URGENT: Estimation Gravity Model by Eviews

[white rose]

hi every body ..

im a new member and i have a problem with my thesis, estimation part exactly .. i wonder if you can help me .. hope you can as soon as possible.

my problem is: i dont know what the type of my data and how i can organise it in eviews..
my model is gravity model .. the dependent variable is exports between 17 countries .. in each observation y = export from country i to j in 1996 , 1997 , 2007, 2008
there are 4 years

the independent variables is
GDPi
GDPj
CAPITA ( the difference between gdp per capita in each pair of countries)
DIST (geographic distance between each pair of countries)
and 3 dummy variables BOR, REGN, GAFTA.

Please tell me what the type of my data and how i can put it in eviews and estimate it ,also how i can deal with dummy varibles?

please if you can help me insert some images to explain the prosses

my data shown in the attached file ..

thank u very much..

[white rose]

please .. help me if u can

[bensamen]

You can use panel data analysis to estimate your gravity model. Please follow this link on creating a panel workfile http://forums.eviews.com/viewtopic.php?f=7&t=74

I downloaded also your data and will analyze it when i have time.

[white rose]

thank you bensamen very much .. but my main problem is:

1) is my data pool data or panel data and what the difference?

2) what the main sectors i must identify in the "pool" object?
like this

_ alg
_ bah
_mor
_sau


or by pairs of counries, like:
_ alg bah
_ alg jor
_ alg mor
_ alg sau

then the second country
_ bah alg
_ bah jor
_ bah mor
_ bah sau

[white rose]

QMS Gareth .. QMS Moderator

please help me ..

[bensamen]

We often loosely use the term panel data to refer to any dataset that has both cross-sectional and time-series dimension. More precisely it’s only data following the
same cross-section units over time. Otherwise it’s a pooled cross-section. After looking at your data, i can suggest to use cross section because you dont first have an regular time series (1996 , 1997 , 2007, 2008).
So far, i dont know the purpose and hypothesis on your studies but when dealing with such data, we practically estimate a cross section regression by year (smpl if year=2007 and then do regressions).

[white rose]

thank you bensamen .. i will try what you tell me

regards,

[nur]

Hi,

I'm sorry, I am very new to eviews, I don't even know how to post a new forum, thus I use a reply to this post as I'm having the same problem with you.

I'm using a gravity model too, to estimate the impact of immigrants on trade. My data consists of 10 countries, spanning from 1997-2008. Thus I guess it is a panel data then.

I have a few questions and I am extremely grateful if any of the users and moderators can help me.

1/ I've done the unit root test for stationary test for each variable. I just click each variable and go to VIEW-UNIT ROOT TEST. Is it the correct procedure? But it states that I did a panel unit root test then, and I want to employ pooled OLS, is ok?

2/ I wan't to employ pooled OLS, rather than Fixed Effect as I have time invariant variable (LANGUAGE, DISTANCE). However, I do not know how to generate time dummy, even as I follow the suggestion (I write a command in the generate series by equation -group dummies @expand(@year) - is this correct?) I still get the respond from eview syntac error. How do I generate the time dummy then?

Thank you so much for your kind attention to this matter. Please help. :eviews6: :eviews6:

p/s: how do attach my file here?

[rickz.k]

Hi there,

I am completely new to E-Views and I really need to carry out my own Gravity Model. The aim of my Gravity Model is to illustrate the impact of internet users on trade flows. I have read many studies and one I came across was http://www.businessjournalz.org/article ... n61617.pdf. I want to focus on the African countries and show the impact the internet can have and I am unsure as to how to carry out the model. If someone could kindly guide me step by step on how to create one I would really appreciate it.

I am thinking of taking 8 African developing countries and 8 developed countries (non-African) and test the relationship using the gravity model. I have data on distance and internet users in the countries. However I do not have a clue on how to create the dummy variables nor what other data I require and how to apply it in E-Views.

I know im asking for a lot here but honestly if someone could help me I would honestly be so grateful!

Kind regards

Ricky

[rickz.k]

Gareth if you could possibly help me it would be much appreciated,

Thanks!!

Ricky

[ditz]

im trying to get the exact Distance between Fiji and Australia...could you help me on how to get the data on this...thank you