synthetic CDOs and CDSs soared. As there was no central registry for synthetic CDOs, it is difficult to estimate the impact. All I have found is a comment to a blog post claiming that the ratio between synthetic and mortgage CDOs was at least 8 to 1; the impact should be of the same magnitude.

© Anders Floderus