Alcohol consumption is the number one modifiable risk factor for disease and disability globally for persons ages 15 to 49 (Lim et al., 2012). Evidence is growing that alcohol advertising and promotion is associated with alcohol initiation, consumption, and health consequences (Anderson, de Bruijn, Angus, Gordon, & Hastings, 2009; Grenard, Dent, & Stacy, 2013; Smith & Foxcroft, 2009). Reducing exposure to alcohol marketing and promotion has been identified globally as a public health priority (World Health Organization, 2010). Even so, in most countries around the world alcohol companies have been granted the privilege to regulate their own advertising and marketing practices (EGTA The Association of Television and Radio Sales Houses, 2011; J. Evans et al., 2003; J. Evans & Kelly, 1999; J. Evans, Marcus, & Engle, 2008; J. M. Evans et al., 2014; Jones & Donovan, 2002). For the most part, these regulations are lax, permitting youth to be disproportionately exposed to alcohol advertising (D. H. Jernigan, J. Ostroff, & C. Ross, 2005). Only one alcohol company has stepped forward to implement more stringent alcohol advertising policies – Beam Global Spirits and Wine (Beam) now part of Suntory. In 2007 in the United States, Beam voluntarily adopted advertising policies that would reduce the proportion of youth in the audience for Beam brand advertisements. This study evaluates the performance of Beam against its unique alcohol advertising policies and estimates the impact of these more stringent policies on both youth exposure and Beam's realized advertising costs.

We begin with a review of the evidence of an association between exposure to alcohol marketing and drinking behavior. Two review studies have been published that summarize findings from 14 longitudinal studies of the association between alcohol advertising and drinking behaviors including drinking initiation, drinking frequency, and binge drinking (Anderson et al., 2009; Grenard et al., 2013; Smith & Foxcroft, 2009), usually defined as consuming five or more drinks in a short period of time (Miller, Naimi, Brewer, & Jones, 2007). Generally, these studies have found that exposure to alcohol marketing is associated with increased risk of drinking initiation, increased consumption, binge drinking, and negative health consequences(Anderson et al., 2009; Grenard et al., 2013; Smith & Foxcroft, 2009).

For the most part, the associations revealed in these studies have been small, leading critics to suggest that the measured effects must be attributable to some unmeasured confounding factor or to a miss-specification of the model (Nelson, 2010). However, advertisers learned many years ago from seminal research conducted by Anheuser-Busch that advertising effects are non-linear, with the greatest impact with the first few advertising exposures and diminishing effects thereafter (Ackoff & Emshoff, 1975a, 1975c; Wind & Sharp, 2009). Most public health studies of advertising effects have not considered these non-linear associations and have measured population samples where rates of exposure are extremely high – factors that may have contributed to findings of small effects (Collins, Ellickson, McCaffrey, & Hambarsoomians, 2007; Connolly, Casswell, Zhang, & Silva, 1994; Ellickson, Collins, Hambarsoomians, & McCaffrey, 2005; Fisher, Miles, Austin, Camargo, & Colditz, 2007; Grenard et al., 2013; Robinson, Chen, & Killen, 1998; Snyder, Milici, Slater, Sun, & Strizhakova, 2006). More recent studies that have examined populations with both high and low rates of exposure and have considered non-linear effects have found stronger associations (Ross, 2014; Ross et al., 2014). Thus the evidence of alcohol marketing's association with drinking behavior continues to grow and strengthen.

Across the globe, alcohol advertising is subject to different levels of regulation. In the World Health Organization's Global Status Report on Alcohol, a survey of 159 countries reported on marketing restrictions for different alcoholic beverages in different media (World Health Organization, 2014). The largest group of countries (39.6%) reported no legislative restrictions on alcohol advertising, leaving the alcohol industry to develop its own advertising self-regulatory guidelines (World Health Organization, 2014). Some countries have limited regulations, for example restricting alcohol advertising to programming after a certain time in the evening (World Health Organization, 2014). However, we have published a study showing that such time restrictions may actually increase adolescent exposure to alcohol advertising by increasing the number of alcohol ads during times when youth are more concentrated in the audience for the program (Ross, de Bruijn, & Jernigan, 2013).

Generally, self-regulatory guidelines for alcohol advertising focus on a maximum underage audience composition (Distilled Spirits Council of the United States, 2011a, 2011c; The Beer Institute, 2011; The Wine Institute, 2011). For example, guidelines may suggest that no alcohol ads be placed on programming when youth under age 18 make up 30% or more of the audience. However, in many countries, youth under age 18 make up considerably less than 30% of the population. For example, in Hungary where a 30% guideline is in place, youth under age 18 make up only 18% of the population and adolescents ages 12-17 make up only 6% of the population (European Commission, 2012). Thus, in Hungary youth may be disproportionately exposed to 30%/18%=1.67 or 30%/6%=5.0 times the advertising weight of other age groups without violating the self-regulatory guidelines.

In 2003 in the United States, trade associations for the beer and distilled spirits companies joined the wine industry in adopting guidelines to advertise only in media where the underage audience composition was below 30%. However, this standard permits alcohol companies to place ads in media in which the youth age 12-20 audience may be double its proportion in the population, causing significant overexposure of this high-risk group (D. Jernigan, J. Ostroff, & C. Ross, 2005). Thus, the Institute of Medicine (Bonnie & O'Connell, 2004) and 24 U.S. state and territorial attorneys general (Shurtleff et al., 2011) have called for more restrictive advertising placement standards.

In 2007, with support from 37 state Attorneys General, Beam Global Spirits and Wine (Beam), then the fourth largest spirits marketer in the U.S., adopted a stricter youth audience composition threshold for print, television, and radio advertising, which was in full force on January 1, 2008 (Beam Global Spirits and Wine, 2007). Under this standard, Beam agreed to place ads only in media where the underage audience composition was less than 25%, which was 5 percentage points lower than the industry standard at that time. Beam also committed that no more than 15% of its aggregate audience for all advertising would be under the legal drinking age (i.e., total youth exposure as a percent of total exposure for all ages <=15%), annually by brand and by medium.

In this study, we evaluated a) whether Beam was able to meet its goal of restricting alcohol advertising to programming where the underage audience composition was less than 25% on television and in magazines; b) whether Beam was able to meet its goal of achieving an average underage audience composition less than 15% per brand per medium per year on television and in magazines; c) whether Beam paid a higher advertising cost relative to other brands for its advertising on programs with more restrictive youth audience limits; and d) whether compliant Beam advertisements generated less youth exposure than matched advertisements from other spirits brands.

From 2008 to 2011, Beam's record with regard to compliance with its new advertising standards has been mixed. Generally, compliance has been better for advertisements placed in magazines than for those placed on television. Almost all Beam brands advertising in magazines met the 25% placement standard but Beam did not consistently achieve an annual aggregate 15% audience composition particularly in 2011. On television, 10 to 32 percent of youth exposure was generated by ads placed on programs with a youth audience composition above 25% and the 15% annual aggregate standard was met by a minority of brands advertising on television in most years. For those ads that did comply with the new standards, we estimate that Beam was able to reduce youth exposure to both its television and magazine ads compared to matched control advertisements, and we did not find any evidence that they paid higher advertising costs than other spirits brands.

Beam's use of two different compliance standards – the 25% maximum youth audience composition placement standard and the 15% annual aggregate composition standard – has some benefits. The 25% placement standard sets a threshold that works well in media such as magazines where audiences are measured annually and the youth composition is relatively stable. It is important to note, however, achievement of the 25% standard does not guarantee achievement of the 15% annual aggregate standard. Beam placed many advertisements in magazines that, for example, had youth audience composition in the range 20-25%. However, if the advertising plan included too many of these magazines, the annual aggregate 15% target was compromised, as occurred for many brands in 2011.

On the other hand, television audience data are measured for each advertising placement and have high variability, particularly for lower rated programs with smaller audiences. Therefore, we found that 10-32% of youth exposure to Beam television ads was generated from advertisements that did not meet the 25% placement standard. Television media present special challenges when planning for advertising purchases since audiences are measured for each advertisement and may vary considerably from one week to the next. However, a usual practice in many industries when attempting to meet performance goals in the presence of highly variable processes is to use a more conservative standard (i.e., a “guardbanded” standard) (Shector, 1992). Beam, and other alcohol companies, could use a 15% underage audience limit when placing ads on lower-rated (smaller audience) cable television programs to improve performance against their published 25% standard on television. In its most recent report on alcohol advertising (J. M. Evans et al., 2014), the Federal Trade Commission (FTC) endorsed the use of a more restrictive (i.e., lower youth audience composition) advertising placement standard when audience measures for a given medium have a high degree of variability.

In previous research, the authors of the current study estimated that youth exposure to alcohol advertising could be decreased by 14% if the Beam standard of 25% were adopted and effectively implemented by all alcohol advertisers (Center on Alcohol Marketing and Youth, 2007). The current study estimated that Beam's ads that adhered to the stricter standard reduced youth exposure by 10.3% relative to matched control ads from other spirits advertisers. The reduction in youth exposure in magazines was lower, possibly as a consequence of the more limited selection of publications with a young-adult audience profile that is desirable to alcohol advertisers.

Determining whether the reduction in youth exposure to advertising for compliant Beam brands has resulted in a reduction in youth drinking of Beam products is beyond the scope of the current study. However, we noted that in a recent national survey of youth alcohol consumption by brand, no Beam brands appeared in the list of 25 products most likely to have been consumed by underage drinkers (M. Siegel et al., 2013). Future research may attempt to assess the impact of reduced youth exposure to brand advertising on youth consumption of the advertised brands.

This study is limited in its ability to find an overall trend relative to the more restrictive placement standards adopted by Beam because of the variation in compliance with the standard across Beam brands, and because of year-to-year variation in alcohol advertising on a brand-specific basis. In addition, television audience data are not stable for many low-rated cable television programs, making it difficult for advertisers to plan their advertising purchases. However, the audience data used in this study are reportedly the same as those used by the alcohol industry to purchase and monitor its advertising performance. Our analysis was also limited to magazine and television media and did not assess compliance with the Beam standards for radio or digital advertising. Alcohol companies are increasing spending in digital advertising but television and magazines still represent the bulk of advertising spending (J. M. Evans et al., 2014). Finally, our cost analysis considered only the advertising costs associated with Beam's more restrictive advertising placement guidelines, and did not take into consideration other opportunity costs Beam may have incurred as a result of adopting this standard ahead of its competitors.

The findings presented in this paper support the recommendation of the National Research Council and the Institute of Medicine in 2003 that compliance with tighter standards can reduce youth exposure to alcohol advertising (Bonnie & O'Connell, 2004). Beam has demonstrated an effective strategy for reducing youth exposure to alcohol advertising that did not result in higher advertising prices. Implementation of stricter advertising standards could make a significant contribution to efforts to reduce youth alcohol consumption, and complement other effective strategies such as raising alcohol excise taxes and reducing alcohol outlet density (Centers for Disease Control and Prevention, 2012).