Advertising Of Alcohol
Similar to the tobacco industry, the alcohol industry in the US is highly concentrated (see Table 2). The US brewing industry, for instance, is dominated by three firms, which account for almost 80% of beer sales. Beer brewers spent approximately $975 million in 2007 on advertising, with the top three firms accounting for 72% of these expenditures. Total advertising and promotional spending for all alcohol companies are on the order of $4 billion (Jernigan and O’Hara, 2004). Advertising by the alcohol industry aims at raising sales through brand differentiation and customer loyalty, and advertising practices are self-regulated, primarily following a set of industry standards. For instance, industry guidelines allow alcohol-related ads to be placed in media where at least 70% of the audience is above the legal drinking age. Advertising messages also cannot directly appeal to under age youth. Some major broadcast networks adhere to a self-imposed ban on liquor advertising, though there are no such restrictions on cable networks.
The issues relating to the promotion and advertising of alcoholic beverages are similar to those discussed above with respect to tobacco, but with one exception. Unlike smoking, the majority of drinkers consume alcohol safely with little external harm. Thus, from a public health standpoint, the key debate with respect to market expansion has centered on problem drinking, which imposes considerable external costs (for instance, motor vehicle fatalities), and centered on the effects of advertising on youth drinking. On both of these fronts, although some studies have indicated that alcohol advertising is associated with more problem drinking and more underage drinking, the evidence is far from conclusive.
Anderson et al. (2009) reviewed 16 longitudinal studies that assessed adolescents’ exposure to media-based advertising and their drinking behavior. They concluded in favor of evidence suggesting that exposure to advertising messages is associated with a higher likelihood that the adolescent will initiate drinking, and associated with higher drinking among baseline drinkers. Many of these reviewed studies, however, are based on small, often nonrepresentative, samples and utilize measures of recalled exposure to ads, which may be potentially confounded with unobserved predisposition toward drinking or pro drinking sentiment.
Saffer and Dave (2006) utilized cross-sectional data from the Monitoring the Future Surveys and longitudinal data from the National Longitudinal Survey of Youth (1997 cohort), both nationally representative, to study the effects of probable advertising exposure on adolescent drinking behavior. They bypassed the problems associated with self-recalled advertising exposure and instead exploited variation across and within markets with respect to the level of alcohol advertising in broadcast and print media. Estimates indicate significantly positive but relatively small effects of media advertising on alcohol participation and binge drinking (elasticity estimates of approximately 0.09 and 0.17, respectively), though there is some heterogeneity in this response across gender and racial groups. The authors simulated the effects of a 28% reduction in total alcohol advertising (based on the range observed in their data) and concluded that the reduction in advertising could decrease adolescent binge drinking from 12% to approximately 10% and decrease monthly alcohol participation from 25% to approximately 23%.
Experimental studies have investigated how individuals’ drinking beliefs and behaviors respond to short-term advertising exposure in a controlled setting. Findings from this literature have been mixed. For instance, Lipsitz et al. (1993) alternately showed televised beer commercials, anti-drinking public service announcements, and soft-drink commercials to three groups of fifth and eighth-grade students. They did not find any significant differences in expectancies regarding drinking outcomes across any of the groups. Slater et al. (1997) examined the responses of high-school students to television beer advertisements embedded in sports or entertainment programs. They found that the responses were split along gender lines, with female students reacting more negatively to the beer advertisements than male students, especially when viewing sports content. The authors also found that white adolescents who responded favorably to the ads were more likely to report current drinking and future intentions to drink, though the effects were relatively small. It is difficult to disentangle causality in this study because favorable reaction to advertising may simply reflect the student’s underlying predisposition to drinking.
Saffer and Dave (2006) also reviewed prior econometric studies on the effects of alcohol advertising on alcohol consumption for the general adult population, according to the source of variation in the advertising measure (time-series, cross-sectional, panel, advertising bans). The vast majority of these studies did not show any substantial positive effects of advertising on overall alcohol consumption. The bulk of these studies though have utilized national time-series data, which often lack variation and confound effects with other unobserved trends. However, given that most individuals consume alcohol without imposing external costs, the more relevant question concerns whether alcohol advertising impacts problem drinking per se. Saffer (1991, 1997) provided indirect evidence on this issue. The study found that countries that ban broadcast alcohol advertisements have lower rates of traffic fatalities as well as alcohol consumption (Saffer, 1991). Saffer (1997) studied the effects of broadcast and outdoor advertising in 75 media markets on motor vehicle fatalities. It was found that a total ban on alcohol advertising could save as many as 5000–10 000 lives, implying an advertising elasticity of between 0.12 and 0.25.
Econometric studies find more consistent and stronger evidence of brand-switching effects in the alcohol industry. Fisher and Cook (1995), for instance, analyzed US data spanning 1970–90 and did not find any evidence that advertising impacts overall alcohol consumption. However, they did find that increased liquor advertising is associated with a reduced consumption of wine, suggesting cross-beverage market share effects. Nelson and Moran (1995) further found that advertising reallocates inter brand market shares, and to a smaller extent also inter beverage market shares, consistent with Fisher and Cook (1995).
Broadcast advertising in the alcohol industry generally aims at brand differentiation, whereas price-based advertising is more common in the print media, especially newspapers. There is some evidence that such price-based advertising leads to pro-competitive effects consistent with the informative view of advertising. Sass and Saurman (1995) indicated that large national brewers gain market share at the expense of smaller firms when states restrict advertising of retail prices. They found that the presence of restrictions on price advertising increased market concentration at the state level, both absolutely and relative to measures of national concentration. Additional restrictions on non-price advertising did not affect market concentration. Milyo and Waldfogel (1999) exploited the US Supreme Court ruling that overturned Rhode Island’s ban on price advertising of alcoholic beverages. Using Massachusetts as a control, they found that price-based advertising substantially reduced the price of the advertised good, though there was little effect on the price of the non-advertised good and no significant effect on price dispersion.
In summary, most evidence points to very weak or nonexistent advertising-induced market expansion effects in the alcohol industry. Several studies do not find strong positive effects of advertising on total alcohol consumption. There is some evidence that this overall nil effect may be masking salient effects for certain subpopulations. For instance, some studies have indicated that alcohol advertising increases indicators of problem drinking (for instance, motor vehicle fatalities) and drinking among adolescents, though in both of these cases the elasticity magnitudes are relatively small (and certainly smaller than estimated price responses). It should be noted that many of these econometric studies have estimated advertising effects conditional on price, which precludes one of the mechanisms through which advertising may impact primary or selective demand – that is, through changes in the retail price. This is especially relevant for price-based advertising. Tremblay and Okuyama (2001), for instance, made the point that if the elimination of advertising restrictions promotes price competition, then elimination of the self-imposed broadcast advertising ban in the liquor industry could cause alcohol consumption to rise even if advertising had no direct effect on market demand. There is more consistent evidence of advertising-induced brand-switching effects both at the brand level and the beverage level. This is in accord with the persuasive view of advertising, wherein the main role of advertising and promotion is to generate potentially spurious brand differentiation and enhance the brand’s monopolistic power. At the same time, there is also some indication from studies based on cross-state restrictions of price-based advertising that such advertising can lower retail prices and have pro-competitive effects.