With significant improvements in its theoretical underpinnings, the gravity model has gained renewed interest in the agro-food trade literature. Notwithstanding, there is a dearth of literature examining the relative trade restrictiveness of tariff barriers across a broad range of agro-food sectors. This represents an important research gap, which this study sets out to fill. Furthermore, this research reconciles the application of zero-inflated models with a sectorally disaggregated analysis. More specifically, employing a fully specified gravity equation, a Poisson estimator and variants of the Poisson model (Negative Binomial, Zero-Inflated Poisson, and Zero-Inflated Negative Binomial) provide statistically significant and theoretically consistent estimates, while allowing for the inclusion of zero-trade values. A panel data model with fixed effects is also employed to improve the estimation of the parameters of interest. Estimation results reveal that in the vast majority of sectors examined, import tariffs are found to be statistically significant, whereas export refunds exhibit a statistically smaller role due to the nonsystematic nature of their application in world food markets. Model simulations of tariff barrier eliminations reveal limited trade gains, although there is encouraging evidence of “low” and “lower middle” per capita income country trade gains in wheat, red meat, dairy, sugar, and (particularly) rice markets.