|
Activity Number:
|
143
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 3, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Risk Analysis
|
| Abstract - #302864 |
|
Title:
|
Model Risk in the Analysis of Personal Credit
|
|
Author(s):
|
Robert Stine*+
|
|
Companies:
|
The Wharton School
|
|
Address:
|
University of Pennsylvania, 444 Huntsman Hall, Philadelphia, PA, 19104-6340,
|
|
Keywords:
|
model selection ; stress testing ; model validation
|
|
Abstract:
|
The finance industry relies on statistical models to estimate the risk of loans granted to businesses and individuals. Various risks are modeled, such as the chance of counter-party default and falling behind in making regular payment. These estimates determine the reserves that the lender must hold, directly impacting profits as well as the cushion against unforeseen charges. Statistical models in this context can err in two broad ways: (1) using an equation with incorrect estimates or the wrong variables and (2) offering an optimistic impression of the accuracy of the fitted model, such as from failing to properly model dependence. Banking regulations (e.g., Basel II) require models to be stress tested in order to gauge the practical accuracy of its estimates, but the results of such tests depend on many discretionary factors and the degree of population drift.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2009 program |
