Conceptualizing the 'Fat Tax': The Role of Food Taxes in Developed Economies

Public health scholars and practitioners in several countries have called for a broad system of taxes on unhealthy foods, possibly combined with subsidies for certain healthy foods. The typical motivation for these fat taxes is the public health perspective, the idea that it is socially valuable to make health outcomes better regardless of how individuals might trade off risky dietary behaviors with those outcomes. This paper assesses possible alternative rationales for such a tax, examining arguments based on behavioral and cognitive failure as well as traditional economic arguments. The strongest role for a broad, health-based tax on foods would be as a health-insurance system component. Ideally, the tax amount would approximate the expected medical cost from consuming each unit of food, and the tax revenues would cover claims for the associated disease conditions as they arose. This tax/claim system would serve four major insurance-based purposes. First, it would reduce moral hazard by causing individuals to make ex ante payments covering the expected covered costs of risky eating behavior. Second, it would tend to alleviate adverse selection problems by providing implicit risk adjustment for the associated maladies. Third, it would address the incomplete markets problem that consumers are restricted to short-term coverage by introducing a long-term component into the system. Finally, it would aid in separating inherent risk over which individuals have no control from risk that is a function of behavior. Serving these purposes would be especially valuable in developed country systems where coverage is universal or a certain minimum treatment package is available regardless of ability to pay. Whether a fat tax served to further insurance system objectives, general public health goals or other purposes, there are substantial implementation and scientific causality issues. The health impact of particular foods is subject to scientific uncertainty and often depends heavily on the other components in each individual's diet. Disease responses to food products are non-linear in quantity consumed and vary substantially across individuals. Individuals also are heterogeneous with respect to their preferences and self-control. A food tax scheme will result in an industry response and may spur both constructive and detrimental innovation. The paper examines many non-insurance-based rationales for a fat tax, including correcting externalities or internalities (problems of self-control), alleviating bounded rationality (inability to be adequately informed at low cost), and addressing various cognitive or behavioral errors. The optimal form of the tax might be affected by such errors. For instance, a tax based on energy density rather than components such as fat might best address the obesity epidemic. Preliminary estimates show that such a tax, if focused on foods that are nutritionally poor as well as energy dense, might be quite large compared to existing prices for many such foods. The paper also considers the possibility of basing the tax partially or wholly on certain biomarkers (weight, blood chemistry, etc.) and briefly considers litigation as an alternative to taxation.