Multiplying two fixed effects variables e.g. i.importer*i.exporter

[Tilo Gashi]

Hi,

Does anyone know if, besides including for example importer and exporter fixed effects in your regression, adding a command that includes a multiplication of these such as i.exporter*i.importer is possible?
I have seen the version that includes a manual procedure with tabulate but then too many variables are created and because I have a lot of observations Stata is blocked. Couldn't find anything by now in Google.

thanks

[Rich Goldstein]

in general, I believe that you will find the answer to your question in "help fvvarlist"; note, however, that your question is very unclear: what estimation command are you using? how many distinct values do each of importer and exporter have? please read the FAQ to increase the probability of getting a useful answer (by asking better questions)

[Tilo Gashi]

in general, I believe that you will find the answer to your question in "help fvvarlist"; note, however, that your question is very unclear: what estimation command are you using? how many distinct values do each of importer and exporter have? please read the FAQ to increase the probability of getting a useful answer (by asking better questions)

Thanks for the help but don't be offensive like this. You are not obliged to answer my question and not required to make comments about the structure of the quesiton. Have fun!

[Carlo Lazzaro]

Tilo:
please, do not take it personally.
Rich was simply highlighting the unclear parts of your query: whenever you post, you implicitly allow other listers to comment on the way you posted (with no hard feelings or bad words, as is clearly reported in the FAQ, that everybody is kindly requested to read before her/his first post).
Even engaging myself in reading behind the lines, I cannot fing any rudeness in Rich's reply.
As Rich said, you're probably looking for an interaction between two terms of your model.
Unfortunately, as you did not post what you typed and what Stata gave you back (as all of us are requested to do, as per FAQ#12), set aside guesswork, I can only echo Rich' advice: please take a look at -fvvarlist-.
I do hope that this misunderstanding will not spoil your fun of participating in this forum.

[Tilo Gashi]

Tilo:
please, do not take it personally.
Rich was simply highlighting the unclear parts of your query: whenever you post, you implicitly allow other listers to comment on the way you posted (with no hard feelings or bad words, as is clearly reported in the FAQ, that everybody is kindly requested to read before her/his first post).
Even engaging myself in reading behind the lines, I cannot fing any rudeness in Rich's reply.
As Rich said, you're probably looking for an interaction between two terms of your model.
Unfortunately, as you did not post what you typed and what Stata gave you back (as all of us are requested to do, as per FAQ#12), set aside guesswork, I can only echo Rich' advice: please take a look at -fvvarlist-.
I do hope that this misunderstanding will not spoil your fun of participating in this forum.

I wanted to avoid this kind of coding:

But, neither exporter#importer nor exporter##importer from the fvvarlist is giving the same result as the one using the code above (so inserting into regression alpha*). And, I don't want to use that code that generates the dummies practically as I have a big dataset and Stata can't cope with it!
Any idea?

[Tilo Gashi]

.

[Clyde Schechter]

I have been following this thread, and I still don't understand what specific difficulty you are having. Have you tried this?

regress /* or other command such as xtreg, whatever */ my_outcome_variable i.importer##i.exporter /* and maybe other variables*/

If so, what results did you get and how do they differ from what you want?

To assure clarity and fidelity, please show us the actual command and output by copying from the Results window or your log file and pasting in a code block. (See FAQ #12 for how to do this.)