OUtput gap

For questions regarding the import, export and manipulation of data in EViews, including graphing and basic statistics.

Moderators: EViews Gareth, EViews Jason, EViews Steve, EViews Moderator

Gohou
Posts: 6
Joined: Mon Apr 19, 2010 10:59 pm

OUtput gap

Postby Gohou » Wed Apr 21, 2010 4:08 am

Hi all,
Could you please show me how to get output gap from GDP data?
i genr log(GDP) but struggling to get get output gap.
could you please show the right equation to write?
Thanks
Best regards

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13307
Joined: Tue Sep 16, 2008 5:38 pm

Re: OUtput gap

Postby EViews Gareth » Wed Apr 21, 2010 7:16 am

Could you define what you mean by "Output Gap"
Follow us on Twitter @IHSEViews

nadja123
Posts: 72
Joined: Thu Aug 06, 2009 10:43 am

Re: OUtput gap

Postby nadja123 » Wed Apr 21, 2010 2:09 pm

Hi Gohou,

I think you need to use a sort of filtering or decomposition. What I know EV offers the Hodrick-Prescott filter. Open your GDP series in a separate window and check out Proc / Hodrick-Prescott filter. Be careful to save the remainder to trend as your outuput gap.

PS I just love the QMS Gareth's sort of irony. If there were any more expressive smilies available I'd just add a few. Really.

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13307
Joined: Tue Sep 16, 2008 5:38 pm

Re: OUtput gap

Postby EViews Gareth » Wed Apr 21, 2010 2:22 pm

No irony intended. I really have no idea what is meant by "output gap". Without knowing what they were talking about, hard for me to tell them how to do it.
Follow us on Twitter @IHSEViews

nadja123
Posts: 72
Joined: Thu Aug 06, 2009 10:43 am

Re: OUtput gap

Postby nadja123 » Wed Apr 21, 2010 2:45 pm

Was this not a question to be replaced under Econometric Discussions? No, okey. Sorry for my PS, I just provoked a bit. I think you and all the moderators are doing a marvellous job and I am happy that this forum exists. :D :D At the moment, for instance, I would really need that my newest post under Programming be replied. Thanks.

Gohou
Posts: 6
Joined: Mon Apr 19, 2010 10:59 pm

Re: OUtput gap

Postby Gohou » Thu Apr 22, 2010 12:18 am

Dear Nadja,
I would thank you a lot for your help. I also understand Gareth's point.
in fact I have the industrial production index (IPI) as a proxy of GDP.
Dear nadja, I tried HP filter and obtained a trend, should I consider these series as output gap?

With the IPI, I did the following to get the output gap.

1) log(IPI) c @trend @trend^2
2) 100*RESID as the output gap to be one of the explanatory variables.

is this make sens?
Thanks for your help!
Gohou
Best regards

nadja123
Posts: 72
Joined: Thu Aug 06, 2009 10:43 am

Re: OUtput gap

Postby nadja123 » Thu Apr 22, 2010 4:09 am

No, it makes no sense.

Try precisely what I wrote: apply the H-P filter to your series (whether it is gross domestic product or industrial production). In EViews 6, you would specify names for the smoothed series (e.g. gdp_hptrend) and the cycle series (gdp_gap) in the H-P dialog. Output gap is the cycle series.

Please, first try to do what I recommended before you post more questions on this. Eventually plot both of the series (original one and the cycle series) and check out "output gap" on the internet. I think it is good for you if you also try to understand what and why you are doing.

Good luck!

Gohou
Posts: 6
Joined: Mon Apr 19, 2010 10:59 pm

Re: OUtput gap

Postby Gohou » Thu Apr 22, 2010 6:39 am

Hi nadja,
thanks a lot! it works! now i have 3 series in the window , hptrend01, my IPI series and finaly the IPI_GAP (as i specified in the "cycle series").
You are the best.
In fact i am trying to regress a african central bank rate decisionagainst inflation, outpu gap, and other explantory variables.
I also learn from the Eviews illustrated book that we should use log values when doing regressions. but my gap series contain negative values.
can i still use log values for all these variables? specially interest rate , lagged interest rate, inflation and output gap (IPI_Gap).

For the inflation series, I used harmonised consumer price index (hcpi) such that :
inflation = 100*(hcpi - (hcpi(-12))/hcpi(-12).
hope this make sens.

Thanks
Best regards

nadja123
Posts: 72
Joined: Thu Aug 06, 2009 10:43 am

Re: OUtput gap

Postby nadja123 » Thu Apr 22, 2010 7:10 am

No, it makes no sense to take logs of interest rate, inflation rate and the computed output gap. Output gap and inflation rate take positive as well as negative values.

You may take logs of so-called real economy variables, such as gross domestic product or your industrial production. But do take the logs of these series as a first step, before you use the procedure (as described in previous post) to compute the output gap .

As for your HICP inflation: If you have "inflation index" data, from monthly data you obtain the year-over-year inflation rate as (index_(t) / index_(t-12) - 1)*100.

nadja123
Posts: 72
Joined: Thu Aug 06, 2009 10:43 am

Re: OUtput gap

Postby nadja123 » Thu Apr 22, 2010 7:13 am

In fact your hcpi formula is okey, just written in a different way.

Gohou
Posts: 6
Joined: Mon Apr 19, 2010 10:59 pm

Re: OUtput gap

Postby Gohou » Thu Apr 22, 2010 7:31 am

Great thanks...
I will rebuild the series first using log of all original variables and second do the HP against log(IPI) to get the IPI_Gap (cycles series).
and se the difference with IPI_Gap of original IPI.
Here you have the data attached.
thanks for your help Nadja.
Attachments
data.xls
(27.5 KiB) Downloaded 762 times
Best regards


Return to “Data Manipulation”

Who is online

Users browsing this forum: No registered users and 16 guests