« Back to Results

Sin Taxes

Paper Session

Saturday, Jan. 5, 2019 10:15 AM - 12:15 PM

Atlanta Marriott Marquis, International 1
Hosted By: American Economic Association
  • Chair: Alex Rees-Jones, University of Pennsylvania

Price Isn’t Everything: Behavioral Response around Changes in Sin Taxes

Alex Rees-Jones
University of Pennsylvania
Kyle Rozema
University of Chicago


Taxes change behavior. But how does this change arise? In traditional economic models, change is achieved through the price channel: assuming all else is held constant, taxes increase prices and thus decrease demand. However, the assumption that all else is held constant may be violated in the course of a legal change, in part because the expressive powers of law can be initiated in the lawmaking process. We examine violations of this assumption in a particular policy domain: discouraging smoking with cigarette taxes. We document a marked increase in related media coverage, lobbying efforts, place-based smoking restrictions, and anti-smoking appropriations in the time period surrounding a tax law change. The intensity of these factors is directly associated with decreases in cigarette consumption, even when controlling for tax changes, and in a manner that could be confused with price effects. Illustrating this effect, we find that cigarette consumption declines well in advance of a vote on a tax law change. Even when a referendum fails, this decline persists despite taxes remaining constant. Our results suggest that price effects may have a surprisingly small role in the behavioral response seen around tax law changes.

Cigarette Tax Avoidance in the United States and Canada: Levels and Trends

Philip DeCicca
Ball State University
Donald Kenkel
Cornell University
Michael F. Lovenheim
Cornell University


In this study, we develop new comprehensive estimates of the levels and trends in cigarette tax
avoidance in the U.S. and Canada from 2003 through 2015. We combine multiple sources of
data: administrative data on tax-paid cigarette sales from the Tax Burden on Tobacco; survey
data on self-reported cigarette consumption and tax avoidance from the Tobacco Use
Supplements to the CPS; scanner data on cigarette purchases from the Nielsen scanner data; and
data from discarded pack studies. We match these data to a newly created data archive of
applicable tax rates by State, Province, locality, and Native American/ First Nations reservation
in the U.S. and Canada. Previous research uses four methods to estimate tax avoidance: (i)
econometric models of the response of tax-paid sales to tax increases as a function of the
distance to lower-tax jurisdictions; (ii) comparisons of tax-paid sales to self-reported
consumption; (iii) self-reported measures of tax avoidance; and (iv) examination of tax stamps
on discarded cigarette packs. Our approach is to combine the four methods in an
econometric model that allows for different sources of measurement error in the self-reported
measures of cigarette consumption and tax avoidance and in the discarded pack data.
We use the results of our econometric model to estimate the current levels and recent trends in
cigarette tax avoidance in the U.S. and Canada. We also use the results to simulate alternative tax
policy approaches. We compare scenarios with tax hikes that preserve existing tax differentials
between states or provinces to a scenario with a harder-to- avoid national tax hike. We also
discuss the implications of the estimates for the possible development of illicit cigarette markets
in response to tobacco product regulations, such as a ban on menthol cigarettes.

Regressive Sin Taxes, with an Application to the Optimal Soda Tax

Hunt Allcott
New York University and Microsoft Research
Benjamin Lockwood
University of Pennsylvania
Dmitry Taubinsky
University of California-Berkeley


A common objection to “sin taxes”—corrective taxes on goods that are thought to be over-consumed, such as cigarettes, alcohol, and sugary drinks—is that they often fall disproportionately on low-income consumers. This paper studies the interaction between corrective and redistributive motives in a general optimal taxation framework. We show that the implications of regressivity hinge on why consumption decreases with income. If the consumption-income relationship is driven by income effects, then regressivity is optimally offset by targeted transfers or income tax reforms, not by moderating the level of the sin tax. If the relationship is instead driven by between-income preference heterogeneity, the optimal sin tax depends on the demand elasticity: if demand is more elastic, then progressive benefits from reduced over-consumption can make the optimal sin tax larger than if there were no distributional concerns, while if demand is less elastic, the optimal tax is reduced. As an application, we estimate the optimal nationwide tax on sugar-sweetened beverages, using Nielsen Homescan data and a specially designed survey measuring nutrition knowledge and self-control. Simulations suggest that current city-level taxes in Berkeley, Philadelphia, and elsewhere are actually lower than the social optimum.

Tax Design in the Alcohol Market

Rachel Griffith
University of Manchester
Martin O'Connell
Institute for Fiscal Studies
Kate Smith
Institute for Fiscal Studies


We study optimal corrective taxation in the alcohol market. Consumption generates negative externalities that are non-linear in the total amount of alcohol consumed. If tastes for products are heterogeneous and correlated with marginal externalities, then varying tax rates on different products can lead to welfare gains. We study this problem in an optimal tax framework and empirically for the UK alcohol market. We find heavy drinkers have systematically different patterns of alcohol demands and that welfare gains from optimally varying rates are higher the more concentrated externalities are amongst heavy drinkers.
JEL Classifications
  • H2 - Taxation, Subsidies, and Revenue
  • I1 - Health