[Tom Zylkin]

Hi KIM SUJIN

As you can see from the second message, your "ban" variable is perfectly explained by the exporter-year and importer-year fixed effects. You should check how this variable is coded.

Regards,
Tom

[KIM SUJIN]

ahh it is coded as 1 or 0 depends on there are sps measures (which is trade ban) between import and export country
it might doesnt recognize that it is related to pair fixed effect??

[Tom Zylkin]

Hi KIM SUJIN

It sounds like one of three things is true:

1. whenever an importer in your data has a trade ban, it applies it to all exporters
2. whenever an exporter is on the receiving end of a trade ban, it receives it from all importers

or

3. some combination of 1 and 2.

What you need for identification is at least one case where a ban is applied to an importer to just some subset of exporters (not all its export partners), with at least some of those exporters not facing bans from other importers.

[KIM SUJIN]

/=could you recommend me any way to regress ban?

thank you

[Tom Zylkin]

Hi KIM SUJIN

Without more information, I cannot. From the sounds of it, I think you need to familiarize yourself more with your data. In particular, you need to spend some time thinking about how the variables you are focusing on are constructed so that you understand where the variation that will identify them should be expected to come from.

Regards,
Tom

[KIM SUJIN]

dear Tom!

yes you are totally right


'Dummies for importer and exporter are included in the sample, excluding one country to avoid perfect collinearity'

maybe this means the same!!


sorry. but do u know how to drop one country..? with stata

[Tom Zylkin]

Hi KIM SUJIN

No this is not the issue. Stata "drops" one dummy regardless here. The issue is that you need to better understand how "ban" is constructed so that you in turn gain some understanding of what type of variation can be used to identify it. With exporter-time and importer-time fixed effects, you cannot identify this variable unless it simultaneously varies by exporter, importer, and time. Currently, stata is telling you that it does not. Perhaps this is a mistake in how you have input the variable or coded it - it's up to you to figure this out (not much I can do for you there unfortunately).

Regards,
Tom

[KIM SUJIN]

thanks a lot!

I appreciate!!!

[KIM SUJIN]

daer tom!

I have just one more question!

if i use different combination of variables
than it could be possible that Ban is not dropped??

Thank you

[Yongzhi Zhao]

Hi KIM SUJIN

When you say time-varying fixed effects, what do you specifically mean? I can tell you that

ppml_panel_sg ..., im(import) ex(export) y(year) trend

will give you exporter-year, importer-year, and exporter-importer fixed effects as well as an exporter-importer-specific time trend. If these is what you want, you should use collapse to collapse your data to an annual frequency.

however, given that your data is monthly, you could just as easily have exporter-month and importer-month FEs in place of the exporter-year and importer-year FEs. In that case, you would replace "year" with a variable identifying the month and year. It depends on what you want to do.

Alternatively, there is another new command I have worked on called ppmlhdfe that provides more flexibility in how you specify the FEs. You could try that and it would not give you an error. However, it would not solve the problem that you need to think more carefully about how your data is structured and what is the right model to apply to it.

I hope this is helpful,
Tom

Hello Tom Zylkin !
I am a PhD in Wageningen University and I am using your command ppml_panel_sg to estimate the gravity equation. I face some problems and hope to hear your comments. Thank you very much in advance!

Here is the thing. I would like investigate the impact of a trade policy (BRI) on bilateral trade flows. I use the ppml_panel_sg with trend options. Now I would like to estimate the Trade Creation and Trade Diversion effect, and I set two "single member" dummies to indicate whether source country (i) and destination country (j) is a member of trade agreement (BRI) respectively. In other words, BRI_i takes value of 1 if i is a member and j is not, zero otherwise. BRI_j takes value of 1 if i is not a member and j is a member, zero otherwise. When I estimate these two dummies, one of them will be omitted. If I understand correctly, in my case, they are collinear with exporter-importer-specific time trend? Are they also collinear with exporter-year FE or importer-year FE? I am a little confused.

Then if I need to include this pair specific time trend terms, can I still use this command to estimate coefficients of these two "single member" dummies?