The Effects of Television and Magazine Alcohol Advertising on Liquor Consumption

There are compelling reasons to study what causes people to drink and how much they drink. After smoking and obesity, alcohol abuse is the third leading cause of preventable US deaths. Further, many American youth drink. In 2014, among youth ages 12 to 20, 23 percent (9 million) reported having had a drink in the past month; 55.6 percent (5 million) of whom consumed 5 or more drinks per occasion [1]. Excessive consumption directly degrades health and imposes costs when people drink and engage in other behaviors. Alcohol is associated with a higher risk of traffic accidents and traffic fatalities of adults [2] and youth [3]; homicides and suicides [4]; accidents of other types; brain damage and neurodegeneration [5]; unprotected sex; transmission of sexual disease; injury; and sexual assaults [6]. The CDC estimates that, from 2006-2010, alcohol contributed to the deaths of roughly 88,000 people (4,200 youth) annually; with a total loss of 2.6 million years of potential life (30 years per alcohol-attributed death) [3]. Researchers estimate that US underage drinking leads to $61.9 billion additional costs annually; roughly $3 per drink [7]. Compelled by these facts, we investigate whether firms induce people to drink and drink more when they successfully expose people to alcohol advertising.

We exploit a quasi-natural experiment using relatively unique data used previously to study youth alcohol consumption and advertising [8]. The 1995-2009 cross-sectional surveys include detailed demographics, the amount and type of past 30-day alcohol consumption, and, most importantly, television programs watched and magazines read. With data on alcohol advertising appearing in each magazine or aired on each program, we estimate the advertising each individual saw.

Technically the study is challenging because it involves a dual-sided selection problem - firms choose when and where to advertise; people choose (particular) media progams/magazines. To mitigate potential selection bias, we exploit an unanticipated advertising policy change. In 1996 the liquor and spirits industry abandoned its 73 year self-imposed ban on television liquor advertising. Consequently, starting in the late 1990s, firms sharply increased television liquor advertising (and decreased magazine advertising).

This policy change allows us to (partially) resolve the dual selection problem because firms shifted advertising dollars to television for reasons uncorrelated with the unobserved factors that lead people to drink and because firms cannot target television audiences as precisely as they can readers of particular magazines.

We examine whether one drinks liquor and the number of liquor drinks consumed if one drinks liquor among people ages 18-20, 21-24, and 25 and older. We find that people of all ages were more likely to drink liquor if they saw more liquor advertisements in both magazines and television. But only the oldest age group increases the amount of liquor they drink when they saw more liquor advertisements on television or in magazines. We also show that liquor advertising exhibits diminishing marginal returns.

Our estimates show that exposure to liquor advertising causes people to consume more liquor. To our knowledge, we are the first to produce evidence of this type.