Frequency Conversion

For requesting general information about EViews, sharing your own tips and tricks, and information on EViews training or guides.

Moderators: EViews Gareth, EViews Moderator

EViews Chris
EViews Developer
Posts: 161
Joined: Wed Sep 17, 2008 10:39 am

Re: Frequency Conversion

Postby EViews Chris » Thu Aug 09, 2012 9:30 am

BTW, here's how you could change week alignment to a particular day of the week without needing to know what day of the week your workfile is currently aligned to:

series dateshift = @datefloor(@date, "w")
pagestruct @date(dateshift)
delete dateshift

That will generate a weekly file starting on Mondays. For other days of the week, simply add an offset to dateshift (+1 for Tuesday, +2 for Wednesday, etc.)

nicvri
Posts: 3
Joined: Wed Sep 12, 2012 5:52 am

Re: Frequency Conversion

Postby nicvri » Wed Sep 12, 2012 6:36 am

I have managed to convert my yearly panel data into 5-year average panel data, but the calculation of the averages have gone in the wrong direction: I would have liked the yearly data for e.g. '91-'95 to be averaged into the new datapoint '1995', but instead my new datapoint '1995' is the average of '95-'99, and '2000' is the average of '00-'04.
Unfortunately I have not been able to find a way to solve the problem, and help is greatly appreciated!

How can I make EViews convert my data such that '1995' is the average of '91-'92-'93-'94-'95, '2000' the average of '96-'97-'98-'99-'00, and so on?

Many thanks in advance!

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13294
Joined: Tue Sep 16, 2008 5:38 pm

Re: Frequency Conversion

Postby EViews Gareth » Wed Sep 12, 2012 7:53 am

A simple way would be to just take your existing averages and lag them by one year.
Follow us on Twitter @IHSEViews

nicvri
Posts: 3
Joined: Wed Sep 12, 2012 5:52 am

Re: Frequency Conversion

Postby nicvri » Wed Sep 12, 2012 8:11 am

EViews Gareth wrote:A simple way would be to just take your existing averages and lag them by one year.

Thanks for the fast reply Gareth!
I thought about this, lagging everything by one period, such that if I want to estimate using the averages '80-'95 I would estimate using my current averages '75-'90.
Data would however be slightly off, wouldn't it? Instead of using '76-'80,...,'91-'95 I would be using '75-'79,...,'90-'94.

Or do you mean instead of using '80, '85, etc. I should use '71, '76, etc. and then lag them by one 5-year-period as mentioned above? Or am I thinking too difficult?
As practice, before applying everything on my own data, I'm trying to precisely reproduce the results of a working paper for which I have the EViews workfile. Hence my nitpicking. :D
Last edited by nicvri on Wed Sep 12, 2012 8:27 am, edited 1 time in total.

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13294
Joined: Tue Sep 16, 2008 5:38 pm

Re: Frequency Conversion

Postby EViews Gareth » Wed Sep 12, 2012 8:25 am

I presume that your resultant page has data that looks something like this:

Year X
1990 Val1
1995 Val2
2000 Val3
2005 Val4


Where Val1 is currently the average of 1990-1994, and Val2 is the average of 1995-1999.

Thus if you lag X by one period, then it should be exactly what you want.
Follow us on Twitter @IHSEViews

nicvri
Posts: 3
Joined: Wed Sep 12, 2012 5:52 am

Re: Frequency Conversion

Postby nicvri » Wed Sep 12, 2012 9:13 am

I'm sorry, maybe I've been staring at EViews for far too long today, but I have difficulty understanding your solution: I'm not really sure what you mean with lagging Year X by one period. :oops: I really do appreciate the help though, and I'm sorry for asking so much of your time!

My resulting series look as follows:
Image

Here, 1990 is indeed the average of 1990-1994, 1995 the average of 1995-1999, etc.
The data in the panel workfile from which I'm copying goes back to 1960, so I would have expected a problem like this when I would convert to 5-year averages starting in 1960, but not when I would start at 1975.

So what exactly do you mean by lagging X one period? Because it seems it still doesn't seem to solve the problem of wanting to have 1995 represent the average of 1991-1995.
The only way I'm currently getting close to what I need is to start the series at 1971 and accepting that 1971 represents 1971-1975 and therefore if I'd want to run estimations I need to start in 1971 to reproduce the paper's 1975. Of course: if it works, it works, but using this method for my own data is less desirable since I'm not the only one that will have to use it.
Last edited by nicvri on Wed Sep 12, 2012 9:16 am, edited 1 time in total.

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13294
Joined: Tue Sep 16, 2008 5:38 pm

Re: Frequency Conversion

Postby EViews Gareth » Wed Sep 12, 2012 9:16 am

Code: Select all

series aid2 = aid(-1)
show aid2 aid
Follow us on Twitter @IHSEViews

d952
Posts: 64
Joined: Tue Nov 22, 2011 6:30 am

Re: Frequency Conversion

Postby d952 » Fri Sep 14, 2012 6:28 am

Dear Eviews

I hope you are still active on this post. I am very confused to choose the appropriate conversion method for my data, I have annual data on "crude oil proved reserves" and going to convert it to quarterly data. This variable doesnt change easily and doesnt have fluctuation, but there is monthly exploration and production of crude oil that reduce the amount of "crude oil proved reserves" with almost a constant rate. It is very appreciated if you guide me which method of conversion is more appropriate for this kind of data, I know the first two ones are rejected (constant ones) but I cant decide about the other four ones.

Looking forward to hear from you
Niaz

Behnam
Posts: 3
Joined: Mon Oct 20, 2014 8:37 am

Re: Frequency Conversion

Postby Behnam » Mon Oct 20, 2014 8:56 am

Hi

I have tried many times to do this , but when ever i do "paste special" in the following window the " frequency conversion options" is still off !
i was wondering if you'd mind helping me to overwhelm this issue :( .

Thank you very much

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13294
Joined: Tue Sep 16, 2008 5:38 pm

Re: Frequency Conversion

Postby EViews Gareth » Mon Oct 20, 2014 8:58 am

Describe exactly what you're doing, and post your workfile.
Follow us on Twitter @IHSEViews

Behnam
Posts: 3
Joined: Mon Oct 20, 2014 8:37 am

Re: Frequency Conversion

Postby Behnam » Mon Oct 20, 2014 10:11 am

First of all i arrange GDP per Capita data in an excel file , then drag it to Eviews 7 , after that create a new page by selecting "Specify by frequency/Range..."
then push OK but in the right column the " Frequency Conversion Options " is off.
i don't know how to fix the problem

Thank you so much

Behnam
Posts: 3
Joined: Mon Oct 20, 2014 8:37 am

Re: Frequency Conversion

Postby Behnam » Mon Oct 20, 2014 10:14 am

Here is the image of my mentioned problem.



Image

isanigr8
Posts: 2
Joined: Mon Dec 26, 2016 11:49 pm

Re: Frequency Conversion

Postby isanigr8 » Tue Jan 03, 2017 12:43 am

I have 15 years GDP annual data, can i convert it in the same manner as mention in above slides, which is more efficient "panel to time series " or "time series to panel "

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13294
Joined: Tue Sep 16, 2008 5:38 pm

Re: Frequency Conversion

Postby EViews Gareth » Tue Jan 03, 2017 1:29 pm

How do you want to do the conversion?
Follow us on Twitter @IHSEViews

Ryan1996
Posts: 1
Joined: Mon Apr 11, 2022 6:14 am

Re: Frequency Conversion

Postby Ryan1996 » Mon Apr 11, 2022 7:56 pm

Hi, thanks for the cool method. BTW, would like to know anyone knows what is the logic or intuition behind the frequency conversion? like how the conversation was done?


Return to “General Information and Tips and Tricks”

Who is online

Users browsing this forum: No registered users and 7 guests