开发者:上海品职教育科技有限公司 隐私政策详情

应用版本:4.2.11(IOS)|3.2.5(安卓)APP下载

金融民工阿聪 · 2021年01月25日

backtest的检验给的置信度都是默认是双尾检验的吗

问题如下:

A risk manager is analyzing a 1-day 99% VaR model. Assuming 225 days in a year, what is the maximum number of daily losses exceeding the 1-day 99% VaR that is acceptable in a 1-year backtest to conclude, at a 95% confidence level, that the model is calibrated correctly?

选项:

A.

3

B.

5

C.

8

D.

10

解释:

B is correct.

考点:Backtesting VaR

解析:The risk manager will reject the hypothesis that the model is correctly calibrated if the number x of losses exceeding the VaR is such that: 12\frac12

where p represents the failure rate and is equal to 1 - 99%, or 1%; and T is the number of observations = 225. And z = 1.96 is the two-tail confidence level quantile. If: x0.01×2250.01×(10.01)×225=1.96\frac{x-0.01\times225}{\sqrt{0.01\times(1-0.01)\times225}}=1.96

Then, x = 5.18. So the maximum number of exceedances would be 5 to conclude that the model is calibrated correctly.

题目也没说是双尾的,有时候一些考题是说单尾,一些说双尾,这要怎么判断呢

1 个答案
已采纳答案

品职答疑小助手雍 · 2021年01月26日

一般来说,var的计算都是单尾的,var回测都是双尾的哦。

  • 1

    回答
  • 1

    关注
  • 849

    浏览
相关问题

NO.PZ2018122701000037 问题如下 A risk manager is analyzing a 1-y 99% VaR mol. Assuming 225 ys in a year, whis the maximum number of ily losses exceeng the 1-y 99% Vthis acceptable in a 1-yebacktest to conclu, a 95% confinlevel, ththe mol is calibratecorrectly? 3 5 8 10 B is correct. 考点:Backtesting V 解析:The risk manager will rejethe hypothesis ththe mol is correctly calibrateif the number x of losses exceeng the Vis suthat: 12\frac1221​ where p represents the failure rate anis equto 1 - 99%, or 1%; anT is the number of observations = 225. Anz = 1.96 is the two-tail confinlevel quantile. If: x−0.01×2250.01×(1−0.01)×225=1.96\frac{x-0.01\times225}{\sqrt{0.01\times(1-0.01)\times225}}=1.960.01×(1−0.01)×225​x−0.01×225​=1.96 Then, x = 5.18. So the maximum number of exceences woul5 to conclu ththe mol is calibratecorrectly. 如标题,谢谢!

2023-11-02 12:14 1 · 回答

NO.PZ2018122701000037 问题如下 A risk manager is analyzing a 1-y 99% VaR mol. Assuming 225 ys in a year, whis the maximum number of ily losses exceeng the 1-y 99% Vthis acceptable in a 1-yebacktest to conclu, a 95% confinlevel, ththe mol is calibratecorrectly? 3 5 8 10 B is correct. 考点:Backtesting V 解析:The risk manager will rejethe hypothesis ththe mol is correctly calibrateif the number x of losses exceeng the Vis suthat: 12\frac1221​ where p represents the failure rate anis equto 1 - 99%, or 1%; anT is the number of observations = 225. Anz = 1.96 is the two-tail confinlevel quantile. If: x−0.01×2250.01×(1−0.01)×225=1.96\frac{x-0.01\times225}{\sqrt{0.01\times(1-0.01)\times225}}=1.960.01×(1−0.01)×225​x−0.01×225​=1.96 Then, x = 5.18. So the maximum number of exceences woul5 to conclu ththe mol is calibratecorrectly. 95%的双尾检验对应的分位点不是1.65嘛,解析用的是1.96那不变成单尾了

2022-11-03 17:46 1 · 回答

NO.PZ2018122701000037 老师你好,题中一会儿99%一会儿95%,实在分不清,这里用95%来test,所以z=1.96,那么公式里的p为什么不等于0.05呢?为什么要用99%level的0.01和(1-0.01)来计算?麻烦老师一下,谢谢

2022-02-19 23:25 1 · 回答