Hi Giovanni and Yves,
My is Eliano Marques, i'm a master degree econometrician from ISEG in
Portugal.
I was trying to use your R package with a specific output i cannot get the
same results that i'm getting in Stata.
Could you please let me know if the below is correct or if i'm missing
anything?
On the below model I want to run a model for 62 companies for a 5 year
period:
log(revenue+2)it = log(revenue+2)(-1)it*b1 + log(Cost+2)(-1)it*b2 +
log(Cost+2)it*b3 + vit
modeloutput<- pgmm(log(Revenue+2) ~ lag(log(Revenue+2), 1)+
lag(log(Cost+2), 0:1)
| lag(log(Cost+2), 2:99) +
lag(log(Cost+2), 2:99) ,
data = BD, effect = "twoways", model = "onestep",
transformation = "ld")
summary(modeloutput, robust = TRUE)
I want to include time.dummies in both equation in level and differences.
I get this output:
Coefficients
Estimate Std. Error z-value Pr(>|z|)
lag(log(Revenue + 2), 1) 0.486593 0.146151 3.3294 0.0008703 ***
lag(log(Cost+ 2), 0:1)0 1.165293 0.069329 16.8081 < 2.2e-16 ***
lag(log(Cost + 2), 0:1)1 -0.565378 0.199199 -2.8383 0.0045360 **
>From Stata, running the following code I get this:
xtabond2 lnrevenue l.lnrevenue lncost l.lncost i.year , iv(i.year )
gmm(l.lnrevenue lncost ) robust small
lnrevenue | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+----------------------------------------------------------------
lnrevenue |
L1. | .358987 .1176716 3.05 0.003 .123688
.594286
|
lncost |
--. | 1.185943 .0363891 32.59 0.000 1.113179
1.258708
L1. | -.4332851 .1551862 -2.79 0.007 -.7435992
-.1229711
|
year |
2009 | -.2100918 .0466175 -4.51 0.000 -.3033094
-.1168743
2010 | -.0936157 .0365742 -2.56 0.013 -.1667503
-.020481
2011 | -.0351653 .036502 -0.96 0.339 -.1081556
.037825
|
_cons | .3372286 .194194 1.74 0.088 -.0510863
.7255435
------------------------------------------------------------------------------
Instruments for first differences equation
Standard
D.(2008b.year 2009.year 2010.year 2011.year 2012.year)
GMM-type (missing=0, separate instruments for each period unless
collapsed)
L(1/4).(L.lnrevenue lncost)
Instruments for levels equation
Standard
2008b.year 2009.year 2010.year 2011.year 2012.year
_cons
GMM-type (missing=0, separate instruments for each period unless
collapsed)
D.(L.lnrevenue lncost)
Do you what I be doing wrong to correct this values?
Many thanks in advance,
Eliano
[[alternative HTML version deleted]]