A********t
发帖数: 240 | 1 One Stock:
If 95% VaR = - Stock * 1.65 * stdev * sqrt(T)
how to calculate 5% VaR ?
Multi Stocks:
IF 95% VaR = - Vector Stock*(Mean-1.65*stdev)
how to calculate 5% VaR ?
是simply change 1.65 to -1.65? 貌似不太对？请大牛点拨一下？ |
p******i
发帖数: 1358 | |
A********t
发帖数: 240 | 3 greates possible profit rather than loss
【在 p******i 的大作中提到】
: 5% var有什么鸟用？
|
a******h
发帖数: 105 | 4 如果你假设stock是对称的分布，5% var和95% var关于mean对称，所以就是2mean-95%
var |
i**********o
发帖数: 5993 | 5 The formula you are using is analytical VaR, which is not very useful.
VaR usually measures downside risk. Nobody cares about 5% VaR. |
g****3
发帖数: 49 | 6 if that, which kind of VaR is considered in the street? Thanks.
【在 i**********o 的大作中提到】
: The formula you are using is analytical VaR, which is not very useful.
: VaR usually measures downside risk. Nobody cares about 5% VaR.
|
i***6
发帖数: 442 | 7 most of the firms use analytical VaR in London. Some use simulation. But I
think simulation (both historical and monte carlo) is better. |
i*****r
发帖数: 1302 | 8 请问下,historical 和 MC simluation有啥区别
有种是用mean,stdev 模拟出比如1000个,然后取第50个就算5% var,这算哪种?
【在 i***6 的大作中提到】
: most of the firms use analytical VaR in London. Some use simulation. But I
: think simulation (both historical and monte carlo) is better.
|
i***6
发帖数: 442 | 9 if the mean and the std are the description of previous volatilities, then
it should be historical var.
if the mean and std are just standard normal or student-t, then it should be
monte carlo. |
i*****r
发帖数: 1302 | 10 那在MC中，用啥mean和stdev做simulation?
be
【在 i***6 的大作中提到】
: if the mean and the std are the description of previous volatilities, then
: it should be historical var.
: if the mean and std are just standard normal or student-t, then it should be
: monte carlo.
|
