In recent years, the European Union is stressing the importance of moni-toring and evaluating its policies, among which the common agricultural policy plays an important role. Policy evaluation, in order to provide reliable results on which to take important legislative decisions, should rely on robust methodological tools. A recent strand of literature casts some doubts about the reliability of the two-way fixed effect estimator when the effect of a treatment is heterogeneous across groups of units or over time. This estimator is widely used in agricultural economics to esti-mate the effect of policies where effect heterogeneity may be at stake. Using the Euro-pean geographical indication (GI) policy, we compared the two-way fixed effects esti-mator with a novel non-parametric estimator that accounts for the issues created by effect heterogeneity. The results show that the two estimators, consistently with the concerns expressed by the technical literature, may lead to different estimates of the policy effect. This suggests that treatment effect heterogeneity is likely a concern when assessing the impact of GI-type policies. Therefore, the use of the standard estimator may lead to misleading conclusions and, as a result, to inappropriate policy actions.