[R] Suitable package for carrying out sigma and beta convergence in panel data

[Roger Bivand]

Amarjit Singh Sethi <set_alt <at> yahoo.co.in> writes:

> 
> Dear all

> I wish to carry out sigma- and beta-convergence analysis in respect of
> panel data [wherein current value of
> one of the variables needs be regressed upon suitably transformed lagged
> values of another
> variable(s)]. I am quite new to the R-language and am not very much aware
> of the availbaility of suitable
> package(s)/ code in the language. 

It is always useful to look at the list of Task Views on CRAN, centrally
on http://cran.r-project.org/web/views/, or on your mirror of choice.

There you will find an Econometrics task view, which is possibly what you are
looking for - although "sigma and beta convergence" is not very obvious at
all except for some flavours of economists. Panel estimation is mentioned
in the paragraph on "Further regression models", specifically the "plm"
package.

> Can any one help me in letting me
> know of appropriate
> package/procedural steps to undertake the anlaysis. Kindly let me
> know as well, format of the input data
> file for such a package.

I'm afraid that you'll have to decide which kinds of sigma or beta convergence
you want to address yourself - the literature is wide and does not give 
clear indications.

The plm package model fitting functions are described in a vignette, accessed
from the package page on CRAN (it will help to have Baltagi's standard 
text to hand). That should show how to pass data to the model fitting 
functions, but you may wish to review the general description of
how to use data frames and formula arguments to model fitting functions
in the online "An Introduction to R" manual (on CRAN, Manual in the left
navigation bar). The R Data Import/Export manual in the same place describes
how to get data into R in general. The manuals are installed locally
with R too, if you prefer to read them offline. The vignette mentioned 
above will also be installed locally if you install the plm package.

Data import, data frame objects, and formula arguments are generic, so 
learning them once lets you use very many other model fitting functions 
too. Pay attention, for example, to what formula handling does to a factor 
(categorical variable) - it generates dummies automatically. 

I hope this will let you get started.

Roger Bivand

> Regards

> Amarjit Singh Sethi
>