Explanation:
At a 95% confidence level, the expected number of exceptions over 250 days is 5% of 250, which equals 12.5. Using backtesting frameworks such as the Kupiec POF test, an outcome of 0 exceptions falls well outside the 95% non-rejection confidence interval (which is approximately 7 to 19 exceptions). Zero exceptions strongly indicate that the VaR model is overly conservative (estimating VaR too high), which means capital is being allocated inefficiently, indicating the analyst is performing poorly. Seven exceptions is at the lower boundary but generally acceptable. VaR does not predict the mean loss of all days or the severity of the maximum loss (options B and C do not inherently invalidate a 95% VaR threshold).
Ultimate access to all questions.
Q.14 A large investment firm has a trading book whose size changes depending on the perceived trading opportunities among traders. According to the firm’s chief risk analyst, the one-day VaR, at the 95% confidence level, is GBP 110 million. Suppose you were asked to evaluate how good a job the analyst is doing in estimating the one-day VaR. Assuming that losses are identical and independently distributed, which of the following would serve as convincing evidence that the analyst is performing poorly?
A
Over the past 250 days, there is no exception
B
Over the past 250 days, the mean loss is GBP 132 million
C
Over the past 250 days, the largest loss is GBP 1.1 billion
D
Over the past 250 days, there are seven exceptions
No comments yet.