NO.PZ2018122701000038
问题如下：
A portfolio risk analyst, who specializes in large capitalization US stocks, is backtesting the firm’s VaR model using two procedures:
Procedure A: Using the "actual return" approach ,the analyst measures the returns on a portfolio based upon the change in market values of the assets hold in the portfolio from the close of each business day to the close of the next business day.
Procedure B: Using the "hypothetical return" approach ,the analyst measures the returns on a portfolio based upon the change in market values of the assets held in the portfolio from the close of each business day to the close of the next business day, keeping all positions fixed.
The two procedures result in significantly different numbers of exceptions. The most likely cause of the different number of exceptions is:
选项：
A. Poor calibration of the VaR model.
B. Intraday trading in the portfolios.
C. Incorrect return distribution assumptions
used in Procedure A.
D. The reduction of hypothetical returns by commission
fees.
解释：
B is correct.
考点Backtesting VaR
解析：Procedure A "actual return" approach 和B "hypothetical return" approach 最大的区别在于Procedure B的最后一句话“keeping all positions fixed”。因为这一假设，所以两种方法的差别很大，有 “significantly different numbers of exceptions”。
keeping all positions fixed的意思是不论交易日当天的收盘价带来的return变化是多少，始终保持各个投资产品的权重不变。而实际中，只要有收盘价格的变化，资产的权重肯定会因为价格而有调整（比如投资的产品有股票和债券，股票价格在当天上涨，债券价格下跌，那么收盘时股票和债券的权重与前一天相比是不同的）。因此，导致两种方法差别很大的原因是B选项，在交易日产生交易行为，导致组合的头寸发生改变。
我们讲到的两种回测方法都没有针对权重有说明，请问为什么这里会考虑权重呢？
