Dr Meier provides an excellent discussion of future directions for research on alcohol marketing. She argues that there is a growing consensus in the literature which supports a link between advertising and alcohol consumption [1]. There are, however, a number of studies which provide no support for this link. For example, Nelson [2] reviews papers which find no effect of alcohol advertising on alcohol consumption. Also, a review by the National Institute on Alcohol Abuse and Alcoholism [3] concludes that research on the impact of alcohol advertising on youth consumption has been mixed. These reviews underscore Dr Meier's point that future research needs to employ more sophisticated methodologies and provide a more detailed examination of the advertising–consumption relationship.

As Dr Meier notes, one important new direction in alcohol advertising research is based on cognitive psychology and neuroscience research. Economists' contribution to this new direction is based on behavioral economics. Neuroscience research, which supports the assertions of behavioral economics, has resulted in an area referred to as Neuroeconomics. Bernheim & Rangel [4,5] present a neuroeconomic theory which is applicable to alcohol advertising and contains empirically testable hypotheses. The neuroeconomic model retains the empirical tractability of a conventional demand model but provides an advance in the understanding of the role played by advertising in alcohol consumption decisions.

The theory begins by explaining that individuals rely upon two key areas of the brain to make alcohol consumption decisions. The individual is assumed to have an addictive stock which reflects consumption of the alcohol over their life-time. One decision area of the brain, which reflects choices based on reasoned thinking, can be referred to as the rational mechanism (RM). When decisions are dominated by this mechanism the individual is defined as being in a ‘cold state’. In the cold state, consumption choices can be characterized by a standard demand model. The second decision area of the brain, which reflects choices based on cravings and impulse, can be referred to as the hedonic forecasting mechanism (HFM). When decisions are dominated by this mechanism the individual is defined as being in a ‘hot state’. The hot state is transient, but in the hot state the individual's consumption of an addictive good is based on the individuals addictive stock and other factors but is always greater than consumption in the cold state.

The switching mechanism between cold and hot states is a function of intensity of environmental cues such as advertising, the individual's addictive stock and other factors. The switch into the hot state is based on a prediction of hedonic pleasure rather than the consequence of actual consumption; that is, an alcohol craving induced by advertising. Neuroscience studies have found that individuals with a higher addictive stock are more likely to enter the hot state due to exposure to advertising. The theory predicts that individuals with a greater addictive stock will be more sensitive to advertising and as a corollary, less sensitive to rational information such as price. A phenomenon known as cognitive control occurs when the RM overrides a choice made by the HFM.

This theory predicts specifically that advertising increases overall demand rather than simply causing brand switching without increasing consumption. This theory also demonstrates that the response to advertising is a learned behavior. A learned behavior can be replaced with new learned behavior which is health-promoting. However, learning a new behavior requires asserting cognitive control, which can be more difficult to achieve in the presence of external cues related to the old behaviour; that is, a high level of alcohol advertising and other environmental cues makes it more difficult for individuals who would like to reduce their alcohol consumption.

Future studies based on behavioral economics might investigate consumption or purchases which occurred immediately following advertising exposure. Data which are not limited to youth and from field rather than experimental situations would be preferable. For example, data collected from individuals buying beer in sports stadiums with varying levels of beer advertising. The theory predicts that a greater number of heavy consumers will be in evidence when there is advertising as opposed to when there is no advertising. Another approach is to include measures of past alcohol consumption in a demand framework. In this case the theory predicts that individuals with a greater history of past consumption will be more sensitive to advertising and less sensitive to price.

Declaration of interest