The standard neoclassical conception of rationality is narrow. It was created mainly for technical purposes such as making economic models internally consistent and mathematically tractable. We advocate a broader notion, especially for normative purposes, of “inclusive rationality.” This rationality concept is consistent with many of the choices and behaviors attributed to cognitive biases by behavioral economists. In fact, these biases can even be useful in helping people achieve their purposes. Although a narrow form of rationality may be more easily falsified, the broader form is the normatively relevant one.

We summarize the arguments against behavioral paternalism laid out in the preceding chapters. Common threads that unify these lines of argument include: the complexity of inclusive rationality, the indeterminacy of paternalistic welfare criteria, the role of incentives and learning, and the rush to policy. We then consider a range of “escape routes” or defenses of behavioral paternalism, including: reverting to objective-welfare paternalism, appealing to obviousness, shifting the burden of proof, loosening the definition of paternalism, relying on the “libertarian condition,” invoking the inevitability of choice architecture, focusing on the irrational subset of the population, relying on extreme cases, treating behavioral paternalism as a toolbox, and invoking fiscal externalities. We conclude with a series of recommendations: replacing neoclassical (puppet) rationality with inclusive rationality, rejecting the paternalism-generating framework, having reasonable expectations of policymakers, and maintaining important distinctions such as public vs. private and coercive vs. voluntary interventions.