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
Credit scores were created in order to distinguish “good” accounts from “bad” accounts. In the process, a credit score ranks the accounts in order between “most good” and “least good”. Given a ranking, a credit-management team can decide on a cut-off point along the scores to distinguish between good and bad. The creation of credit scores for ranking accounts is a well-developed field with an extensive literature (Thomas et al 2002; Mays 2004; Anderson 2007; Thomas 2009).
Because credit scores were the first successful models in retail lending analytics, practitioners have historically thought about accounts first and left portfolio dynamics as an afterthought. This book has reversed the usual flow in order to emphasise the importance of time. Lifecycles and macroeconomic impacts are temporal phenomena and by understanding them first for the portfolio, we have a better foundation upon which to revisit credit scoring.
Appearing last does not imply that account-level modelling is less important. We have shown throughout the previous chapters that portfolio dynamics can be understood and harnessed without going to the account level, but strategic portfolio decision-making must