[R] plm issues: error for "within" or "random", but not for "pooling"

[Liviu Andronic]

Dear all
I am working on unbalanced panel data and I can readily fit a
"pooling" model using plm(), but not a "within" or "random" model.
Reproducing the examples in vignette("plm") and in the AER package I
encountered no such issues.

##unfortunately I cannot disclose the data, and it is too big anyway
> dim(ibes.kld.exp.p[x.subs , ])
[1] 13189    34
> summary(ibes.kld.exp.p[x.subs , ]$ibes1y.meanest)
total sum of squares : 28058
      id     time
0.752284 0.018656
> summary(ibes.kld.exp.p[x.subs , ]$employee_kld)
total sum of squares : 9146.5
      id     time
0.637098 0.073421

##fitting a pooling model works OK
> x <- plm(ibes1y.meanest ~ employee_kld, ibes.kld.exp.p[x.subs , ], model="pooling")
> summary(x)
Oneway (individual) effect Pooling Model

Call:
plm(formula = ibes1y.meanest ~ employee_kld, data = ibes.kld.exp.p[x.subs,
    ], model = "pooling")

Unbalanced Panel: n=3041, T=1-16, N=13189

Residuals :
   Min. 1st Qu.  Median 3rd Qu.    Max.
 -6.530  -0.871  -0.189   0.629  13.200

Coefficients :
             Estimate Std. Error t-value Pr(>|t|)
(Intercept)    1.5607     0.0127  122.73  < 2e-16 ***
employee_kld   0.1118     0.0152    7.35  2.2e-13 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Total Sum of Squares:    28100
Residual Sum of Squares: 27900
F-statistic: 53.954 on 1 and 13187 DF, p-value: 2.17e-13
> plmtest(x, "individual")

	Lagrange Multiplier Test - (Honda)

data:  ibes1y.meanest ~ employee_kld
normal = 1675.7, p-value < 2.2e-16
alternative hypothesis: significant effects


##fitting a within or random model fails
> x <- plm(ibes1y.meanest ~ employee_kld, ibes.kld.exp.p[x.subs , ], model="within")
Error in Tapply.matrix(x, effect, mean, ...) : subscript out of bounds
> x <- plm(ibes1y.meanest ~ employee_kld, ibes.kld.exp.p[x.subs , ], model="random")
Error in Tapply.matrix(x, effect, mean, ...) : subscript out of bounds


Would this be an issue with my data (which is a bit specific, since
employee_kld is categorical)? Or perhaps there is an issue in plm()
for unbalanced data?

Please let me know your opinion
Liviu


> sessionInfo()
R version 2.10.1 (2009-12-14)
x86_64-pc-linux-gnu

locale:
 [1] LC_CTYPE=en_GB.UTF-8       LC_NUMERIC=C
 [3] LC_TIME=en_GB.UTF-8        LC_COLLATE=en_GB.UTF-8
 [5] LC_MONETARY=C              LC_MESSAGES=en_GB.UTF-8
 [7] LC_PAPER=en_GB.UTF-8       LC_NAME=C
 [9] LC_ADDRESS=C               LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C

attached base packages:
 [1] tcltk     grid      splines   stats     graphics  grDevices utils
 [8] datasets  methods   base

other attached packages:
 [1] RcmdrPlugin.sos_0.2-0    tcltk2_1.1-1             RcmdrPlugin.Export_0.3-0
 [4] Hmisc_3.7-0              xtable_1.5-6             Rcmdr_1.5-5
 [7] car_1.2-16               ggplot2_0.8.5            digest_0.4.2
[10] reshape_0.8.3            plyr_0.1.9               proto_0.3-8
[13] plm_1.2-3                sandwich_2.2-5           zoo_1.6-2
[16] MASS_7.3-5               Formula_0.2-0            kinship_1.1.0-23
[19] lattice_0.18-3           nlme_3.1-96              survival_2.35-8
[22] fortunes_1.3-7           sos_1.2-4                brew_1.0-3
[25] hints_1.0.1-1

loaded via a namespace (and not attached):
[1] cluster_1.12.1 tools_2.10.1



-- 
Do you know how to read?
http://www.alienetworks.com/srtest.cfm
http://goodies.xfce.org/projects/applications/xfce4-dict#speed-reader
Do you know how to write?
http://garbl.home.comcast.net/~garbl/stylemanual/e.htm#e-mail