# VAR Q09

http://www.analystforum.com/phorums/read.php?13,967998,968101#msg-968101 *Spoiler* VAR Question Posted by: Dwight (IP Logged) Date: May 20, 2009 12:54PM Given a daily 5% VAR of \$5 million, which of these statements is/are correct and why? 1, “VAR is a measure of maximum loss, which in this case means we are 95% confident that the maximum 1-day loss is \$5 million.” 2, “VAR is a measure that combines probabilities over a certain time horizon with dollar amounts, which in this case means that one expects to lose a minimum \$5 million five trading days out of every 100.”

1. you can likewise state that there is a 5%chance that you will lose minimum of 5mil in one day

1 is correct. Time horizon is specified ( instead of frequency in 2. ) . probability range is mentioned

1 is incorrect, it should read “95% confident that the maximum 1-day loss will not exceed \$5 million.”

Both are correct. I think it can be stated as % or number of days. Historical method anyone?

1

1 is incorrect, same reason that wake posted

2 is correct Reason: Statement 2 gives the probability of loss, gives the dollar amount expected to be lost an gives the specified time over which the loss is expected to happen. Statement 1 does not give the period over which this VAR is expected to occur.

bpdulog Wrote: ------------------------------------------------------- > 1 is incorrect, same reason that wake posted That is semantics and only applies when we deal with hypothesis…

it say 95% confident that the maximum 1-day loss is \$5 million." isnt this implicit that it will not exceed?

Have you guys even read the curriculum? VAR is not a measure of maximum loss, that one its shortfalls. VAR often underestimates the magnitude and fre- quency of the worst returns, although this problem often derives from erroneous assumptions and models. (Level III Volume 5 Alternative Investments, Risk Management, and the Application of Derivatives , 4th Edition. Pearson Learning Solutions 245).

pimpineasy Wrote: ------------------------------------------------------- > it say > > 95% confident that the maximum 1-day loss is \$5 > million." > > isnt this implicit that it will not exceed? This VAR, is it a maximum 1-day loss expected in one week, one month or one year?

bpdulog Wrote: ------------------------------------------------------- > Have you guys even read the curriculum? VAR is not > a measure of maximum loss, that one its > shortfalls. > > VAR often underestimates the magnitude and fre- > quency of the worst returns, although this problem > often derives from erroneous assumptions and > models. > (Level III Volume 5 Alternative Investments, Risk > Management, and the Application of Derivatives , > 4th Edition. Pearson Learning Solutions 245). > Given a daily 5% VAR of \$5 million, this can also be interpreted to mean that the expected maximum loss “given a 95% confidence” is \$5 million. It is in the curriculum. I read it there…

Wake - "1 is incorrect, it should read “95% confident that the maximum 1-day loss will not exceed \$5 million.” I agree that statement 1 is wrong, but I don’t think your correction is right. It should read…“95% confident that the MINIMUM 1-day loss will BE AT LEAST \$5 million.” Actual losses exceed VAR a lot. I see VAR as a base case level…statistically we would expect to lose AT LEAST \$5 million 5% of the time. Or 95% confident we will lose \$5 million…but we could lose a lot more than that if some crazy unlikely/unexpected event occurs (which is why VAR failed) during the crisis.

june2009 Wrote: ------------------------------------------------------- > Wake - "1 is incorrect, it should read “95% > confident that the maximum 1-day loss will not > exceed \$5 million.” > > I agree that statement 1 is wrong, but I don’t > think your correction is right. > > It should read…“95% confident that the MINIMUM > 1-day loss will BE AT LEAST \$5 million.” > > Actual losses exceed VAR a lot. I see VAR as a > base case level…statistically we would expect to > lose AT LEAST \$5 million 5% of the time. Or 95% > confident we will lose \$5 million…but we could > lose a lot more than that if some crazy > unlikely/unexpected event occurs (which is why VAR > failed) during the crisis. Only the last paragraph of this post is right. The first three paragraphs are at variance with the curriculum.

:-))) this thread is really funny… anyway, I think both are correct. and why? it is maximum/minimum loss with a given probability (over a period), both sentences specify all of this, minimum or at least and maximum or not exceed is same. but it does not say maximum loss possible. (100 % confidence)

Both are correct and are two sides of the same story, in cliche-speak!!! 1. Tells you that your daily loss will not exceed \$5m on 95 out of 100 days 2. Claims that your daily loss will exceed \$5m on 5 out of 100 days I should know. I work in Market Risk in a bank

me.tega is right So… 5% probability you will lose at least \$5 million - 5 out of 100 days you will lose more than 5 95% confident loss will not exceed \$5 million - we won’t lose more than \$5 million on 95 out of 100 days Is that thinking better? I know Schweser is not best for this, but they start out by saying “VAR is used as an estimate of the minimum expected loss (alternatively, the maximum loss)” WTF…is it minimum or maximum??

This is directly from EOC questions. States that VAR can be explained as 95% probability of not losing more than 5M in a given period. or Minimum loss in 5% of instances will be 5MM. The book stated both ways to explain VAR are correct in the EOC question.

5% probability you will lose at least \$5 million - 5 out of 100 days you will lose more than 5 95% confident loss will not exceed \$5 million - we won’t lose more than \$5 million on 95 out of 100 days I like this.