13.1 Price elasticity of demand for tobacco products

Show / hide chapter menu

A fundamental building block of economic theory is the fact that increasing (or decreasing) the price of a commodity reduces (or increases) demand for that commodity. Price elasticity of demand refers to the extent to which use of a product falls or rises after increases or decreases in its price. If price elasticity of demand for a product were very low—that is, if it were inelastic—then demand would fall or rise only slightly in response to price changes. For instance, if price elasticity for a particular good were about –0.1, then demand for that good would fall by only 0.1% for every 1% increase in price. Demand would fall by 1% for a 10% increase in price, by 2% for a 20% price increase and so on. Demand for a good with high price elasticity would fall much more sharply in response to price increases. If price elasticity of demand for a good were about –1.0, then demand for that good would fall by 1% for every 1% increase in price. Demand would fall by 10% for a 10% price increase, 20% for a 20% price increase, 100% for a 100% price increase and so on.

While demand for tobacco products is not as elastic as demand for many other consumer products1 research has consistently demonstrated that increases in the price of tobacco products are followed by moderate falls in both the percentage of people smoking and the amount or number of tobacco products that remaining smokers consume.2 Because increases in tobacco taxes result in higher tobacco prices for everyone, the effect of even small resulting reductions in tobacco use can be very large across the whole population.

13.1.1 Theories of demand

Because many people are highly dependent on tobacco-delivered nicotine and persist in smoking despite recognising that tobacco products are dangerous, many early economists theorised that tobacco use was an irrational behaviour and not suitable for conventional economic analysis.3 Other economists simply ignored the addictive nature of tobacco products. More recent studies explicitly address the addictive nature of cigarette smoking. The economic models of addiction underlying such research can be divided into three groups:4 imperfectly rational models of addictive behaviour; models of myopic addictive behaviour; and models of rational addictive behaviour.

Imperfectly rational addiction models assume that the rational, far-sighted part of a person wants good health and a long life but that his efforts to quit are repeatedly undone by the 'wayward' part of his personality which quite simply 'adores' smoking. Proponents of imperfectly rational addiction models hypothesise that consumption will fall sharply in response to price increases, but will then drift back again with time.

The myopic addiction model assumes that addicted smokers are short-sighted. Myopic addiction theorists predict that factors such as price and income will affect uptake of smoking but that once addicted patterns of consumption are well-established, individuals tend to ignore or discount future costs (both monetary and health-related). They hypothesise that, while decreases in price will increase consumption and increases in price will reduce consumption, the effect of price increases will be much smaller than the effect of any price decreases.

The rational addiction model by contrast rejects the proposition that smokers behave myopically. It asserts that even addicted individuals do take into account future costs. The model assumes that addicted smokers make a rational choice, weighing up the pleasure of current smoking and the unpleasantness of withdrawal that comes with quitting on the one hand and the cost of current and continued smoking and the long-term health effects on the other. Different people will make different decisions depending on how much they value good health, how unpleasant they believe it will be to quit, and how much financial pressure they are under. Individuals also differ in the extent to which they prefer short-term over long-term benefits. Nevertheless the choice an individual makes will take all relevant factors into account and be a rational one. Proponents of the rational addiction model such as Becker and Murphy5 have demonstrated in support of the model that current consumption of an addictive good tends to be inversely related not only to the current price of the good but also to the past and predicted future prices.6, 7 The model also suggests that more educated and older people will be responsive to both new information and to price increases, and that less educated and younger people will be much less responsive to information about long-term effects and relatively more responsive to immediate changes in price.

The rational addiction model was popular among many theorists and researchers over the late 1980s and the 1990s, but recently has been criticised on several grounds. First, it implies that individuals have good foresight, a very accurate picture of what the future is going to be like. In reality however, it is evident some people give little thought to the future. While they may be able to recite some of the diseases caused by smoking, they do not fully appreciate the nature and extent of health risks and may not be able to accurately envisage what their life would be like if they became very ill or disabled due to smoking. Second, the model would predict that individuals rarely regret past decisions about consumption, a theory not borne out in interviews with current smokers, almost all of whom regret ever having started smoking.8 Third, critics argue that it might be that smokers do not actually choose future consumption. Rather by continuing to smoke they are choosing only current consumption: future consumption happens to them rather than being chosen by them.

13.1.2 Estimates of elasticity

The extent to which demand for a particular product responds to changes in price is an empirical question, the answer to which can be ascertained by measuring trends in consumption as prices and other relevant factors change.

Depending on the size of the price increase, reduced consumption of tobacco products following increases in tobacco taxes can be quite substantial. In 1999, a World Bank review concluded that, all else being equal, price rises of about 10% would on average reduce tobacco consumption by about 4% in developed countries and about 8% in developing countries.9, 10 In their more recent meta-analysis reviewing 86 studies published to the year 2001 which examined the price elasticity of demand for tobacco products, Gallet and List11 found a mean price elasticity of –0.48, meaning that, on average, a 10% increase in price will be followed by a decrease in consumption of 4.8%. Subsequent research has found similar levels of elasticity.12-17[1]

13.1.3 Various methods of measuring elasticity

Price elasticity can be calculated in many different ways, and various methods have different advantages and different problems—see Section 13.1.7.

Many studies examining the price elasticity of demand have used population-level data—that is, the total amount of tobacco purchased or taxed for the entire population. Some of these studies have analysed changes in taxable or reported sales of all tobacco products, and some have looked at cigarettes alone. Some have measured the weight of product sold, others the number of units. Some studies have compared consumption in the same jurisdiction with different prices over time (time-series analyses). Others have compared consumption in jurisdictions with different prices at the same points in time (cross-sectional studies).

Other studies make use of surveys conducted regularly among school students, adults or households. These might be surveys that ask about smoking habits or they might be surveys about spending on a whole range of household items.[2] A small number of studies have looked at consumption as reported by a panel of smokers interviewed (about their consumption and various factors which might affect it) before and after tax increases over time. In recent times, behavioural economists have examined the response of individuals not to real-life price increases, but rather to price increases simulated in behavioural experiments in laboratories.

Researchers have also used a variety of statistical models for specifying demand and for estimating elasticity.

In their comprehensive review, Gallet and List11 found that resulting estimates of price elasticity from all these sorts of studies were generally fairly similar and did not differ systematically depending on the design or method of analysis used. They did find, however, that studies that took into account the effects of smuggling found lower price elasticity (–0.36). Studies that measured responses to price changes in the short term tended to report lower elasticity than studies that reported long-run estimates (–0.40 compared with –0.44). Studies published more recently and in high-quality journals also tend to report slightly lower estimates, reflecting increasing sophistication in modelling and the greater number of factors also affecting consumption being taken into account in more recent research.11

13.1.4 Price sensitivity among various population groups

Studies collecting data from individuals allow researchers to assess how different population groups tend to respond to changes in price.18

Since young adults more commonly than older adults are occasional rather than regular smokers, it is theoretically possible that they may be less influenced by price changes. However teenage and younger smokers generally earn lower wages and are less dependent on tobacco, both of which would tend to make them more price sensitive.19 Early cigarette demand studies by Lewit and colleagues using individual-level data20, 21 found a strong inverse relationship between price sensitivity and age, with younger smokers reducing consumption much more than adults in response to the same price increases. This finding was thrown into doubt by an influential Rand study by Wasserman et al in 1991.22 Most recent studies however 23-26 have supported the earlier conclusions of Lewit et al. that younger smokers tend to be more price sensitive.

In the US Surgeon General's review of the literature published in 1994,27 estimates of elasticity of youth cigarette demand ranged from –0.9 to –1.5, (about three times the level of elasticity of adult cigarette demand). Across all the studies published up until 2001, Gallet and List also found greater responsiveness among younger people, with an average price elasticity of –1.43 for teenagers, –0.76 for young adults, and –0.32 for adults.11 At least one major study published subsequent to the Gallet and List review also supported the conclusion of higher price sensitivity among young people, though this included a somewhat lower estimate than either the Gallet and List or US Surgeon General reviews (between –0.67 and –1.02).28[3] A study published in April 2007 provides the most recent and possibly the best estimate to that date, of price elasticity among American high-school students. Using data from more than three quarters of a million high-school students who were surveyed in Youth Risk Behavior Surveys over the period 1993 to 200529 Carpenter and Cook found price elasticity of between –0.23 and –0.56. Importantly, their analysis controlled for anti-smoking sentiment, shown by DeCicca and his colleagues in 2006 to be an important factor in explaining changes in youth smoking in periods following tax increases.30

Studies have differed as to findings about the relative price sensitivity of women compared with that of men. Townsend et al (1994) found that women in the late 1980s and early 1990s in Britain were more sensitive than men to increases in tobacco prices. Chaloupka and Pacula31 however found that the prevalence elasticity[4] for young American men was almost twice as large as that for young women. Across all the studies published until 2001, Gallet and List11 found an average price sensitivity of –0.50 for men, and –0.34 for women.

Studies have also tended to show greater price sensitivity among lower socioeconomic groups. Low-income groups were more responsive to increases in cigarette prices in the United Kingdom in the 1980s and early 1990s.31-33 In the US, smoking rates of young black men are significantly more responsive to changes in price than are smoking rates among young white men.34 See Section 13.10 for further details.

Smokers in poorer nations also tend to be more sensitive to price17. However, Blecher and Walbeek conclude that, when changes in affordability rather than changes in price are taken into account, the level of sensitivity in developing countries may be similar to that observed in developed countries.35

13.1.5 Impact of price increases on uptake, continued consumption and quitting

Individual-level data also allow researchers to examine the impact of price increases on both smoking participation (that is, whether people smoke at all) and smoking intensity (that is, how much people smoke).

Early researchers assumed most of the effects of price on young people smoking were on participation, mostly by reducing smoking initiation. Among older people, price increases were believed to be more likely to affect both intensity and participation, the latter mainly through cessation rather than preventing uptake. While the results of studies examining these factors have varied,36, 30 those which have used the longest panels of data25, 26 have generally supported these assumptions.4

In their analysis of national data from 17,287 students in 200 schools throughout the US published in 2003, Ross and Chaloupka found that higher cigarette prices reduced probability of youth cigarette smoking and that there was also evidence of negative price effects on intensity of youth smoking.28 Based on their findings, for each 10% increase in average cigarette prices in each state, the percentage of high-school students smoking could be expected to fall by 3.5%. Among students who continue to smoke at least monthly, the number of cigarettes smoked each month could be expected to fall 1.38% to 1.99%. For every 10% increase in average cigarette prices specific to students' school locations and typical brand choices, the percentage of students smoking could be expected to fall by 4.92%. Among students who continue to smoke, the number of cigarettes smoked each month could be expected to fall by between 5.21% and 5.62%.

Looking at a group of young adults over time in the US, a study published by Tauras in 200437 found price elasticity for cessation of –0.35, suggesting that, in addition to discouraging initiation among children, a significant increase in cigarette taxes may also be one of the most effective ways open to government agencies to encourage cessation among young people who have already taken up smoking.

13.1.6 Price elasticity for tobacco products other than cigarettes

US,38 Canadian39 and Finnish40 research has concluded that an increase in the price of manufactured cigarettes can lead to an increase in sales of hand-rolled cigarettes and other tobacco products. Increases in the price of those products without simultaneous increases in the price of cigarettes are also followed by drops in consumption.40 In Australia, use of roll-your-own tobacco has increased in recent years following substantial increases in tax levels on all tobacco products—see Chapter 2 Section 2.5 . Increases in use of roll-your-own have been even more pronounced in countries where taxes on loose tobacco have not kept pace with taxes on cigarettes.41

13.1.7 Limitations of studies assessing price sensitivity

Both studies that use individual-level data and those that use aggregate data are subject to various limitations.

The reliability of individual-level data is very much dependant on adequate sampling, and recall and honesty by respondents about factors such as purchase of illicit tobacco products. Under-reporting of numbers of cigarettes (or amount of other tobacco products) smoked is also a significant problem.42

Using aggregate data on taxable product sales avoids these problems and ensures that researchers are looking at the impact of price changes on the entire market, however studies which use such data are subject to a number of different limitations.

First, without asking a representative sample of consumers about the prices of the products they are able to buy over time, it is sometimes difficult to determine exactly what the average or typical price increase is—see Section 13.3.2 for further details. Manufacturers use a variety of methods of cushioning consumers from the effect of the increases in taxes. It is by no means straightforward to assess the extent of price increases across a market where retailers engage in various degrees of discounting, 'specialling' and bulk-selling and where tax-free and illicit tobacco products are available.

Specifying changing quantities of tobacco products is similarly not straightforward. Studies that measure the weight of product may not sufficiently account for tobacco companies gradually reducing the taxable weight of cigarettes (without corresponding decreases in delivery of toxins).43 Studies that analyse changes in the number of products generally do not take into account any consumer shift to higher tar, possibly more toxic tobacco products.44

Finally, studies often do not adequately take into account all the less-easily measured factors that affect tobacco consumption, such as negative (unpaid) publicity about tobacco in the media and anti-smoking sentiment in the community. Often changes in such factors coincide with or precede tax increases, and the effects of such factors may multiply rather than simply add to the effect of price increases.[5]

Despite debates on the finer points of theory and despite all these methodological limitations, there is no doubt that price exerts a profound impact on tobacco consumption. Falls in consumption following large price increases are consistently much more rapid and more significant than falls following implementation of most other tobacco control policies, and the World Bank has stated that increasing tobacco taxes is the single most effective measure that governments can take to reduce health burdens.2 Increasing taxes on tobacco is a key plank in every major international tobacco control strategy.45-52