EViews.com

I HP filtered a panel with 21 nations. Then I unstacked the panel, I took the raw variables and I HP filtered the series one by one (I mean country by country). Surprisingly the filtered output is not the same, why? any hints?

Thanks for helping me.

Hi.

I'm sure I understand what you mean by this question. HP filtering is not allowed on panel workfiles (EViews will produce an error message stating this if you try it).

How did you perform the HP filtering?

EViews 5

a) workfile with panel structure
b) open one series
c) Proc/HP Filter
d) you got the original series filtered

Thanks a lot for your prompt reply

What is the build date of your copy of EViews 5?

(You can check this by clicking on Help->About EViews).

"Standard Edition - Aug 20 2004 build"

You're very out-of-date. You should update your copy.

Hi all,

I would like to know any update available on this issue since EViews version 5.

I'm using EViews version 7.2, but having a trouble in HP-filtering using panel data.

My code is: where I have a series named ln_gdp_sa for several countries. I attached my sample data only containing two countries.

Please let me know if there is any way to do HP-filtering for panel data using EViews.

Thank you very much for your help,
Futoshi

You'll have to copy the series into a non-panel workfile.

Hi Gareth,

Thank you so much for your very quick reply!

I'm trying to implement what you advised, but could you fix the following code?

The problem is I cannot copy back the HP filtered series to the paneled page, matching the crossid that is being worked on.

Hi all,

I figured it out by myself, but thank Gareth anyway for letting me know what I have to do.

I needed more lines and thoughts than I expected to fix my codes above, but anyway the following codes worked. Hope this would be a help to somebody else who wants to do panel HP-filtering using EViews.
Please let me know any easier ways and/or any mistakes. The sample data was attached again just for convenience.

Best Regards,
Futoshi

Hi
please help me :( . i have a panel data containing 14 countries for 25 years. i must specify oil shock in those countries. am i true or not? i have to estimate it one by one in all 14 countries? So how should I use them in panel data to estimate asymmetric effect of oil shock in GDP growth?

it doesn´t work to me :(

message error "copy(smpl={strsmpl}) ln_gdp_sa nonpanel\ @src @date @dest @date": strsmpl is not a valid string or scalar name


Please help, im build a panel data set and i want to do HP filter for each company

Hi all,

I figured it out by myself, but thank Gareth anyway for letting me know what I have to do.

I needed more lines and thoughts than I expected to fix my codes above, but anyway the following codes worked. Hope this would be a help to somebody else who wants to do panel HP-filtering using EViews.

Code: Select all

' Open DATA workfile. open ./sampledata.wf1 smpl @all ' Create a page to conduct HP-filtering pagecreate(page=nonpanel) q 1990q1 2012q2 pageselect Untitled1 ' Work on counry by country series temp = @crossid !ncross = @max(temp) delete temp for !i=1 to !ncross ' Copy the series of interest to a page without a panel structure. %strsmpl ="1990q1 2012q2 if @crossid = !i" copy(smpl={%strsmpl}) ln_gdp_sa nonpanel\ @src @date @dest @date pageselect nonpanel ' Detrend using HP-filter. ln_gdp_sa.hpf(lambda=1000) gdp_hp_tmp ' Copy the results to the main page. copy gdp_hp_tmp Untitled1\ @src @date @dest @date pageselect Untitled1 smpl if @crossid = !i series gdp_hp = gdp_hp_tmp delete gdp_hp_tmp next smpl @all

Please let me know any easier ways and/or any mistakes. The sample data was attached again just for convenience.

Best Regards,
Futoshi

Did you copy his line exactly?

yes i do, exactly

The error message would indicate you're missing a %