Sample size for historical VaR

Discussion on Value-at-Risk Models
Forum rules
DISCLAIMER: We do not warrant or represent that this forum or its content is free of viruses, worms or other code that might be contaminating or destructive. We cannot guarantee that documents or files downloaded from the Site will be free from viruses and we do not accept any responsibility for any damage or loss caused by any virus. Accordingly, for your own protection, you must use virus-checking software when using the forum. You must not post or provide to us via the forum, any document or file which you believe may contain a virus. You must virus check any document or file which you intend to post or provide to us via the forum. You must ensure that any document or file you intend to post to the forum does not contravene any applicable laws or contravene any person's legal rights. We do not accept any responsibility for any damage or loss you may suffer.
Posts: 1
Joined: Sun Apr 28, 2013 8:59 pm

Sample size for historical VaR

Postby lawhow » Fri May 03, 2013 5:03 am

Dear Prof Alexander,

I'm a reader of your book, Market Risk Analysis - VaR and find it extremely beneficial. Here I have one question on the sample size for historical VaR and hope you can help.

In the book, Section IV.3.2.2, you mentioned that for a 1% VaR, we need at least 2000 daily observations, to have at least 20 data points in the 1% tail, while you agreed that there are practical problems with such requirement.

while by "Risk-Based Capital Guidelines: Market Risk" from the US OCC and Fed,
( ... arket-risk)
in page 12, it states that
"Quantitative Requirements for VaR based Measure. Like the January 2011 proposal, the final rule does not change the existing quantitative requirements for the daily VaR-based measure. These include a one-tail, 99.0 percent confidence level, a ten-business-day holding period, and a historical observation period of at least one year."

so that means a sample size of around 250 points.

Can you please help me understand the differences? By your opinion, what is the theoretical limit on the sample size for a basic historical VaR, without other techniques involved? is there a sample size with which we should simply forget about the historical VaR and go apply parametric VaR or MC?

Or maybe we should call OCC and Fed and let them change. :)

Kind regards,

Return to “Volume IV”

Who is online

Users browsing this forum: No registered users and 1 guest