|
Activity Number:
|
365
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2003 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Business & Economics Statistics Section
|
| Abstract - #301589 |
|
Title:
|
Population Shift and Imminence Adjusted Assessment of Portfolio Risk
|
|
Author(s):
|
Timothy Hyungchul Lee*+ and Sung-Chang Jung
|
|
Companies:
|
Equifax Decision Solutions and Chonnam National University
|
|
Address:
|
595 Croydon Ln., Alpharetta, GA, 30022-8026,
|
|
Keywords:
|
classification ; population shift ; time series
|
|
Abstract:
|
A popular analytic tool for performance forecasting such as a scoring algorithm using a two-group classification method produces discriminants or scores. This implies likelihood of event (e.g., default in payment or response to solicitation) based on the current profiles that are being assessed. Some users expect that the scores would change such that they reflect stronger signs of events as the event approaches. It is also expected that the difference between event and non-event observed on key variables would become greater as it approaches the event. In this study, (i) we observed how bankrupt and nonbankrupt groups are separated on each key variable for each observed time point before and after the filing, (ii) investigated the relationship between the time to the event and the change of scores; (iii) tried to incorporate the imminence of event with scores for portfolio decision makings, and (iv) explored and discussed the ways of using short-term time series information for classification. For the analysis, time series data are simulated from observed data for six months before and six months after bankruptcy filing for small and medium firms.
|
