EViews.com

Hi, thanks for the videos! I am trying to make a regression to estimate economic convergence betwen various countries I got the model but I don't know how combine o how to relate the data in eviews in order to do it. any information will be helpfull!

You're going to have to provide considerably more detail on what you're trying to do.

Thanks, sorry for the delay I have been working on this, I want to estimate economic convergence across four countries, the teoric model is stated as;

(1/T)*(Ln(Yi,t/Yi,0)=c+B*Ln(Yi,0)+e

where the left side is the growth of GDP during the whole period and in the right side we have an intercept and as variable independent the log of the initial GDP. the theory states that there is economic convergence if there is a inverse relation betwen GDP growth and the inicial level of GDP. (B <0)

I have realize that maybe a good way to do it is using panel data models but I dont know how to set the equation on eviews, I have series of 24 years but following the model is it possible to tell the program to calculate the growth in the way of the model is written and then that relate it only with the initial gdp and how do i write it on the eviews? after a first regression I will add other variables such as Foreing investment, trade, etc, should I use rates of growth of those variables or how do I include then in this regression.

any suggestions are welcome,

Thanks

Hi there! any help? I would appreciate it very much!

I have no idea what you're trying to do.

I think the OP just needs to know how to get the first observation for each cross-section in panel.

So something like ...
?

Thanks, I already set my panel regression getting the results in the attachment the thing is that the coeficients are not significant and I wanted to try the fixed effects metod, getting the error "Near Singular Matrix".

You are only using 4 observations. That's probably not what you want.

Yes the thing is that those are the only four countries I need to study, later i will introduce more countries (but not more than 20) is there a way to make this result more robust? thank you.

You can't used fixed effects if you only have one observation per country.

So I guess I have to change the model but that means that I could not apply or test the convergence theory in the basic form wich says that average gdp growth depends on the inicial gdp. or use other methods.

Right. Or at least it says you can't use fixed effects, which don't make any sense here anyway.

Thank you, I confess I am not experienced in this at all and I am working on this for mi thesis, do you think the results I got are suitable for an economic thesis or not?, even if i make it clear that the coefficients obtained are not significant?

Can you recommend me another econometric way I can test this theory with only four countries and still respecting the theorical model?;
(1/T)*(Ln(GDPi,t/GDPi,0)=c+B*Ln(GDPi,0)+e

Thanks a lot!