[R] issues in plm using random effect model

[Nina Schönfelder]

Dear Kailas Gokhale,

The negative individual variance is not a problem with your code or plm. 
It a property of your data. Please check the posts of Giovanni Millo on 
this topic:

[R] R: plm random effect: the estimated variance of the individual 
effect is negative
Millo Giovanni Giovanni_Millo at Generali.com
Sat Jan 5 10:10:01 CET 2013

You can find the posts in the archive by rseek.org.

Kind regards,

Nina Schönfelder

-----
FernUniversität in Hagen
Fakultät für Wirtschaftswissenschaft
Lehrstuhl für Volkswirtschaftslehre,
insbes. Makroökonomik
58084 Hagen

E-Mail: Nina.Schoenfelder at FernUni-Hagen.de
Telefon: +49 2331 987 - 2379
Fax: +49 2331 987 - 391

Hausanschrift:
Informationszentrum (IZ, ehemals TGZ)
Universitätsstr. 11
Raum B110

Am 06.06.2017 um 12:00 schrieb r-help-request at r-project.org:
> Message: 1
> Date: Mon, 5 Jun 2017 15:41:23 +0530
> From: Kailas Gokhale<kls.gokhale at gmail.com>
> To:r-help at r-project.org
> Subject: [R] issues in plm using random effect model
> Message-ID:
> 	<CAMO91N4oCvTf1ZkHLqQ4t3SAbVw=VYLM-6JKt49CPbw+kUBiGg at mail.gmail.com>
> Content-Type: text/plain; charset="UTF-8"
>
> Dear Sir,
>
> Thank you for accepting my request for registration on this site.
> I am trying to solve panel data problems using plm package , but while
> suing random effect model i am getting following messege saying
> "
>
> Warning message:In sqrt(sigma2) : NaNs produced
>
> "
>
> In some other cases i am getting message saying where TSS = NA , that I am
> not understanding
> I am sending you my code along with out put.
>
> Kindly help me .
>
> I am sending you my code and output for your kind reference. data file is
> also attached
>
>
> rm(list=ls())
> library(MASS)
> library(bdsmatrix)
> library(zoo)
> library(nlme)
> library(sandwich)
> library(car)
> library(lmtest)
> library(plm)
>
> data1<- read.csv(file.choose(),header=TRUE,sep=",")
> D<-na.omit(data1)
> attach(D)
> Pdata<-plm.data(D, index=c("CT","T"))
>
> pool11<- plm(Y~X1+X2,data = Pdata, model="pooling")
> # pool21<- plm(Equity.dividend.1~ Profit.after.tax.1+ LaggedDivd.1+
> log(Size.1)+factor(CompanyName)-1,data = Pdata, model="pooling")
> fixed.mod1<- plm(Y~X1+X2,data = Pdata, model="within")
> rand.mod1<- plm(Y~X1+X2,data = Pdata, model="random")
>
>
> summary(pool11)
> summary(fixed.mod1)
> summary(fixef(fixed.mod1))
> summary(rand.mod1)
>
>
> ##################output####################
>
> Oneway (individual) effect Pooling Model
>
> Call:
> plm(formula = Y ~ X1 + X2, data = Pdata, model = "pooling")
>
> Unbalanced Panel: n=6, T=4-6, N=34
>
> Residuals :
>     Min. 1st Qu.  Median 3rd Qu.    Max.
> -19.400  -9.810  -0.648   8.490  23.900
>
> Coefficients :
>               Estimate Std. Error t-value  Pr(>|t|)
> (Intercept) 25.229162   6.858418  3.6786 0.0008847 ***
> X1           0.016438   0.046905  0.3504 0.7283744
> X2          -2.231250   2.220346 -1.0049 0.3227198
> ---
> Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1
>
> Total Sum of Squares:    5082
> Residual Sum of Squares: 4892.7
> R-Squared:      0.037241
> Adj. R-Squared: 0.033955
> F-statistic: 0.599566 on 2 and 31 DF, p-value: 0.55529>
> summary(fixed.mod1)Oneway (individual) effect Within Model
>
> Call:
> plm(formula = Y ~ X1 + X2, data = Pdata, model = "within")
>
> Unbalanced Panel: n=6, T=4-6, N=34
>
> Residuals :
>      Min.  1st Qu.   Median  3rd Qu.     Max.
> -24.0000  -8.0400  -0.0795   6.6300  25.1000
>
> Coefficients :
>      Estimate Std. Error t-value Pr(>|t|)
> X1  0.065306   0.060090  1.0868   0.2871
> X2 -3.082215   2.514602 -1.2257   0.2313
>
> Total Sum of Squares:    4791.9
> Residual Sum of Squares: 4380.7
> R-Squared:      0.085822
> Adj. R-Squared: 0.065628
> F-statistic: 1.22042 on 2 and 26 DF, p-value: 0.31146>
> summary(fixef(fixed.mod1))  Estimate Std. Error t-value  Pr(>|t|)
> A  33.2672    10.4360  3.1877 0.0014340 **
> B  21.9300     9.2930  2.3598 0.0182831 *
> C  27.6590     7.9522  3.4781 0.0005049 ***
> D  21.9369     9.4271  2.3270 0.0199647 *
> E  17.6243     8.6149  2.0458 0.0407766 *
> F  23.8578     9.2198  2.5877 0.0096625 **
> ---
> Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1>
> summary(rand.mod1)Oneway (individual) effect Random Effect Model
>     (Swamy-Arora's transformation)
>
> Call:
> plm(formula = Y ~ X1 + X2, data = Pdata, model = "random")
>
> Unbalanced Panel: n=6, T=4-6, N=34
>
> Effects:
>                   var std.dev  share
> idiosyncratic 168.49   12.98  1.117
> individual    -17.60      NA -0.117
> theta  :
>     Min. 1st Qu.  Median    Mean 3rd Qu.    Max.
> -0.6366 -0.6366 -0.6366 -0.5983 -0.6366 -0.3106
>
> Residuals :
>     Min. 1st Qu.  Median    Mean 3rd Qu.    Max.
> -21.000 -11.400   0.913   0.114   9.040  24.000
>
> Coefficients :
>                Estimate Std. Error t-value  Pr(>|t|)
> (Intercept) 26.3963638  6.5706835  4.0173 0.0003482 ***
> X1          -0.0066621  0.0433425 -0.1537 0.8788364
> X2          -2.1087903  2.1533566 -0.9793 0.3350111
> ---
> Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1
>
> Total Sum of Squares:    5548.4
> Residual Sum of Squares: 5382.1
> R-Squared:      0.037133
> Adj. R-Squared: 0.033856
> F-statistic: 0.479037 on 2 and 31 DF, p-value: 0.62389Warning
> message:In sqrt(sigma2) : NaNs produced
>
>
>
> with warm regards
> kailas D. Gokhale