Credit Risk Quantification Project for IMSM 2011
Jared Bogacki
ERM Analytics
BB&T
Winston-Salem, NC
Jeff Scroggs
Department of Mathematics
N.C. State University
Raleigh, NC
Project Description
The challenge for the Credit Risk Quantification group is to develop systematic approaches to the identification of a given counterparty's NAICS code, and to utilize the NAICS code (and other factors) to identify an appropriate risk rating in the context of commercial banking.
Risk from credit is the deviation of the performance of a portfolio of loans from its expected value. Credit risk is diversifiable, but it is difficult to eliminate completely. This is because portions of default risk result from exposure to systematic risks (market risks). In addition, the idiosyncratic nature of some portion of these losses remains a problem for creditors in spite of the beneficial effect of diversification. This is particularly true for banks that lend in local markets. Credit risk arises due to uncertainty in a counterparty's ability to meet its obligations in accordance with agreed upon terms. Banks are required to maintain capital that will cover an amount predicted by a Value-at-Risk calculation. Quantification of credit risk helps banks manage diversification, and also helps in the development of adequate controls over risk. Analytical techniques, such as those built into automated credit scoring, are designed to assign a risk rating to each debtor. Clustering companies into groups, may facilitate the assignment of a risk rating. An example of credit ratings for groups of companies appears in Table 1 of [Santomero]. A suggested approach to grouping companies uses NAICS codes, where the NAICS code may be identified using Latent Semantic Indexing (LSI).
Participants are encouraged to contact Dr. Scroggs ([email protected]) with any questions they might have.
References
Technical
1. Latent Semantic Indexing (LSI) (know the definition and general ideas)
A. Text to Matrix Generator (TMG) (this code will be used)
2. Santomero (bank risk management)
3. Engelmann & Rauhmeier (risk modeling and default probabilities)
4. Rasmussen & Clausen (mortgage loan portfolio optimization, optional)
5. Dietsch & Petey (credit risk model, optional)
Non-technical & Data
6.Office of Federal Housing Enterprise Oversight (OFHEO)
A. House Price Index
B. S&P Case-Shiller Housing Price Index (Wikipedia link)
7.North American Industry Classification System (NAICS)
A. US Census Bureau NAICS information
B. Wikipedia description of NAICS
C. 2007 definitions and index file
8. Small Business Administration Loan Failure Rates (by Industry)
9. Real Estate Investment Rating
Terms (know these)
10. Value at Risk (VaR)
11. Conditional Value at Risk (CVaR)
12. Loss Given Default (LGD)
13. Coherent Risk Measure
14. Commercial Bank (compare with investment bank)
15. Copula