题目
Analyst Sue observes that equity returns exhibit leptokurtosis but are symmetrical; i.e., skew is zero. Her value at risk (VaR) model assumes returns are normal. Which of the following errors is MOST LIKELY?
选项
A.Her VaR model will neither under- nor over-estimate the actual VaR
B.Her VaR model will under-estimate VaR at all confidence levels
C.At low (high) confidence levels her VaR model will over-estimate (under-estimate) actual VaR
D.Her VaR model will over-estimate VaR at all confidence levels
答案
C
解析
At low (high) confidence levels her VaR model will over-estimate (underestimate)actual VaRA leptokurtic distribution has a higher peak and heavy tail than the normal density with the SAME variance. Therefore, depending on the extent of excess kurtosis, normal VaR will overstate actual VaR up to a certain confidence level, then as the heavy-tails dominate, will certainly underestimate actual VaR.在低（高）置信度下，她的VaR模型将高估（低估）实际VaR矮峰分布的峰和重尾比具有相同方差的正常密度高。 因此，根据过度峰度的程度，正常VaR会高估实际VaR到一定的置信度，然后随着重尾支配，肯定会低估实际VaR。