endogeneity test in PPMLHDFE

[Juan Quicana]

Hi Tom Zylkin Joao Santos Silva,

I am looking for the impact of non-tariff measures on Peruvian exports in 2016. For that purpose, I used the gravity model and the PPMLHDFE command with fixed effects (for importer and HS2 chapter). So, I was worried about the possibility of endogeneity in my model (reverse causality between the non-tariff variable and the level of exports). Thus, I ask you if there is some endogeneity test for the PPML or I could use the Hausman test for that purpose. If there is the case, how can I implement the Hausman test after estimating with PPMLHDFE command?

Here is my command,

ppmlhdfe exports lndistance border language lnRER landlocked PTA ln_tariff prevalence_index_TOT prevalence_index_TOT_PTA, a(importers hs2) vce(cluster distance)

[Joao Santos Silva]

Dear Juan Quicana,

I am not aware of any such tests. The standard in this literature is to use panel data a a three-way gravitiy equation (with importer-time, exporter-time, and pair fixed effects) to account for possible endogeneity of policy variables.

Best wishes,

Joao