Alcohol advertising bans, consumption and control policies in seventeen OECD countries, 1975–2000

Abstract

This article uses cross-country panel data to study the effects of advertising bans and other control policies on alcohol demand. The null hypothesis is that advertising bans do not decrease alcohol consumption. The study addresses several shortcomings in four previous cross-country studies. First, an explanatory variable is included for other alcohol control policies. Second, the study examines the history of advertising bans in OECD countries. Third, the study also examines differences in cross-country trends that characterize developed countries, including aging of the population, increased tourism, higher unemployment rates and increased consumption of wine. The Mediterranean wine-drinking countries are shown to be categorically distinct from the beer-drinking countries and Nordic spirits-drinking countries. Fourth, the study examines the panel data for unit roots and employs model specifications that correct for nonstationary data. The empirical results indicate a significantly negative effect for the control index and the alcohol price. Using alternative model specifications and estimation methods, the results indicate that advertising bans do not reduce alcohol consumption.

Acknowledgments

I wish to thank an anonymous referee and the editor for helpful comments on an earlier draft. The usual disclaimer applies.

Notes

¹ For reactions to the conclusions on advertising and promotion in the WHO report, see, e.g., Jernigan (2003) and the response by Babor et al. (2003b).

² Article 15 of the EU's Television Without Frontiers Directive of 1989 (89/552/EEC) requires that TV advertising of alcohol shall not be aimed specifically at minors; link alcohol to physical performance or driving; link alcohol to social or sexual success; claim that alcohol has therapeutic qualities; encourage immoderate consumption; or place emphasis on high alcohol content as a positive quality.

³ For more on EU trade and alcohol policies, see Holder et al. (1998) and Osterberg and Karlsson (2003).

⁴ Consumption data are available for Germany and Japan, but these countries are excluded due to lack of adequate price data. Luxembourg is excluded due to the importance of cross-border purchases and tourism.

⁵ I also estimated regressions (1)–(3) in Table 6 with a common AR(1) term, which provides a different control for serial correlation. Neither advertising variable was significant in these regressions. Estimating regressions (4)–(6) with the country-specific time trends, rather than the year fixed-effects, did not alter the results.