An economic theory of cigarette addiction

https://doi.org/10.1016/S0167-6296(98)00037-XGet rights and content

Abstract

In this paper we present a model in which individuals act in their own best interest, to explain many behaviors associated with cigarette addiction. There are two key features of the model. First, there is an explicit representation of the withdrawal effects experienced when smokers attempt to quit smoking. Second, there is explicit recognition that the negative effects of smoking generally appear late in an individual's life. Among the things we use the model to explain are: (1) how individuals can become trapped in their decision to smoke; (2) the conditions under which cold-turkey quitting and gradual quitting may occur; and (3) a reason for the existence of quit-smoking treatments.

Introduction

Economists model human behavior by assuming that individuals act in their own best interest. We describe agents as maximizing utility when they make rational choices consistent with their preferences, given their beliefs. Yet classes of widespread behaviors seem to prima facie contradict this assumption. Millions of consumers smoke cigarettes, overindulge in food and drink, pay taxes too soon, procrastinate, and save or exercise too little. To the extent that these choices are not what people want, they present a puzzle for standard economic theory.

An economic modeling answer to this puzzle is found in rational addiction models, which build on earlier analyses in which the individual's past consumption patterns affect present choices (see, for example, Pollak, 1970, and Stigler and Becker, 1977). In the addiction analysis of Becker and Murphy (1988) they assume that agents choose an overlife consumption path to maximize expected utility and show that addiction is consistent with rational behavior.

Another theory of addiction is the `multiple selves' approach that embodies competing preferences. In these models a smoker would have one `self' who wants to smoke competing against her non-smoker `other self'. (examples of such approaches, not necessarily focusing on addiction, include Schelling, 1984 and Thaler and Shefrin, 1981).

In the medical literature it is sometimes suggested that addictions are `diseases' arising out of biological predispositions. This approach places the choice about using addictive substances beyond the individual's control.1 Finally, psychologists sometimes describe addictions using `primrose path' theories. In this view individuals are lured into addiction either because the latent costs are initially hidden or because of a faulty reasoning process.2

In this paper we focus on one common addictive behavior—cigarette smoking. We present a model of cigarette consumption in which self-interested individuals choose to become addicted. The model incorporates many of the same elements from the Becker–Murphy rational addiction analysis. However, we reinterpret or re-characterize some of the effects.

For example, the Becker–Murphy analysis suggests that potential addiction arises whenever consumption of a good displays adjacent complementarity. This implies that someone is addicted to a good only when past consumption of the good raises the marginal utility of present consumption. Becker–Murphy note further that adjacent complementarity is closely related to the principle of reinforcement, which means that greater current consumption raises future consumption, and that reinforcement is a feature of many addictive goods. However, an additional implication of adjacent complementarity is a negative utility effect whenever an individual with a past consumption stock attempts to curtail present consumption. In our model we incorporate this negative utility effect explicitly and interpret it as a quitting or withdrawal cost. The presence of quitting costs is shown to be the critical element in making a good addictive.3

The assumption of quitting costs is consistent with the medical literature which indicates that smoking cessation imposes recognizable negative utility effects. Harris (1993), a practising physician and a health economist, provides a useful summary of the medical evidence of withdrawal effects in the book Deadly Choices. Harris argues that quitting smoking requires management of `nicotine withdrawal', a `well-defined condition'. Those trying to quit “become impatient or frustrated or anxious or nervous or restless…. They go through an experience that is nearly universal among millions of people every time they try to quit. They have the well-documented symptom of a recognized medical syndrome”. (Harris, pp. 163–164) Harris argues that, without some form of nicotine replacement, “virtually every regular smoker who abstains from cigarettes will go through his or her own version of nicotine withdrawal” (p. 165) In a broader sense, withdrawal occurs even if there is nicotine replacement, reflecting both deprivation (relative to actual smoking) in the dose and speed of delivery of nicotine, and behavioral/psychological deprivation.4

High relapse rates provide evidence that smokers who want to quit, have extreme difficulty doing so. Harris indicates that, of those who begin an attempt to quit smoking in week 0, 46% are still abstaining after 1 week, 25% after 4 weeks, and only 8% after 1 yr. Harris asserts that, “(w)hen it comes down to it, what caused my patients' symptoms and what precipitated their relapse was their inability to manage their nicotine withdrawal symptoms”.5 Prochaska et al. (1992), publishing in the psychology literature, note that “relapse is the rule rather than the exception with addictions”. (p. 1104). They cite an earlier study indicating that people who successfully stop smoking “make an average of three to four … attempts” before they succeed (p. 1104).

Perhaps the most essential aspect of addiction is the habitual behavior it generates. However, repetitive usage of a good is not sufficient to call consumption of that good an addiction. For example, regular consumption of a good such as breakfast cereal is rarely described as addictive even though it is consumed habitually. Instead, addiction requires that someone would like to either cease or reduce habitual consumption but is unable to do so. In this paper we show that explicit recognition of quitting costs can explain why people seemingly wish to do one thing but choose another.

Another feature in the Becker–Murphy model is the direct effect that the past stock of cigarette consumption has on present utility. If the potentially addictive goods are harmful, as with cigarettes, then current consumption will have a detrimental effect upon future utility and earnings. Becker–Murphy note that harmful goods imply a form of tolerance (another feature of many addictions) since higher past consumption lowers the present utility for a fixed consumption level. For cigarettes this negative effect also clearly relates to the detrimental health consequences that are likely to arise from consumption. Another feature of cigarette consumption, not explicitly incorporated into the Becker–Murphy framework, is that many of the detrimental effects, especially the development of various cancers, will occur in the distant future. In fact, for purposes of simplifying the exposition of our model, we will make the extreme assumption that the entire detrimental effect of smoking comes `at the end of life'. Subsequently, we will show that, with some modifications, our results hold even under the more realistic assumption that smoking affects the probability of death at every age.

In order to enhance the intuition of our results we make some simplifying assumptions about the nature of the choice problem faced by the individual. Whereas Becker–Murphy assume that the individual chooses a lifetime consumption path to maximize the present value of lifetime utility, we shall assume that the individual chooses only how much to consume today. While we do allow the individual to take account of the first-order effects of today's decision on future utility, we do not allow him to make the superhuman calculations that are necessary to form a fully consistent lifetime consumption path. As such our agent may be thought of as being `boundedly rational' in the sense of Herbert Simon (1978).6

In addition to the improvement in intuition, our model can explain behavior that, seemingly, cannot be explained using any other single approach. For example, this model can clearly show why utility maximizing individuals may be unhappy about being smokers, express a sincere desire to quit, yet find themselves unable to do so.7 We can also explain why merely growing older is sufficient to induce some smokers to quit. In addition, the model can generate cold-turkey quitting without relying on an exogenous shock or stressful event as in Becker–Murphy. The motivation to quit is based instead on changes in the smoker's perspective as he grows older. The model also shows that some individuals may quit smoking by gradually reducing consumption over time, a result not explicitly derived by Becker–Murphy. Finally, this model provides a simple explanation for `quit-smoking' treatments like Smokenders and nicorette gum used by smokers attempting to quit. In the rational addiction models it is not clear why smokers would ever seek out such treatments.

We begin in Section 2with a complete description of the model. In Section 3we use the model to explain the decision to begin smoking and to explain habitual usage. In Section 4we explain how both cold-turkey quitting and gradual quitting may arise. In Section 5we summarize the results of the alternative model developed in the appendix. In Section 6we consider the responsiveness of cigarette demand to exogenous shocks such as price changes. In Section 7we apply the model to explain behavior modification treatments and the potential effects of advertising. Finally we conclude in Section 8with a summary and suggestions for further research.

Section snippets

The model

We assume that the effects of smoking at age A can be decomposed into three additively separable components, current benefits (BA), future losses (LA), and an adjustment cost (CA).

Starting to smoke

For a young person whose friends do not smoke and who has little interest in, or understanding of, the physical effects of nicotine, the current benefits of smoking will be non-existent or unknown, so there appears to be no incentive to begin smoking. To generate such an incentive, consider the teenager whose friends begin lighting up cigarettes in a secluded location and encourage participation. Suddenly current benefits exist that did not exist previously. Alternatively, exposure to

Breaking the addiction to cigarettes

As a cigarette smoker becomes older, the future loss function, LA, which is a function of age, begins to change. To show how it changes we differentiate Eq. (3)with respect to age (A) to get,∂LA∂A=T′(A)+1WT(A)+A−αSAe−rT(A)−αSAT′(A)+1WT(A)+A−α(SA+s)e−rT(A)−α(SA+s)+T(A)+A−α(SA+s)T(A)+A−αSAre−r(t−A)Wtdt≥0.Future losses rise as one gets older because the discount factor used to weight end-of-life utility rises as aging draws one closer to the terminal date. The effect of the rising loss function

A `more realistic' set of assumptions

The modeling above used the idea that each cigarette reduces one's life by a fixed number of minutes. While this assumption is easy to understand, and facilitates a more intuitive diagrammatic exposition of a complex modeling process, it may not be the best way to assess the future health consequences of smoking. In particular, some deaths attributable to smoking occur when an individual is relatively young, such as when a smoker dies of a smoking-induced heart attack in his thirties.

This idea

Demand responsiveness to shocks

One area of controversy within the addiction literature involves the responsiveness of demand for cigarettes to changes in price or other exogenous shocks. For some who consider addiction an irrational behavior, policies such as higher consumption taxes or greater dissemination of health information will have little to no effect upon the demand for cigarettes. Empirical studies by Townsend (1987), Becker et al., 1991, Becker et al., 1994 and Chaloupka (1991), among others, show that cigarette

Applications of the model

Because this model explicitly incorporates quitting costs, it can explain the existence of behavior modification treatments such as Smokenders or nicorette gum. These treatments are chosen because it is hoped they will make quitting easier, that is to reduce quitting costs. Furthermore, these treatments may be experience-goods. If so, then a decision to try, say, nicorette gum, would initially be based on the smokers expectation of its likely effects. If the actual effects turn out less

Conclusion

This paper uses a model in which individuals act in their own self-interest to explain many seemingly irrational choices by cigarette smokers. Our model is similar to the Becker–Murphy rational addiction model in that past consumption of cigarettes influences current choices. However, we reinterpret or re-characterize a number of assumptions. Among these changes are the explicit modeling of quitting costs and incorporation of the fact that the effects of smoking are concentrated later in one's

Acknowledgements

We would like to thank Bryan Boulier, David Goldbaum, Chris Snyder, participants in seminars at American University and the University of Maryland, Baltimore County, several referees and the editor for helpful comments.

References (32)

  • F. Chaloupka et al.

    Price, tobacco control policies, and smoking among young adults

    Journal of Health Economics

    (1997)
  • E.M. Lewit et al.

    The potential for using excise taxes to reduce smoking

    Journal of Health Economics

    (1982)
  • G.C. Winston

    Addiction and backsliding: a theory of compulsive consumption

    Journal of Economic Behavior and Organization

    (1980)
  • J.J. Barendregt et al.

    The health care costs of smoking

    New England Journal of Medicine

    (1997)
  • G.S. Becker et al.

    A Theory of rational addiction

    Journal of Political Economy

    (1988)
  • G.S. Becker et al.

    Rational addiction and the effect of price on consumption

    American Economic Review

    (1991)
  • G.S. Becker et al.

    An empirical analysis of cigarette addiction

    American Economic Review

    (1994)
  • F. Chaloupka

    Rational addictive behavior and cigarette smoking

    Journal of Political Economy

    (1991)
  • Chaloupka, F., Grossman, M., 1996. Price, tobacco control policies, and youth smoking. Working Paper No. 5740. National...
  • Cross, J.G., Guyer, M.J., 1980. Social Traps. University of Michigan Press, Ann...
  • S. Douglas

    The duration of the smoking habit

    Economic Inquiry

    (1998)
  • D.S. Hamermesh et al.

    Does perception of life expectancy reflect health knowledge?

    American Journal of Public Health

    (1983)
  • Harris, J., 1993. Deadly Choices: Coping with Health Risks in Everyday Life. Basic Books, New...
  • Herrnstein, R.J., Prelec, D., 1992. A theory of addiction. In: Lowenstein, G., Elster, J. (Eds.), Choice Over Time....
  • T.A. Hodgson

    Cigarette smoking and lifetime medical expenditures

    Milbank Quarterly

    (1992)
  • Manning, W.G. et al., 1991. The Cost of Poor Health Habits. Harvard Univ. Press, Cambridge,...
  • Cited by (100)

    • Differential changes in quitting smoking by daily cigarette consumption and intention to quit after the introduction of a tobacco tax increase and pictorial cigarette pack warnings in Korea, 2013–2017

      2020, Drug and Alcohol Dependence
      Citation Excerpt :

      According to a review by Vangeli et al., intention to quit (ITQ) smoking consistently predicted attempts to quit, and dependence on tobacco was inversely related to successfully quitting smoking (Vangeli et al., 2011). The effects of tobacco control policies also varied depending on smokers’ smoking dependence and motivation to quit (Suranovic et al., 1999; Willemsen, 2005; Hammond et al., 2003; Ross et al., 2011; Park et al., 2015; Tabuchi et al., 2017; Nagelhout et al., 2013; Levy et al., 2007). The Korean government recently introduced two new national-level tobacco control policies sequentially.

    • Today's decisions, Tomorrow's outcomes: Does self-control explain the educational smoking gradient?

      2018, Social Science Research
      Citation Excerpt :

      The negative consequences of smoking are decades in the future and abstract; in contrast, the social and psychological rewards are immediate and concrete, and with further use and addiction, the immediate rewards become more numerous and intense. Even with no expectation of a long-term habit (Moffat and Johnson, 2001), with repeated use smoking becomes increasingly enjoyable compared to not smoking (Schelling, 1992), and the desire to smoke in the short term becomes a stronger behavioral incentive than poor health decades in the future (Becker and Murphy, 1988; Suranovic et al., 1999). In some cases, smoking becomes a social identity, which encourages continued use (Denscombe, 2001; Graham, 2012; Haines et al., 2009; Pampel, 2006).

    • Do consumers substitute opium for hashish? An economic analysis of simultaneous cannabinoid and opiate consumption in a legal regime

      2015, Drug and Alcohol Dependence
      Citation Excerpt :

      The analysis utilizes consumption and price data for opium, bhang and charas from 20th century British India. Basic economic models of addiction link the consumption of a drug in the present time with its past consumption (Stigler and Becker, 1977; Boyer, 1983; Becker and Murphy, 1988; Becker et al., 1991; Suranovic et al., 1999). The driver of addiction in the economic model is the role that past consumption plays in increasing the reward (i.e., the marginal utility) of consuming the drug in the present (Becker et al., 1991; Becker and Murphy, 1988).

    View all citing articles on Scopus
    View full text