[Joao Santos Silva]

Dear Aqib Khursheed,

You certainly can do that, but typically imp and exp fixed effects are interacted with the year dummies.

Best wishes,

Joao

[DONG HANGHANG]

Dear Professor Joao Santos Silva,
I would like to ask you something regarding the structural gravity model in Stata.

I got only one exporter country (China) with its trading partners (40) countries.

In this case, if I wanna apply PPML (importer-fixed effects and exporter-fixed effects) to control for the structural items. How can I generate exporter-fixed effects for my case?

And how can I generate the command to control for the Multilateral Resistance Terms (MRT) in structural gravity model?

Thanks very much for ur kind help.

[Joao Santos Silva]

Dear DONG HANGHANG

In your case, you only need importer fixed effects. If you xtset your data by trading partner, you can just use xtpoisson with the fe option. Otherwise, try ppmlhdfe.

Best wishes,

Joao

[hoda assem]

Dear Joao Santos Silva

I have a question about differences between xtpoisson and ppml. I ran the code you gave above and I got different standard errors.

Dear CARLOS ABREO,


The following code should help you to understand the use of these commands:

clear all
cls
sysuse auto
xtset rep78
xtpoisson price mpg weight, fe robust
xi: ppml price mpg weight i.rep78, cluster(rep78)

Best wishes,

Joao

I also ran the same regression using poisson command and ppmlhdfe command and I got the same result as ppml.
My question is why are the standard errors for xtpoisson different from the other commands?

This is the code I ran

clear all
cls
sysuse auto
xtset rep78
*xtposisson*
xtpoisson price mpg weight, fe robust
*ppml*
xi: ppml price mpg weight i.rep78, cluster(rep78)
*poisson*
poisson price mpg weight i.rep78, cluster(rep78)
*ppmlhdfe*
ppmlhdfe price mpg weight, absorb(rep78) cluster(rep78)

Thanks in advance,
Hoda

[Joao Santos Silva]

Dear hoda assem,

To be honest, I do not know the answer to your question. My guess is that one of the methods uses a small sample adjustment that will become irrelevant in large samples. You should be able to check that in the Methods and Formulas section of the manual.

Best wishes,

Joao

[Severin Tamwo]

Dear Professor Santos,
I working on migration, i use PPML regression to obtain migration instruments. I predicted a new migration variable. Now, I would like to obtain a new variable by summing the predicted variable depending to the 20 destination countries to cancel the destination effect. I can't do it on stata.
Best regards,
Sévérin

[Joao Santos Silva]

Sorry, Severin Tamwo, I do not work on migration and therefore I am afraid I do not understand what you want to do.

Best wishes,

Joao