Securitization — the bundling of bank loans to create tradeable bonds — started in the mortgage industry in the 1970s, when Government Sponsored Enterprises (GSEs) began to pool relatively safe, conventional, "conforming" or "prime" mortgages, create "mortgage-backed securities" (MBS) from the pool, sell them to investors, guaranteeing these securities/bonds against default on the underlying mortgages. This "originate-to-distribute" model had advantages over the old "originate-to-hold" model, where a bank originated a loan to the borrower/homeowner and retained the credit (default) risk. Securitization removed the loans from a bank's books, enabling the bank to remain in compliance with capital requirement laws. More loans could be made with proceeds of the MBS sale. The liquidity of a national and even international mortgage market allowed capital to flow where mortgages were in demand and funding short. However, securitization created a moral hazard — the bank/institution making the loan no longer had to worry if the mortgage was paid off — giving them incentive to process mortgage transactions but not to ensure their credit quality. Bankers were no longer around to work out borrower problems and minimize defaults during the course of the mortgage.
With the high down payments and credit scores of the conforming mortgages used by GSE, this danger was minimal. Investment banks however, wanted to enter the market and avoid competing with the GSEs. They did so by developing mortgage-backed securities in the riskier non-conforming subprime and Alt-A market. Unlike the GSEs the issuers generally did not guarantee the securities against default of the underlying mortgages.
What these "private label" or "non-agency" originators did do was to use "structured finance" to create securities. Structuring involved "slicing" the pooled mortgages into "tranches", each having a different priority in the stream of monthly or quarterly principal and interest stream. Tranches were compared to "buckets" catching the "water" of principle and interest. More senior buckets didn't share water with those below until they were filled to the brim and overflowing. This gave the top buckets/tranches considerable creditworthiness (in theory) that would earn the highest "triple A" credit ratings, making them salable to money market and pension funds that would not otherwise deal with subprime mortgage securities.
To use up the MBS tranches lower in payback priority that could not be rated triple-A and that conservative fixed income market would not buy, investment banks developed another security—known as the collateralized debt obligation (CDO). Although the CDO market was smaller, it was crucial because unless buyers were found for the non-triple-A or "mezzanine" tranches, it would not be profitable to make a mortgage-backed security in the first place. These CDOs pooled the leftover BBB, A-, etc. rated tranches, and produced new tranches — 70% to 80% of which were rated triple A by rating agencies. The 20-30% remaining mezzanine tranches were sometimes bought up by other CDOs, to make so-called "CDO squared" securities which also produced tranches rated mostly triple A.
This process was later disparaged as "ratings laundering" or a way of transforming "dross into gold" by some business journalists, but was justified at the time by the belief that home prices would always rise. The model used by underwriters, rating agencies and investors to estimate the probability of mortgage default, was based on the past history of credit default swaps, which unfortunately went back "less than a decade, a period when house prices soared".
Image By TechCrunch50-2008 (Flickr) [CC-BY-2.0 (http://creativecommons.org/licenses/by/2.0)], via Wikimedia Commons