Dairy Trade Reconsidered: Gravity and the Distributional Effect of Non‐Tariff Barriers and Distance

This study examines the average and distributional effects of trade policy and friction variables on dairy trade flows. To capture average effects, we utilize the recently extended Poisson Pseudo Maximum Likelihood (PPML) methods, which rigorously address potential bias for coefficient and standard error estimates due to the incidental parameter problem in two‐way and three‐way gravity models. Additionally, we are among the first to apply a new distribution regression method to study counterfactual distributional effects in a two‐way gravity model. Using PPML methods, we find that sanitary and phytosanitary (SPS) measures hinder dairy trade, while technical barriers to trade (TBT) boost it on average. Findings show that the model specification with separate intranational and international distance variables better fits the dairy trade data than using a combined distance variable. Using the distribution regression method, we find that the effect of TBT is positive but statistically insignificant. In contrast, trade agreements exhibit positive and heterogeneous effects, particularly in the upper quantiles of dairy trade flows. Additionally, distance shows a strong negative heterogeneous impact on the upper quantiles of dairy trade flows. Our findings could guide policymakers in formulating trade policies to improve global dairy trade.