The devastating implications of the theory of second best, despite being well recognized a half century ago, do not have much influence on the current way we teach public economics at undergraduate level, as this paper shows. The current practices basically stick to the first‐best policy framework, and the theory of second best, even if mentioned, is at most marginally important. Some justifications for this textbook framework are called for if we do not think second best is a bad idea. This paper argues that this textbook framework cannot be justified by the absence of irremovable distortions in the real world, and should be considered within the social‐welfare framework to maintain coherence with other parts of the framework. Without calling for a major revamp of the current way of teaching, I suggest that the theory of third best can be adopted to justify this way of teaching in a realistic environment: when policy‐makers are informationally constrained, and when the unaccounted‐for distortion is not large, or when a conservative attitude is adopted for interventions.