Introduction
What do economists add to the multidisciplinary discussion of addiction? In this article, economic theories of addiction, statistical evidence produced by economists on addictive behaviors, and resulting policy implications are described.
The manner in which economists approach addictive behaviors differs in some ways from the approaches of other disciplines. Medical and public health research often views addiction as, by definition, maladaptive. Addicts passively submit to urges rather than actively make rational consumption decisions. Consumption of an addictive good is itself beyond the control of the individual. The National Institute on Drug Abuse uses the following definition:
Addiction is defined as a chronic, relapsing brain disease that is characterized by compulsive drug seeking and use, despite harmful consequences. It is considered a brain disease because drugs change the brain – they change its structure and how it works. These brain changes can be long lasting, and can lead to the harmful behaviors seen in people who abuse drugs.
By this definition, addiction is characterized by physiological changes and research often focuses on the neurological and psychological mechanisms underlying those changes (see Redish et al., 2008 for a cross-disciplinary review of addiction research). Alcohol and other drug addictions are found to cause physical changes in body functioning, such as reductions in functioning of neurotransmitter activity like dopamine, and these neurotransmitters are part of the brain’s reward system (Koob and Le Moal, 2008). These physiological changes are often observed in conjunction with, and indeed difficult to disentangle from, psychological changes such as increased depression and anxiety (Newlin, 2008).
Economists differ in generally focusing on models intended to reveal how social phenomena involving addictive behaviors emerge, which requires models suitable for investigating the manner in which addicts alter their behaviors as incentives change. By how much do smokers change their cigarette consumption if tobacco taxes increase, and over what time period? Do illicit drug addicts change their behavior as criminal penalties imposed on drug possession vary, and if so, how is the market for illicit drugs affected? What are the private and social costs of addictive behaviors? How are addictive behaviors related to income? Which policies tend to reduce harms to addicts and to nonaddicts? These sorts of questions are better addressed using a combination of abstract behavioral models combined with statistical evidence on addictive behaviors, prices, and other incentives than by detailed exploration of physiological mechanisms.
Following Becker and Murphy (1988), economists often use the following definition:
Addiction: A good or activity is addictive for a given person at a given time if an increase in the person’s consumption today causes an increase in consumption tomorrow, other things equal.
Loosely speaking, you are addicted to cigarettes in the economic sense if smoking more today causes you to smoke (or want to smoke) more tomorrow. Increased consumption of a nonaddictive good, however, does not cause you to want to consume more today if you happened to consume more of it yesterday; your desire to drink milk today is independent of your past milk consumption. Note that this notion of addiction does not require the addiction to operate through an action of the drug on the brain, although it is consistent with such an action. Nor does this definition require that the activity is maladaptive; a person may be addicted in the economic sense to, for example, health enhancing exercise. Finally, whether a given good or activity is addictive may vary across people and over time within a given person’s life.
The economic definition of addiction is a purely behavioral definition, as opposed to alternate conceptions involving physiological processes. Nonetheless, in Becker and Murphy’s canonical model, people exhibit reinforcement and tolerance, elements of alternate conceptions of addiction. Reinforcement here means that increasing consumption of an addictive good today increases the marginal value that is given to consumption of the addictive good tomorrow. Tolerance suggests that consuming a given amount of the addictive good today yields less utility when consumption of the addictive good yesterday was higher.
The implications of this apparently straightforward notion of addiction are surprisingly complex. In the next Section Perfectly Rational Addiction, the canonical addiction model in economics, the rational addiction model of Becker and Murphy (1988), is discussed. This model is highly stylized, imposing strong assumptions about preferences and information, but the model is able to mimic many aspects of addictive behavior, make predictions that are possibly surprising but verified by evidence, and provide a framework for empirical analysis of taxation and other policies intended to limit consumption of addictive goods. Research building on this framework to incorporate more realistic behavioral and information assumptions is considered in Sections Imperfectly Rational Models of Addiction and Irrational Models of Addiction. Following Cawley (2008) economic models are distinguished as falling into one of the three categories: models of perfect rationality, models of imperfect rationality, and models of irrationality. Finally, in Sections Empirical Evidence and Policy Implications of Addiction Perspectives the statistical evidence and policy implications stemming from this line of research are discussed.