TRUE OR FALSE QUESTIONS

1. Suppose BankWest uses a back test to test the accuracy of their VaR(99%, 1-day) model. Using 250 days of data, they observe 3 failures where the losses exceed the VaR estimates. Based on Kupiec’s test statistic this means BankWest should reject their VaR model.

2. In a bivariate normal model, market risk event are usually assumed to be independently distributed.

3. If losses consistently exceed the VaR and such events are clustered, this suggests that banks have difficulties in forecasting conditional changes in the volatility of profits and losses.

4. Using the first-order approximations, the variance of a bond is simply D2 σ2dY

5. In back testing, the frequency of violations of the conditional VaR model will each time be higher than that of the unconditional VaR model.

6. At a 95% confidence level, there is no statistical difference between 250-day historical VaR figures of -$45,000 and -$46,000.

7. Historical credit data provided by rating agencies can be used to predict the probability of default (PD) of the issuer. Such predictions are most accurate for issues with a triple-A credit rating.

8. Monte Carlo VaR can be regarded as a more forward looking estimator of VaR than either percentile VaR or normal-VaR.

9. The goal of credit risk management is to maximise a bank's risk-adjusted rate of return by maintaining credit risk exposure within acceptable parameters.

10. Academic evidence shows that modern banks tend to systematically overstate their VaR numbers due to incorrectly measuring risk.

11. Portfolio VaR is more than the sum of its parts.

12. Fraud is not part of operational risk.

13. “Spreadsheet errors” (from decimal points to formula entries) relate to operational risk.

MULTIPLE CHOICE QUESTIONS

14. Which of the following is NOT a major factor in the credit rating process conducted by agencies such as Moody’s Investors’ Service and Standard and Poor’s? a) Sovereign risk of the country of the issuer. b) Industry risk associated with the type of issuer. c) Financial capacity of the issuer. d) Return attached to the particular issue.

15. If the loans in a bank’s portfolio are all negatively correlated, what will be the impact on the bank’s credit risk exposure?

a) The loans’ negative correlations will decrease the bank’s credit risk exposure because lower than expected returns on one loan will be offset by higher than expected returns on other loans.

b) The loans’ negative correlations will increase the bank’s credit risk exposure because higher returns on less risky loans will be offset by lower returns on riskier loans.

c) The loans’ negative correlations will decrease the bank’s credit risk exposure because higher returns on less risky loans will be offset by lower returns on riskier loans.

d) There is no impact on the bank’s credit risk exposure.

16. The timeline of the historical GARCH(1,1) estimates of a trading portfolio are as follows:

Which of the following statements is correct with respect to the VaR (99%, 10-day) for 14 February 2009?

a) The 250-day percentile VaR will exceed the GARCH VaR.

b) The 250-day percentile VaR will be less than the GARCH VaR.

c) The 250-day percentile VaR will be identical to the GARCH VaR.

d) None of the above answers is correct.

17. What conclusion can you draw from this graph regarding the VaR model employed?

a) The model systematically overestimates losses.

b) The model systematically underestimates losses.

c) The estimates are from the delta-normal model with constant variance.

d) The estimates are from the delta-normal model with time-varying variance.

18. The measurement error in VaR, due to sampling variation, will be greater with

a) More scenarios

b) Fewer scenarios

c) More scenarios and lower variance of the problem

d) Fewer scenarios and higher variance of the problem

19. Which one of the following four statements correctly defines credit risk?

a) Credit risk is the