Global Effects of Smoking, of Quitting, and of Taxing Tobacco

First, the risk is big. Large studies in the United Kingdom, the United States, Japan, and India have examined the eventual effects on mortality in populations of men and of women in which many began to smoke in early adult life and did not quit.^11–16 All these studies showed that in middle age (about 30 to 69 years of age), mortality among cigarette smokers was two to three times the mortality among otherwise similar persons who had never smoked, leading to a reduction in life span by an average of about 10 years (Figure 1). This average reduction combines zero loss for those not killed by tobacco with an average loss of well over a decade for those who are killed by it.

Second, many of those killed are still in middle age, losing many years of life. Some of those killed in middle age might have died soon anyway, but others might have lived on for decades. On average, those killed in middle age by smoking lose about 20 years of life expectancy as compared with persons who have never smoked.¹

Third, stopping smoking works. Those who have smoked cigarettes since early adulthood but stop at 30, 40, or 50 years of age gain about 10, 9, and 6 years of life expectancy, respectively, as compared with those who continue smoking.

Tobacco is the biggest external cause of noncommunicable disease and is responsible for even more deaths than adiposity both in high-income countries such as the United States²⁰ and globally.²¹ The risks in middle age are much greater for smokers who started in early adulthood than for those who started later. This means that the ratio of mortality among smokers to that among persons who have never smoked is much more extreme now (Figure 1, and the 50-year trends shown in the Supplementary Appendix, available with the full text of this article at NEJM.org) than it was half a century earlier, when the epidemic of smoking-attributable deaths was at an earlier stage.^11–15

Cigarette smoking was uncommon throughout the world in 1900, but smoking rates increased substantially in many high-income countries during the first half of the 20th century, first among men and then, in some countries, among women.²² By 1950 in the United States and the United Kingdom, substantial numbers not only of men but also of women smoked, and rates of lung cancer were increasing steeply, particularly among men.¹ In 1950, major studies in both countries^23,24 showed that smoking was a cause of most deaths from lung cancer, and subsequent reports showed that smoking caused even more deaths from other diseases than from lung cancer.^25,26

After 1950, cigarette consumption continued to rise for some decades in high-income countries, and it has risen among men (though generally not among women) in many low- and middle-income countries. Although there has been widespread cessation in many high-income countries (in some, consumption per adult has been halved since the 1970s),²² about 1.3 billion people worldwide now smoke, most in low- and middle-income countries where cessation is uncommon.⁴ Two thirds of all smokers live (in descending order of numbers of smokers) in China, India, the European Union (in which central tobacco legislation can influence 28 countries), Indonesia, the United States, Russia, Japan, Brazil, Bangladesh, and Pakistan (Table 1).^27,28 In India, manufactured cigarettes are now displacing bidis (locally manufactured small cigarettes).²⁹ Cigarette consumption in China continues to rise steeply and now accounts for more than 2 trillion of a worldwide total of about 6 trillion cigarettes smoked per year.³⁰ A useful approximation suggested by studies in high-income countries is that 1 ton of tobacco yields about 1 million cigarettes and causes about 1 death, so just 1 trillion cigarettes consumed a year will eventually cause about 1 million deaths a year.

One reason why the mid-century evidence of hazard was not at first taken seriously, even in countries where it was generated, is the delay of about half a century between widespread adoption of smoking by young adults and the main effect on mortality in later life.^1–3 Among all U.S. adults, for example, cigarette consumption averaged 1, 4, and 10 per day in 1910, 1930, and 1950, respectively, after which it stabilized. The long-delayed result of this increase in consumption during the first half of the century was seen only in the second half of the century; tobacco caused about 12% of all U.S. deaths in middle age in 1950 but about 33% of such deaths in 1990.¹ A similar pattern was seen about 40 years later among Chinese men, who consumed about 1, 4, and 10 cigarettes per day in 1952, 1972, and 1992, respectively. In 1990, tobacco caused about 12% of all deaths among middle-aged Chinese men, and it could well cause about 33% in 2030.^31,32 (Tobacco causes few deaths in Chinese women, because less than 1% of Chinese women born in each decade since 1950 smoke.^27,31)

Because men started smoking before women, the effects in middle-aged men are now apparent in most high-income countries. The full eventual effects of persistent smoking in women, however, can be assessed directly in only a few countries (e.g., the United States and the United Kingdom) and only in the present (21st) century. The ratio of mortality from lung cancer among U.S. women who currently smoke to the (constant) mortality among women who have never smoked has increased greatly during the past half-century: it was only 3 in the 1960s, but it was 13 in the 1980s and 26 (similar to that among men) in the 2000s.¹⁴ The reason for the jump from a ratio of 3 to a ratio of 26 is that in the 2000s many U.S. women in their 60s who were smokers had smoked ever since early adulthood, whereas in the 1960s few women in their 60s who were smokers had done so.

Even though mortality from lung cancer among U.S. women was still low in the 1960s, women who were then in their 20s and who continued to smoke without quitting faced substantial hazards 40 years later.^13,14 Similarly, among men in low- and middle-income countries where many smoke but the death rates in middle age from smoking are not yet substantial, a full decade of life expectancy will eventually be lost by young adults who continue to smoke. Tobacco already accounts for about 12 to 25% of deaths among men in low- and middle-income countries such as China,^31,32 India,^16–18 Bangladesh,³³ and South Africa³⁴; given current smoking patterns, these proportions are likely to increase. Worldwide, about half a billion of the children and adults younger than 35 years of age already smoke or will do so if current uptake rates persist, and given current cessation patterns, relatively few will quit.²⁷ In all countries, young adults who smoke face about a decade of life lost if they continue and hence have much to gain by stopping.

Whereas tobacco-attributable mortality increases slowly after the uptake of smoking, the effects of cessation emerge more rapidly.^11–15 Persons who began smoking in early adulthood but stopped before 40 years of age avoid more than 90% of the excess risk during their next few decades of life, as compared with those who continue to smoke, and even those who stop at 50 years of age avoid more than half the excess risk, although substantial hazards persist (Figure 2).^11–15

The ratio of former smokers to current smokers in middle age is a useful measure of the success of tobacco control. Among persons 45 to 64 years of age in the European Union and the United States, there are now about as many former smokers as current smokers^28,35; by contrast, in most low- and middle-income countries (with the notable exception of Brazil), there are far fewer former smokers than current smokers (Table 1). Cessation is the only practicable way to avoid a substantial proportion of tobacco-attributable deaths before 2050, because a substantial reduction by 2025 in uptake by adolescents will have its main effect on mortality only after 2050.^2,3

Comprehensive tobacco-control programs using several price and nonprice interventions can substantially raise smoking-cessation rates and decrease initiation of smoking.⁴ Uruguay implemented most of the FCTC provisions and reduced consumption more rapidly than otherwise similar Argentina, which implemented only a few of the provisions.³⁶ Large increases in specific excise taxes on tobacco are particularly important, because they can have a substantial and rapid effect on consumption.^6–9 Reviews of comprehensive control programs in various U.S. states^37,38 and other high-income areas³⁹ concur that higher prices account for much, but not all, of the decline in smoking.

Similarly, an International Agency for Research on Cancer review of more than 100 econometric studies confirmed that tobacco taxes and consumption are strongly inversely related.⁹ It concluded that a 50% increase in inflation-adjusted tobacco prices reduces consumption by about 20% in both high-income countries and low- and middle-income countries,^6–9 corresponding to a price elasticity (percent consumption change per 1% price change) of about −0.4. Hence, doubling inflation-adjusted prices should reduce consumption by about one third (in which case revenues would increase, because the effect of reduced demand would be outweighed by the extra revenue per pack). Some of the effect among adults is due to quitting (or not starting), and some is due to reduced consumption per smoker.⁹ Higher taxes are particularly effective in poorer or less educated groups^6–9,39 and help prevent young people who are experimenting with smoking from becoming regular smokers.⁴⁰

The two major types of tobacco tax are specific excise taxes (which, being based on quantity or weight, are difficult for the industry to manipulate) and ad valorem taxes (which are based on manufacturer-defined price and can be manipulated more easily). In many high-income countries, about 50 to 60% of the retail price of the most-sold brand is a specific excise tax on tobacco or some variation of it (as in the European Union), but in low- and middle-income countries, this proportion is typically only about 35 to 40% (Figure 3).^4,6 A low specific excise tax on tobacco is the main reason that cigarettes are about 70% cheaper (even after adjustment for purchasing power) in many low-income countries than in high-income countries. Moreover, rapid income growth in many low- and middle-income countries is making the lower-priced tobacco products more affordable⁴¹ and helping cigarettes to displace bidis in India.²⁹

A low reliance on specific excise taxes on tobacco by China,⁴² India,²⁹ Indonesia,⁴³ and most low- and middle-income countries^4,6 means that the prices of commonly sold cigarette brands vary greatly within each country (by a factor of more than 10 in China, as compared with a factor of only about 2 in the United Kingdom and the United States), and this continued availability of low-cost brands discourages smoking cessation. In contrast, high specific excise taxes on tobacco of all brands encourage cessation rather than switching (by narrowing the price gap between the most and least expensive cigarettes), are easier to administer than ad valorem taxes, and produce a steadier revenue stream.⁹ In many low- and middle-income countries, although specific excise taxes on tobacco account for less than half the total retail price of cigarettes, tripling them approximately doubles the retail price, partly by triggering smaller increases in other taxes (e.g., sales tax) and markup. In most high-income countries, specific excise taxes on tobacco already account for more than half the retail price, so even just doubling them would approximately double prices.

The United States and the United Kingdom took more than 30 years to halve cigarette consumption per adult.²² With the use of large tax increases, however, France and South Africa halved consumption in less than 15 years (Figure 4).^3,44,45 From 1990 to 2005, France tripled inflation-adjusted cigarette prices by raising taxes 5% or more every year in excess of inflation, halved cigarette consumption, and doubled inflation-adjusted tobacco revenues. Today, the ratio of former smokers to current smokers in France comfortably exceeds the European average.^28,35 Over a similar period, South Africa also tripled the inflation-adjusted price of cigarettes, halved cigarette consumption, and doubled tobacco revenues.⁴⁵ Additional revenue can be used to fund tobacco-control programs or broader health efforts; much of the revenue from the 2009 U.S. taxation increase of 53 cents per pack of 20 cigarettes is allocated to expand children's health insurance.⁴⁶

Though tobacco advertising is banned throughout the European Union, China, and some other countries, cigarettes are still among the most heavily advertised and promoted products in the world, with spending on tobacco marketing reaching $8.6 billion annually in the United States alone.⁴⁷ In 2011 Australia, which had already banned advertising, introduced plain packaging for tobacco products, removing all brand imagery. The brand is printed only in small standard lettering below a pictorial warning. Recent evidence suggests that plain packaging increases cessation attempts.^48,49 New Zealand will introduce plain packaging in 2014, and the United Kingdom is considering it. Plain packaging goes beyond the prominent, rotating pictorial warning labels on tobacco products that have helped increase cessation attempts in Canada, Thailand, and elsewhere.⁵⁰ Pictorial warnings can reach even illiterate persons, and half the deaths from tobacco in India occur among the illiterate.²⁹

In the United States and the United Kingdom, bans on tobacco advertising on television coincided with the start of the long-term downturn in sales,⁵¹ although these partial bans on advertising allowed the industry to shift to other forms of advertising or promotion. More comprehensive bans on all direct and indirect advertising or promotion of any tobacco goods or trademarks further help to reduce consumption^52,53 and have the advantage of severing any dependence of the media on the tobacco industry. Bans on smoking in public places reduce nonsmokers' exposure to tobacco smoke and can also help decrease overall consumption,^54,55 as can mass-media campaigns.^51,56 In populations with many long-term smokers, low-cost epidemiologic studies of various types that monitor the changing extent to which tobacco is causing premature death help to raise political awareness of tobacco hazards and to provide information for the individual smoker.^1,16,33,34

Throughout the world, most former smokers managed to quit unaided, but physician support or telephone-based or Internet-based counseling and support can increase the likelihood of success.⁵⁷ In motivated persons, pharmacologic treatments or electronic cigarettes, or e-cigarettes, can also increase quit rates.^57,58 The eventual role of e-cigarettes remains uncertain, however, particularly if the tobacco industry controls the marketing of both traditional and e-cigarettes.

The WHO reports⁴ that although many countries now use nonprice interventions, only a few (including Mauritius, Mexico, the Philippines, Poland, and Turkey) have been using large increases in specific excise taxes on tobacco to reduce smoking.⁶ A large increase in inflation-adjusted price is, however, a key component of any realistic strategy to reduce smoking substantially during the 2010s or 2020s. The Bill and Melinda Gates Foundation, Bloomberg Philanthropies, the World Bank, and the Asian Development Bank are therefore providing technical advice for some ministries of finance to counter misleading tax advice from the tobacco industry.^29,42,43,59 Manufacturers' worldwide profits of about $50 billion in 2012⁶⁰ (approximately $10,000 per tobacco-attributable death) yield enormous political influence that is used, among other things, to try to prevent large tax increases.

Smuggling is a concern when tobacco taxes rise; about 10% of all cigarettes manufactured worldwide are already untaxed.⁶¹ Use of specific excise taxes on tobacco (rather than ad valorem taxes), stronger tax administration, and practicable controls on organized smuggling can, however, limit the problem.⁶² Even with some smuggling, large tax increases can substantially reduce consumption and increase revenue (Figure 4), especially if supported by better tax enforcement.⁶¹

Tripling inflation-adjusted specific excise taxes on tobacco would, in many low- and middle-income countries, approximately double the average price of cigarettes (and more than double prices of cheaper brands), which would reduce consumption by about a third and actually increase tobacco revenues by about a third. In countries in which the government owns most of the industry, as in China, the distinction between taxes and profit is fairly arbitrary, but doubling the average prices would still substantially reduce consumption and increase revenue. Worldwide, raising specific excise taxes on tobacco to double prices would raise about another $100 billion (in U.S. dollars) per year in tobacco revenues, in addition to the approximately $300 billion that the WHO estimates governments already collect on tobacco.⁴ Conversely, if a decrease in smoking by about a third were somehow achieved without increasing the inflation-adjusted price, tobacco tax revenues would decrease by about $100 billion.⁶

The main argument for reducing smoking is, however, the hundreds of millions of tobacco-related deaths if current smoking patterns persist. Indeed, in reviewing options to achieve a grand convergence by 2035 among the risks of premature death in low-, middle-, and high-income countries, the Lancet Commission on Investing in Health⁶³ recently identified a substantial increase in specific excise taxes on tobacco as the single most important intervention against noncommunicable diseases, as did the 2013 World Health Assembly.⁵ Losses or gains in tobacco revenue are of secondary importance; indeed, tobacco taxes are a small percentage of overall revenue in most countries (except China), and money not spent on tobacco is spent on other taxable goods or services.⁷ Attainment of the WHO target of a decrease of about a third in the prevalence of smoking by 2025, involving major decreases not only in high-income countries but also in populous low- and middle-income countries, would prevent several tens of millions of tobacco-attributable deaths during the next few decades^2,3,63 and about 200 million tobacco-attributable deaths during the century as a whole, mostly among people who are already alive, both by helping smokers to quit and by helping adolescents not to start.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

No potential conflict of interest relevant to this article was reported.

We thank Jillian Boreham and Hong-Chao Pan for the Supplementary Appendix on 50-year smoking-attributed mortality trends in the United Kingdom, United States, and Poland; Judith MacKay for comments and Cathy Harwood, Leslie Newcombe, and Joy Pader for editorial assistance on an earlier draft of the manuscript; Catherine Hill and Corne van Walbeek for French and South African data; and Ayda Yurekli for WHO tax-revenue estimates.