|
Activity Number:
|
238
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Business and Economics Statistics Section
|
| Abstract - #302513 |
|
Title:
|
Anticipating Extreme Changes in the Performance of Long Repayment Term Consumer Loan Portfolios Through Eigenvector Analysis of Markov Chain Matrices
|
|
Author(s):
|
Alex Strounine*+
|
|
Companies:
|
Strounine Thompson Technology Group
|
|
Address:
|
PO BOX 990934, Boston, MA, 02199-0934,
|
|
Keywords:
|
consumer credit ; Markov chains ; recession ; credit risk ; risk management ; forecasting
|
|
Abstract:
|
Predicting credit performance of consumer loans that have long (12 to 20 years) repayment terms can be a challenging task. When loans' past prepayment and default rates are taken as best approximations of the future outcomes, the forecast might be outdated regardless of modeling techniques applied. Competitive activity and changes in the economic environment are common factors that undermine the accuracy of such predictions. We present a technique that uses short term models and Markov chain analysis to monitor inconspicuous movements in pools of new loans to detect trends that signify qualitative changes which will result in extreme deviation of actual loan performance form the predicted one. Whether we expect a recession, an economic crisis, or a revolution, the method provides tools to detect an extreme change sooner and to define strategies to mitigate potential credit loss.
|
