[Joao Santos Silva]

Dear Maira javed,

The missing values are likely to represent zero imports, so maybe you can replace them with zeros. Anyway, you can and should use PPML. As for the rest, it is up to you :-)

Best wishes,

Joao

[Maira javed]

Thanks Joao
I am new to the stata forum,can you tell me that after setting panel data in stata and taking logs of variables, should i take log of dependent variable?
And equation for ppml should be this
ppml Mij lGDPi lGDPj lCPIi lCPIj lEPIi lEPIj lDis lpop

[Joao Santos Silva]

Dear Maira javed,

Do not take log of the dependent variable when using PPML, and use the command ppmlhdfe rather than ppml; make sure you check the examples in the help file.

Best wishes,

Joao

[Maira javed]

This is my data file. should i run PPML with FE or only PPML. And to run PPML with FE i have to create dummies for country-pair fixed effects and country-specific effects as well as with country-year effects?
These are my commands
egen id1=group (Id id)
xtset id1 Year
which command i should run
1-ppml Mi lGDPi lGDPj CPIi CPIj EPIi EPIj lDis lpop,cluster(id1)
2-ppml Mij lGDPi lGDPj lCPIi lCPIj lEPIi lEPIj lDis lpop, absorb(Id#Year id#Year Id#id) cluster(Id#id)
3-ppml Mij lGDPi lGDPj lCPIi lCPIj lEPIi lEPIj lDis lpop, absorb(i.Id#c.Year i.id#c.Year Id#id) cluster(Id#id)

i am confused about what command i should run

[Joao Santos Silva]

Dear Maira javed,

I suggest you use the user-contributed ppmlhdfe command, which allows you to deal with the fixed effects using the "absorb" option (ppml has no such option). As always, check the help file before using.

Best wishes,

Joao