Association for Public Policy Analysis & Management

Well-designed public policies that correct market failures improve economic efficiency and social welfare. Standard examples include regulations or taxes to reduce negative externalities. Behavioral economics research provides evidence that in addition to market failures, under predictable circumstances individuals fail to make fully optimizing decisions about their own welfare. Tobacco use is a potential example. Present bias, projection bias, or other biases might mean that individuals fail to fully account for the adverse consequences: their current tobacco use imposes internalities on their future selves. In this paper, we explore the benefits and costs of using regulations or taxes to correct smoking internalities. We use a revealed preference/ sufficient statistics/ reduced‐form approach to behavioral welfare economics (Mullainathan et al. 2012, Chetty 2015). We examine two types of tobacco regulations: regulations that “nudge” consumers to make more fully optimizing decisions; and regulations that use “soft paternalism” to reduce tobacco use. We compare these regulations to an excise tax. As in standard cost-benefit analysis, our approach infers consumers’ preferences as revealed by their cigarette consumption choices. However, to capture internalities we distinguish “suspect choices” that might not be reliable indicators of preferences from “non‐suspect choices” that are presumed to reflect true experienced utilities. When nicotine is addictive, reducing non‐optimal cigarette consumption by youths improves their future welfare, but the welfare gains depend on whether adult smokers continue to make non‐optimal choices. When adult smokers are nonaddicted or make optimizing decisions given their addiction, the envelope theorem implies their future behavioral responses do not contribute additional welfare gains. Relying on the envelope theorem allows us to express the social welfare impacts of tobacco regulations and taxes in terms of sufficient statistics – reduced‐form elasticities rather than deep structural preference parameters (Chetty 2009).

Levels of nicotine addiction and the extent of non‐optimizing cigarette consumption are difficult empirical questions. We use data from the Tobacco Use Supplements to the Current Population Survey to bring evidence to bear on these questions. We use the empirical evidence to estimate the costs and benefits of tobacco regulations and taxes. We find that a well-targeted nudge regulation could create substantial positive net benefits. Soft paternalism is likely to create negative net benefits, because it is like a tax that does not yield any revenue. An excise tax creates positive net benefits only if smoking internalities are relatively frequent and large. We conclude with a discussion of whether the FDA regulations made possible under the 2009 Tobacco Control Act are likely to be well-targeted nudges or soft paternalism