Friday, July 23, 2010

Verifying Historical Volatility Calculation

              
This article was written for you, Bryan.

Since you brought up the question of whether volatility is an arbitrary value in your comment on my previous article, I've been curious to find out if I am still half as good as I was a few years back :) .  Since most of the files I did last time were lost, I spent many hours trying to find and finally managed to dig out an annualized 30 day historical volatility of KLCI from 2001-2006 that I've downloaded from Bloomberg (industry standard other than Reuters) last time.

The first thing I did was to plug-in the KLCI index value into the Excel that I've created to calculate the volatility of FCPO and put Bloomberg's data side by side to compare.  Then I tried to find the period that has been used by Bloomberg to calculate the annualized 30 day historical volatility. 

After a few trial and error, I managed to get the same exact value of KLCI's annualized 30 day historical volatility value with Bloomberg.  Bloomberg uses 260 days instead of 252 trading days in a year.

As usual, the screen capture is as below:




Cheers & Happy trading!
                                          

7 comments:

  1. You are welcome. Actually I also want to make sure that I did it correctly.

    ReplyDelete
  2. Dear Mr

    I have read you blog from the beginning and found it very helpful to improve my knowledge about trading. I am based in Barcelona-Spain.
    Do you offer any online courses?


    How is possible to attend your courses?


    Regards

    Aydin

    ReplyDelete
  3. Hi Aydin, I am very glad that you find my blog helpful and congrats for winning the world cup!

    At the moment I don't offer online courses yet, they are all done only in Malaysia. Maybe someday... Cheers & Happy trading!

    ReplyDelete
  4. Hi,

    What is the formula you used for volatility in the volatility column of the excel spreadsheet picture above?

    I am trying to understand how bloomberg calculates historical volatility..

    Thanks,

    Andrea

    ReplyDelete
  5. =STDEV(last 29 logrelative observation)*SQRT(260)

    ReplyDelete
  6. Hi Anonymous,

    Thanks for your excellent reply! I was too busy with my Microstructure research and forgot to come back and reply ;)

    ReplyDelete