岳同学2019-12-10 15:45:07
請問老師,如果加入了rare negative event會不會造成風險被高估? 在原版書中提到 "If our data series includes even one observation of a rare event, we may substantially overstate the likelihood of such events happening in the future. Within a finite sample, the observed frequency of this bad outcome will far exceed its true probability." 但是在視頻中 slide 12-114提到 This issue could also lead to an underestimation of risk when sample data include rare negative events.
回答(1)
Dean2019-12-13 15:47:07
同学你好,前面这段话是针对一种风险计量方式VaR而言的,因为通常来说var是有一个期限,一年或三年,那数据量是有限的,可能没有发生过什么大型的尾部事件,如果加入像2008年的这种数据,会导致风险被高估。因为这类风险往往是十年一遇,二十年一遇的,那如果在样本数据选取的过程中没有出现过类似事件,会导致风险被高估,这跟样本选取是有关的。
- 评论(0)
- 追问(0)


评论
0/1000
追答
0/1000
+上传图片