EViews.com

Dear Eviews fanatics!

I'm currently working on my dissertation about Realized GARCH models (a, relatively new (Hansen, Huang, Shek 2010), extension of the GARCH model).
This model uses "realized measures" (i.e. high frequency data). I've got an .csv file with data from a lot of different indexes, running from 1996 to 2009,
for example dow jones returns, nikkei returns, etc. After opening the file with eviews (6), some series are labeled as an alpha
series, while others are normal. I've noticed that for the alpha series, a lot of data is missing and have values NaN (not a number).
As high frequency data wasn't available until, for example, 1999 in the netherlands. Any suggestions how I could change it into a regular series?
I've tried exchanging th NaN values for zero's, but no succes.

Greets from amsterdam!

Posting your csv file is probably a good idea.

of course, my bad..
the file is too large to upload (+-4mb>2mb allowed)
the file can be found on
http://realized.oxford-man.ox.ac.uk/data/download

thanks!

EViews has a couple of different methods of reading files. Try dragging-and-dropping the file onto EViews. When I did that everything came in as numerics with NAs in the right places.

Thanks! But unfortunately it doesn't work.
FALSE:=I cut off the first 3 years of the data set and saved it in an excel file. After that I could import it in eviews and no more the alpha series!
(of course I miss 3 valuable years of indices that did have high frequency data at the time)
It seems eviews labels series as an alpha one if a certain amount of consecutive points in time have letters in them.

Well I'll keep on looking for the solution :eviews6:

*edit: still alpha series after removing first three years, don't know what I was thinking, probably som false-saturday-night-excitement going on in my brain :series: *

See if the attached file helps.

Wow impressive! Thanks a lot! :D
Do you perhaps care to ellaborate for a bit? It's always nice to have some extra data manipulation tricks up in the sleeve.

Again thanks

To be quite honest, I just dragged the csv file onto EViews.
If you're not using 7.1, there's a chance something has improved...

enough said lol!

cheers!