Starting at the Top
Setting Up the Problem
Considering Multiple Vintages
Volatility Analysis and Economic Capital
Credit Scores and Account Management
Analysis of the US Mortgage Crisis
An Example Using [email protected] Data
Examples of Modelling Vintages
Epilogue: it’s about time
The US mortgage crisis serves as a perfect example of where models, systems and judgement can fail in retail lending. In this chapter, we examine the onset and progression of the crisis, incorporating the results of analysing two industry data sets. The first was a conforming-loan data set with performance data from 1999 to 2006 and vintages from 1984 to 2006. Conforming loans are those that meet the standards for purchase by the government sponsored enterprises (GSEs), namely the Federal National Mortgage Association (FNMA or Fannie Mae) and the Federal Home Loan Mortgage Corporation (FHLMC or Freddie Mac). The results of this analysis were previously published in Breeden et al (2008).
The second data set contains approximately 90% of the non-agency loans securitised in the US. The performance data runs from 2001 to 2009 with vintages from 1999 to 2007. Non-agency securitisations include most of the subprime, Alt-A, jumbo, low documentation and alternate products that have been targeted as possible causes of the crisis. These results have not been previously published.
Both data sets were analysed with the DtD approach described in Chapter 3. This serves an example for applying