[R] R: resdiuals of random model estimated by plm function

[Millo Giovanni]

Dear Alfonso,

in a RE model you do not explicitly estimate every single individual effect, but only the variance of the distribution they have been "drawn from". Hence the only (pointwise) residual you can estimate ex-post is the composite one: i.e., the sum.

Best,
Giovanni 

Giovanni Millo, PhD
Research Dept.,
Assicurazioni Generali SpA
Via Machiavelli 3,
34132 Trieste (Italy)
tel. +39 040 671184
fax  +39 040 671160

-----Messaggio originale-----
Da: alfonso.carfora at uniparthenope.it [mailto:alfonso.carfora at uniparthenope.it] 
Inviato: mercoledì 6 novembre 2013 17.30
A: r-help at r-project.org
Cc: Millo Giovanni
Oggetto: resdiuals of random model estimated by plm function

Hi all,


I have estimated a random panel model using plm function.

I have a question about the vector of resduals obtained with the object $residuals.

example:

data("Produc", package = "plm")
zz <- plm(log(gsp) ~ log(pcap) + log(pc) + log(emp) + unemp, model="random", data = Produc, index = c("state","year"))

res<-zz$residuals # vector of the residuals.

the vector res is the sum of the idyosiyncratic (eit) and individual
(ui) component or is only the idyosiyncratic (eit) component?

Thanks
Alfonso





 
Ai sensi del D.Lgs. 196/2003 si precisa che le informazi...{{dropped:12}}