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
Portfolio optimisation has largely been a failure in retail lending. The failure is not in the optimisation, but in the input to optimisation and failing to clearly specify what is being optimised.
Portfolio managers put a great deal of effort into account management, a significant effort into forecasting and are starting to take stress testing seriously. However, at the most strategic level of portfolio optimisation the most primitive models tend to be used. Many optimisations are even run with simple extrapolations of previous years’ portfolio totals. Optimisation is most critical in times of rapid change and yet that is when performance extrapolations fail most dramatically.
Using the analysis of the previous chapters, this need not be the case. This chapter will focus on putting the right input into some existing optimisation approaches to get useful answers. Of course, we also need to review what those results really mean for retail portfolios.
Before any work begins, we must define exactly what the goals are and how success will be measured. For example, are we trying to optimise for next year or any year? These lead to fundamentally different answers. Neither is entirely