Abstract:
Campos et al. (2021) (Economics Letters) show that including domestic trade flows in the gravity model is important for estimating the effect of trade agreements (TAs), regardless of how these flows are measured. Without domestic trade flows, TA effects are small, negative, and statistically insignificant. The opposite holds once domestic flows are included. Using the Campos et al. (2021) data, we show that their results are caused by the big size difference between international trade and domestic trade flows in combination with the Poisson Pseudo Maximum Likelihood (PPML) estimator. We find that the TA variable is not statistically significant in the Campos et al. (2021) data.