Hi all,

I have a panel dataset with firms like the one I am posting here:

Code:
* Example generated by -dataex-. To install: ssc install dataex
clear
input int y float id int year byte var5 float(var3 var1 var2 var4 USfirm)
 0  75 1995 0 1.9559152  2.878459  3.548952   3.49954 0
 0  75 1996 0   1.70938  2.892308  3.562962  3.516405 0
 0  75 1997 0 2.0171318   2.91069  3.581572  3.554847 0
 0  75 1998 0  2.662187  2.894193   3.55883  3.547519 0
 0  95 1995 0  2.041887  2.899392   3.62577  3.505842 0
 2  95 1996 0 1.6876222  2.914558  3.650194  3.522758 0
 0  95 1997 0 1.8329792  2.942667  3.664291  3.568601 0
 5  95 1998 0 2.4560704   2.91326 3.6302755 3.5552595 0
48  96 1995 0 2.2747307  2.947563  3.677085 3.5749686 0
57  96 1996 0 1.5987836  2.965781  3.694177  3.598127 0
59  96 1997 0 2.2408223 2.9902856  3.713735  3.642683 0
61  96 1998 0  2.873193  2.942477  3.652357  3.608142 0
 0 102 1995 0  1.785547  3.000866 3.7701905  3.586181 0
 0 102 1996 0  .5521408 3.0160716  3.792323  3.606914 0
 0 102 1997 0 2.3070471  3.038387 3.8095646 3.6571555 0
 0 102 1998 0  2.380391  2.962264  3.710243  3.586995 0
end
format %ty year
What I want is to estimate a regression like that: y = exp(α+ β1*var1 + β2*var2 + β3*var3 + β4*var4 + β5*var5 + β6*USfirm + δt + δi + εi,t )

where δt is time fixed effects and δi is firm fixed effects.

and I have two questions:

1) I have been asked to forward the dependent variable by two lags. What is the appropriate way to do that?

2) The exponential expression means the same as when we do IRR regressions?

Thank you very much in advance.