plm IV regression does not respect suppressed intercept in instrument matrix

[tappek]

plm's IV estimation does not respect suppressed intercept in instrument matrix. Reproducible example below.

library("plm")
library("ivreg")
library("AER")

data("SeatBelt", package = "pder")
SeatBelt$occfat <- with(SeatBelt, log(farsocc / (vmtrural + vmturban)))
pSB <- pdata.frame(SeatBelt)

formiv1 <- occfat ~ log(usage) + log(percapin) | log(percapin) + ds + dp + dsp
formiv2 <- occfat ~ log(usage) + log(percapin) | log(percapin) + ds + dp + dsp + 0

plm(formiv1, SeatBelt, model = "pooling")
#> 
#> Model Formula: occfat ~ log(usage) + log(percapin) | log(percapin) + ds + dp + 
#>     dsp
#> 
#> Coefficients:
#>   (Intercept)    log(usage) log(percapin) 
#>       7.56412       0.17686      -1.17226
AER::ivreg(formiv1, data = SeatBelt)
#> 
#> Call:
#> AER::ivreg(formula = formiv1, data = SeatBelt)
#> 
#> Coefficients:
#>   (Intercept)     log(usage)  log(percapin)  
#>        7.5641         0.1769        -1.1723
ivreg::ivreg(formiv1, data = SeatBelt)
#> 
#> Call:
#> ivreg::ivreg(formula = formiv1, data = SeatBelt)
#> 
#> Coefficients:
#>   (Intercept)     log(usage)  log(percapin)  
#>        7.5641         0.1769        -1.1723

plm(formiv2, SeatBelt, model = "pooling")
#> 
#> Model Formula: occfat ~ log(usage) + log(percapin) | log(percapin) + ds + dp + 
#>     dsp + 0
#> 
#> Coefficients:
#>   (Intercept)    log(usage) log(percapin) 
#>       7.56412       0.17686      -1.17226
AER::ivreg(formiv2, data = SeatBelt)
#> 
#> Call:
#> AER::ivreg(formula = formiv2, data = SeatBelt)
#> 
#> Coefficients:
#>   (Intercept)     log(usage)  log(percapin)  
#>       0.57379       -0.02646       -0.47789
ivreg::ivreg(formiv2, data = SeatBelt)
#> 
#> Call:
#> ivreg::ivreg(formula = formiv2, data = SeatBelt)
#> 
#> Coefficients:
#>   (Intercept)     log(usage)  log(percapin)  
#>       0.57379       -0.02646       -0.47789

[tappek]

fixed by f3359ff