The Economics of Tobacco: Myths and Realities Tob Control: First Published As 10.1136/Tc.9.1.78 on 1 March 2000

78 Tobacco Control 2000;9:78–89

The economics of tobacco: myths and realities Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from

Kenneth E Warner

Introduction papers, insurance coverage, and transportation Each side in the debate about tobacco and shipping). Excise (and other) tobacco control—the tobacco industry and the public taxation generates many tens of billions of dol- health community—wields a seemingly power- lars annually.2 ful set of economic arguments, the industry Individual countries’ tobacco economies claiming that economic considerations urge a vary greatly. Nearly half of the world’s tobacco “go slow” approach to tobacco control, the farmers (an estimated 15 million) live in public health community insisting they recom- China, the world’s largest producer and mend an aggressive stance. Each of the most consumer of tobacco; 3.5 million reside in prominent arguments presented by both sides India. Other countries exhibit substantial has a kernel of truth to it; yet each, in its own tobacco sectors as well, including Zimbabwe, way, represents only a half truth. The industry Indonesia, Turkey, Bangladesh, Egypt, the uses its economic appeal with full knowledge of Philippines, and Thailand. These countries are and intent to exploit its ability to deceive and the exceptions to the rule, however. Tobacco mislead policy makers and the public. In farming constitutes a modest source of contrast, tobacco control advocates frequently employment in most countries and tobacco employ their economic rationale without full manufacturing employment constitutes well appreciation of its limitations. To inform both under 1% of total manufacturing employment tobacco control advocates and policy makers in most countries.1 Several countries derive more fully, this paper identifies the principal 10% or more of total government revenues economic myths concerning tobacco and from tobacco taxation. However, in most discusses their realities. The myths are countries, tobacco taxes generate only a few associated with their purveyors by initials: TI percent of total revenues.3 for Tobacco Industry myths, and TC for myths The public’s perception of the industry’s perpetuated by the Tobacco Control contribution to economic activity often community. http://tobaccocontrol.bmj.com/ exceeds the reality. For example, the economies of the six state “tobacco bloc” in the Myth 1 (TI) southeastern United States are perceived by Regardless of its health consequences, tobacco is many Americans to be heavily economically crucial to a nation’s (or region’s) economy. reliant on tobacco growing and manufacturing. Without the cultivation of tobacco, manufacture of tobacco products, and distribution and sale of prod- In fact, only 1.6% of jobs in these six states are associated with the core tobacco sectors of the ucts, a country’s economy will suVer devastating 4 economic consequences. Jobs will be lost, incomes economy. Almost half of the tobacco counties will fall, tax revenues will plummet, and trade in the US derive less than 1% of their income surpluses will veer dangerously in the direction of from tobacco farming, and the vast majority of deficits. tobacco farmers work oV their farms, most holding full time jobs elsewhere. Indeed, there REALITY on September 30, 2021 by guest. Protected copyright. This is the tobacco industry’s favourite is only a single farm dependent county in the economic myth. Conveyed to legislators and entire US that derives a majority of its farm cabinet ministers (and journalists), its intent is revenues from tobacco.56 According to a US to encourage the development of an Department of Agriculture economist, “to- indigenous tobacco industry within a given bacco plays a minor economic role in most country, or to discourage the adoption of local economies where it is grown”.6 tobacco control policies likely to decrease Although the importance of tobacco’s role is tobacco product consumption. often exaggerated in the public’s mind, there is Department of Health no denying that the number of people whose Management and The half truth in this myth resides in the fact Policy, School of that tobacco farming and product manufac- livelihoods depend on tobacco, at least in part, Public Health, ture, distribution, and sale do constitute is substantial, both globally and in selected University of significant economic activities in many countries, as noted above. Many of the Michigan, Ann Arbor, Michigan, USA economies, and in the world as a whole. estimates emanating from tobacco industry K E Warner Globally, according to tobacco industry public relations documents tend to inflate estimates, 33 million people farm tobacco, these numbers considerably, however, by Correspondence to: Dr Kenneth Warner, albeit many of them part time and most in including “expenditure induced employment” Department of Health addition to other crops.1 Approximately half in addition to core tobacco sector jobs and Management and Policy, School of Public Health, that number work in tobacco product those of industry suppliers. “Expenditure University of Michigan, 109 manufacture, distribution, and retailing. In induced employment” refers to jobs created in S Observatory, Ann Arbor MI 48109-2029, USA; addition, another 10 million or so are all sectors of the economy when tobacco work- [email protected] employed in supplier industries providing ers spend their incomes on other goods and Received 20 July 1999. materials and services to the tobacco industry services. Because this income “recycles” as it is Accepted 27 October 1999. (for example, harvesting tools and cigarette spent over and over again, the expenditure Economics of tobacco: myths and realities 79

induced employment eVect tends to dominate the transitional problems that have accompa- Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from estimates of total employment associated with nied many far more rapid shifts in industrial tobacco sales. Yet similar employment would activity, such as the decline in the steel and be created by workers from any industry shoe industries in the US. spending a similar amount of income. Certainly, there are a few countries that are The myth in the tobacco industry’s so dependent on tobacco that any substantial economic importance argument is that a declines in their tobacco industries would rep- significant economic presence necessarily resent genuine and important shocks to their implies significant economic dependence. economies, most notably Malawi (in which Implicit in the industry’s argument (occasion- tobacco accounts for 60% of total export earn- ally explicit) is the notion that a decline in ings) and Zimbabwe (23%).1 Note that the tobacco economic activity will entail a compa- problem for these countries lies not in the area rable decline in the economy of the country in of domestic tobacco control policy—declining question. In point of fact, when resources are domestic tobacco spending would not aVect no longer devoted (at all or as much) to a given the trade balance—but rather in the global economic activity, they do not simply demand for exported tobaccos. Thus, disappear into thin air—the implication of the campaigns to reduce domestic tobacco industry’s argument. Rather, they are consumption could benefit the health of the redirected to other economic functions. If a citizenry of these countries without signifi- person ceases to smoke, for example, the cantly damaging their economies.18 money that individual would have spent on For selected countries, such as China, Brazil, cigarettes does not evaporate. Rather, the and India, each with a large indigenous person spends it on something else. The new tobacco industry and each including a large spending will generate employment in other proportion of smokers, making the transition industries, just as the spending on cigarettes from a tobacco dependent to a tobacco generated employment in the tobacco industry. independent economy poses some special tran- Studies by non-industry economists in several sitional challenges. The issues diVer, too, countries have confirmed that reallocation of between those countries in which tobacco spending by consumers quitting smoking products are manufactured and distributed by would not reduce employment or otherwise government monopolies and those in which significantly damage the countries’ the industry is privatised. Recognising, economies.7–13 however, that even the most successful tobacco The industry understands this reality. control campaigns tend to reduce consump- Pressed by a journalist, a US Tobacco Institute tion gradually, the real issue for all such coun-

vice-president concurred that declining spend- tries is to reconcile the need to combat http://tobaccocontrol.bmj.com/ ing on tobacco would not necessarily mean smoking for public health purposes with the that overall economic activity would be need to ensure a smooth transition toward adversely aVected.14 The industry’s own alternative economic enterprises. economic consultants clearly perceive this Globally, concerns about the transitional truth, acknowledging it in their reports to the costs associated with declining tobacco industry.15 16 However, no industry public rela- consumption evaporate when one recognises tions document has ever mentioned that alter- that tobacco consumption is rising, not falling. native economic activity would replace declin- The World Health Organization predicts that ing tobacco activity. the number of tobacco consumers will increase The industry’s retort to dismissal of its prin- from 1.1 billion at present to 1.6 billion by the cipal economic argument is to appeal to year 2025.1 Thus, for the foreseeable future,

concerns about the high transition costs that any realistic conception of successful on September 30, 2021 by guest. Protected copyright. declining consumption might create, focusing international tobacco control must focus on on tobacco industry workers thrown out of reducing the rate of increase in tobacco use, jobs. A particularly poignant image is that of rather than producing substantial absolute poor tobacco farmers, their livelihood decreases in smoking. Globally, therefore, the constantly in jeopardy because of the public “transitional costs” of successful tobacco health assault on tobacco. Certainly, any rapid control will involve less rapid expansion of the decline in tobacco consumption could create tobacco industry, rather than dire economic transitional problems, for example brief straits attributable to its contraction. periods of unemployment for cigarette plant Elements of the industry’s economic manufacturing workers before they found new argument deserve the serious attention of jobs, some in the industries that expand country policy makers, especially the issue of because of the reallocation of consumers’ tax revenues. Declining cigarette smoking spending. However, the types of declines in frequently will mean declining revenues, as tobacco consumption witnessed in the major cigarettes are among the relatively few industrialised nations are so gradual that they commodities singled out for product speciﬁc create few transitional problems of any excise taxes. Any country now heavily reliant consequence. As economist Thomas Schelling on tobacco excise taxation and embarking on a has observed, the principal eVect of such dimi- national tobacco control campaign will need to nution in tobacco use is not that tobacco farm- revisit the nation’s tax policy. Note, however, ers will be thrown out of work, but rather that that taxes represent transfers of money from the children of tobacco farmers will be less citizens to their governments. They do not rep- likely to go into tobacco farming than were resent economic activity, in the sense of actual their parents.17 This stands in stark contrast to consumption of valuable resources. As such, 80 Warner

the issue confronting countries will be how to iting smoking in public places with the Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from fulfill revenue needs by redistributing the tax argument that such laws would penalise burden in an equitable manner, once tobacco aVected businesses, depriving them of smoking tax revenues decline. The crux of the issue will customers who would desert them, or increase not be a question of inflicting damage on the their costs of doing business. With regard to country’s economy. the latter, opponents claimed that workplace There is one important exception to this smoking restrictions would make it more diY- concern, examined in greater detail in cult to recruit and retain workers and cause consideration of industry myth 4 below; at least smoking workers to take more frequent and during the short and medium term, reducing longer breaks to replenish their nicotine tobacco consumption by raising tobacco excise depleted bodies. taxes will increase government revenues at the The workplace is invariably an important same time that it diminishes the disease burden battlefield for the tobacco industry, but associated with tobacco use.119 perhaps inevitably it represents a losing fight The bottom line on myth 1 is that the for the industry. In the US, evidence on the tobacco industry’s principal economic argu- health eVects of environmental tobacco smoke ment is not valid. With but the rare exception, (ETS)21–23 raised the prospect of potentially aggressive tobacco control campaigns can pro- ruinous lawsuits brought by non-smoking ceed full speed ahead without fear of damaging employees who claimed to have been sickened a country’s economy. Particularly if a country on the job by ETS. Further, in many settings, is a net importer of tobacco or cigarettes, the especially white collar workplaces, non- potential exists that modest short term smokers constituted the majority of employees economic gains will accompany the transition and they preferred a smoke free environment. to a populace less hooked on tobacco. Reduced The highly educated (and low smoking) spending on imports will permit more domes- leadership of companies often favoured restric- tic spending, with more employment thereby tions as a matter of principle. And evidence generated within the country. accumulated that banning (or severely restrict- Ironically, the tobacco industry itself is ing) smoking in the workplace actually paid frequently responsible for more lost jobs in a economic dividends: contrary to the industry given country’s domestic tobacco industry generated predictions, it often eased problems than the most successful of national tobacco associated with worker recruitment and reten- control campaigns. Industry induced job losses tion, and it reduced a myriad of other costs as derive from at least three sources: (1) mechani- well, including cleaning and maintenance of sation of cigarette production plants, in which equipment and facilities.24 20 technology supplants factory workers ; (2) In recent years, the eVects of laws banning http://tobaccocontrol.bmj.com/ purchase of imported tobaccos, replacing smoking in restaurants (and most recently bars domestically grown tobaccos raised by local in California) have been the most publicised farmers; and (3) wholesale price hikes that and controversial. In part this reflects the reduce cigarette sales, thereby diminishing highly politicised battles over the passage of industry employment while raising profits for such laws25 26; in part it resulted from the cigarette companies. publication of empirical studies that demonstrated the absence of the dire consequences Myth 2 (TI) predicted by organisations representing the Specific tobacco control policies will cause severe aVected businesses.27–29 economic hardship in specific non-tobacco The argument against the laws has been that industries. prohibiting smoking in restaurants in a given

REALITY jurisdiction (for example, a city) would cause on September 30, 2021 by guest. Protected copyright. A variant on myth 1, the second industry myth customers to flee to neighbouring jurisdictions is more targeted to both policy type and that had laws more hospitable to smokers, aVected sector of the economy. Rather than thereby damaging the local restaurant industry. arguing that policy induced reduction in Similarly, cities (or states) “burdened” with tobacco consumption will hurt an entire “draconian” laws banning smoking in economy, the industry claims that a specific restaurants would lose both domestic and policy, such as a ban on smoking in foreign tourists to more accommodating restaurants, will hurt a specific non-tobacco locales, producing declining payrolls and prof- industry or sector of the economy, such as the its, failing businesses, and unemployment. restaurant industry of the community covered The empirical evidence has not supported by the ban, or tourism within that community. the claim. In one study after another, covering Whereas myth 1 focuses on tobacco industry multiple states within the US, analysts have employment, myth 2 self consciously focuses found no adverse eVect of smoking on the economic implications of policy on restrictions, including complete bans, on local non-tobacco industries. This myth relies on the restaurants’ business. Indeed, several of the belief that non-tobacco businesses hurt by studies have found a tendency for smoking tobacco control policies will be seen by legisla- restrictions to increase business. Similar tors as bystanders in the war on smoking which findings derive from analysis of the eVects of become its undeserving casualties. smoking restrictions in bars (presumed the Myth 2 arose in the venue of workplaces and smoker’s sacred territory) and of the impacts of public accommodations in the US. The restaurant and bar restrictions on tourism.29 30 tobacco industry, and citizen groups and Restaurant and bar associations and industries working with it, battled laws prohib- smokers’ rights organisations have generated Economics of tobacco: myths and realities 81

numerous rebuttals to the research findings, (fewer) years of life. Critics of the tobacco con- Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from focusing on alleged methodological or data trol community’s economic argument cite flaws or the contrary results of their surveys of studies that indicate that the net costs of the opinions of owners of the aVected business smoking—the costs of treating smoking related establishments.30 Unlike the public health illness minus the additional expenditures on literature, however, these challenges have never non-smokers because they live longer—are themselves involved empirical analysis of small or non-existent.34 objective experience. Whether smoking adds to or subtracts from Another example of myth 2 concerns restric- aggregate medical expenditures remains a mat- tions on cigarette advertising. The tobacco ter of dispute. A series of studies has produced industry argues that specific ad restrictions will contradictory results, reflecting diVerences in have dire consequences for employment within the studies’ underlying assumptions, methods, a number of non-tobacco economic sectors. and data. The most recent published analysis35 When a ban on cigarette advertising was essentially supported the original conclusion of proposed in Hong Kong, for example, the Leu and Schaub36 that the net cost is eVectively Association of Accredited Advertising Agen- zero, but critiques of the new study have found cies issued an analysis by Coopers & Lybrand important flaws in it.37 In another prominent that estimated that from 1450–1600 workers in analysis, Hodgson concluded that smoking did non-tobacco industries would lose their jobs.31 add significantly to the net health care costs of Interestingly, only 50–100 of the job losses the US. He even found that net smoking would occur among “primary service related expenditures might be positive during providers”—media providers and advertising the years of senior citizenship (65 and older).38 agencies. The balance would be split between Hodgson’s seemingly contrarian finding “support industries” (a third) and, especially, resulted in part because he considered all of “community beneficiaries” (1000 jobs). In a smokers’ medical costs associated with US tobacco industry contracted study of the consumption of cigarettes, not merely those eVects of advertising restrictions proposed by associated with the principal smoking related the Food and Drug Administration, the diseases (cardiovascular disease, lung cancer, authors estimated substantial job losses, with and chronic obstructive pulmonary disease). printing and publishing firms “suVer(ing) the Studies that have found no significant impact most.” The authors acknowledged, however, on net costs have limited consideration to these that if these specific industries suVer job losses, major smoking related diseases.33 other industries will realise compensating job No such study ever compares the gains.32 expenditures to the number of people benefit-

ing from them. The entire reason that old age http://tobaccocontrol.bmj.com/ Myth 3 (TC) expenditures unrelated to smoking might Tobacco imposes an enormous financial burden on balance out earlier (and concurrent) smoking a country, greatly increasing health care costs to related expenditures is that smoking ensures treat smoking produced diseases and reducing that there are fewer people around to utilise productivity. health care services! If, instead of comparing REALITY aggregate expenditures, one asked how per This is the tobacco control community’s capita expenditures compare with and without favourite economic argument. Not only does smoking, one would likely find that per capita tobacco wreak havoc with physical health, the expenditures are considerably greater with argument runs, it damages a society’s fiscal smoking than without it. The net cost perspec- health as well. Thus, far from representing an tive also ignores how much people contribute

economic asset on society’s balance sheet, as to the health care system, through taxes and on September 30, 2021 by guest. Protected copyright. the industry suggests, tobacco constitutes a insurance premiums. Because a disproportion- sizable liability. ate number of smokers die young, their contri- Like the industry’s principal economic argu- butions to health care funding sources are ment, the validity of this argument rests on its smaller than those of longer lived non-smokers. interpretation and application. We begin with These caveats notwithstanding, it remains consideration of the health care expenditure true that once the full story is appreciated, to implications of smoking and later turn to the appeal to the high medical costs of smoking as issue of lost productivity. a fundamental reason to reduce smoking seems Most societies devote a significant pro- at least a bit disingenuous. Smoking may well portion of their health care resources to impose a financial burden on health care budg- treating people made ill by smoking and other ets, but its net impact is likely modest. use of tobacco products (at least 6–8% in the Further research on that burden is still US,33 for example). It is certainly reasonable clearly warranted. Such research should that a country should want to reduce smoking include the medical costs associated with produced disease so that it could devote these disease produced by ETS. Although the few resources to other health and social welfare estimates of such costs published to date have needs. generally been relatively small,39 no published It is also true, however, that non-smokers study has as yet considered the costs of ETS live longer than smokers, and thus that the produced heart disease, which growing health care costs of non-smokers during the evidence suggests may be by far the single most “extra” years of their lives (compared to smok- important source of ETS related mortality.23 40 ers) balance, at least to some extent, the higher Country specific research must also consider costs smokers experience during each of their how much of the financial expenditure for 82 Warner

smoking related medical care represents a bur- ated alongside the health care and productivity Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from den on the society and how much on individu- cost issues. However, care must be taken to als privately. Most of the literature on the distinguish smoking consequences that repre- medical costs of smoking does not distinguish sent real resource consumption, such as smok- between publicly borne and private costs. ing related medical care, from those such as Many economists are quick to point out, how- social security transfers that produce only dis- ever, that only the former technically constitute tributional consequences. true social costs.34 39 Although the public health Smoking may well impose a net economic perspective on social costs does not concur burden on a nation, but the magnitude of that with this view,41 it warrants serious fiscal concern pales in comparison with the consideration. Clearly, governments’ budget- enormity of the burden smoking exacts from ary interest in the medical costs of smoking will the public’s health. The tobacco control depend on the extent of their involvement in community must consider whether emphasis financing their citizens’ health care. In many on the economic argument demeans the true countries, particularly the poorer nations of the importance of the battle against smoking. world, government responsibility for treatment of chronic disease is modest at best. In numerous studies of the social costs of Myth 4 (TI) smoking, productivity losses attributable to A large tax increase is undesirable because it will smoking exceed medical expenditures. Produc- reduce government revenues by decreasing legal tivity losses result from smoking induced work cigarette sales. This will result from decreased absenteeism by employees sickened by smoking and increased smuggling of lower priced smoking, and from foregone productivity cigarettes from neighbouring countries. among workers who die prematurely because REALITY of smoking produced disease.42 Although this outcome is not impossible, it is The problem of appealing to these costs as extremely unlikely. Nearly all politically representing a social loss is that, as most conceivable tax increases will generate economists view the matter, the costs are increased revenues in virtually all countries. incurred privately, by the smokers and their The myth relies on two phenomena. First is families, as is the case for privately financed the matter of how revenue reductions resulting smoking related medical care. In other words, from reduced purchase of cigarettes will the fruits of the smokers’ labour represent pri- compare with revenue increases from the vate benefits to their families; conversely the higher unit tax rate. If the proportionate loss of smokers’ labour imposes costs on the decline in tobacco consumption exceeds the

smokers’ families, not on the society at large. proportionate increase in the tax rate, total rev- http://tobaccocontrol.bmj.com/ (This perspective is incomplete, since it fails to enues will fall. If the opposite holds, total consider that taxes on smokers’ earned income revenues will rise. represent beneﬁts of their productivity derived In industrialised nations, the consensus esti- by the larger society.) Of course, there are mate is that a 10% increase in cigarette price exceptions. If the premature demise of a will produce approximately a 4% decline in the smoker forces a previously self suYcient family quantity of cigarettes demanded by smokers. onto the dole, the burden of the loss of support More limited data from developing countries will be borne by the greater society through tax suggest that the impact may be in the order of funded health and welfare programs; but this twice as large.1 represents only a small fraction of the total Note, however, that tax constitutes only a smoking attributable productivity loss. fraction of total price. Thus, if tax constitutes

There are additional reasons that the half of the price of a pack of cigarettes, for on September 30, 2021 by guest. Protected copyright. productivity measurements typically overesti- example, a doubling of the tax rate would mate the loss, even if one construes these increase retail price by only 50%. As such, con- private costs to represent social losses. For one, sumers would have to be extraordinarily price the labour market adjusts to reflect the excess sensitive for consumption declines to exceed smoking related work loss. This is never figured the rate of increase in tax, far more so than into the calculations of productivity loss that found in any study to date. For all politically are found in the literature. feasible tax increases, revenue increases would If one’s objective is to tally the financial con- be expected in nearly every country in the sequences of smoking, one must also consider world, at least for some period of years. smoking’s implications for pension systems. In Empirical experience supports this the US, research has demonstrated that smok- expectation.19 ers subsidise non-smokers’ retirements. The second element of myth 4 relates to the Because smokers die earlier on average than phenomenon of cross border smuggling. non-smokers, they receive less in social security Clearly, inter-jurisdictional discrepancies in benefits, despite having made roughly the prices charged for a product, if suYciently comparable contributions to the system during large, create an incentive for people to buy the their lifetimes.43 In Great Britain, in contrast, product in the low price jurisdiction and trans- smokers’ greater disability leads to net pension port it for sale in the high price jurisdiction. payments to them.44 Thus, the implications of Particularly where borders are not closely smoking for disability and pension obligations guarded, substantial diVerences in cigarette of government will vary from country to coun- prices, often attributable primarily to try. In a full examination of the fiscal diVerences in tax rates, create conditions consequences of smoking, they must be evalu- conducive to cigarette smuggling. Economics of tobacco: myths and realities 83

Whenever governments contemplate to- Myth 5 (TI) Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from bacco tax increases, the tobacco industry and Even if a tax increase would raise government rev- its allies vigorously raise the spectre of a fiscal enues and decrease smoking, it is fundamentally and criminal disaster. In 1994 this argument unfair because its burden would fall disproportion- persuaded the Canadian government to cut its ately on the poor. Taxes should be proportional or taxes substantially in response to the progressive (that is, taking the same (proportional) perception of dramatic losses in cigarette excise or a larger proportion (progressive) of the income of tax revenues attributable to smuggling from the aZuent); they should not be regressive (taking the US.45 In 1998 Sweden likewise slashed a larger share of the income of the poor than of the taxes in response to the same concern, follow- wealthy). ing a 44% increase in Sweden’s already high REALITY taxes from December 1996 to August 1997.46 In most nations, more of the poor smoke than The states in the US invariably confront an do the rich. As a consequence, the poor often spend a substantially larger proportion of their armada of tobacco industry lobbyists making incomes on cigarettes than do the aZuent. the same argument each time an excise tax This means, in turn, that the poor bear a increase is under consideration. disproportionate share of the burden of a ciga- Evidence shows that the fears raised by the rette tax. Increasing the tax, it stands to reason, industry often greatly exaggerate the will increase this disproportionate burden even magnitude of the problem. In the Canadian more. (This will not always be true. case, federal and provincial government tax Particularly in poor countries, aZuent smokers revenues fell C$1.2 billion the year after the tax are likely to smoke many more cigarettes per cut, contrary to expectations raised by the day than do poor smokers. If the number of industry’s campaign, and smoking among chil- cigarettes smoked rises more rapidly than does 45 dren increased. Before its reduction, the income, the tax burden will fall proportionately Swedish tax increase had increased tax more heavily on the aZuent smokers. This revenues by 9% and decreased smoking among possibility will be enhanced if, as is often the both sexes and both young and older case, aZuent smokers consume premium, pos- smokers.46 sibly foreign brands of cigarettes, while poor Smuggling remains a legitimate concern, in people smoke much less expensive domestic part because the tobacco industry itself appears brands. If taxes are tied to price, rather than to tolerate and actively encourage it, as number of cigarettes, the burden will fall more indicated by recent court cases in which heavily on smokers of the high priced brands.) tobacco company executives have been found As this regressivity argument appears to be 46 47 guilty of complicity in smuggling operations. compelling, it is often the source of great con- http://tobaccocontrol.bmj.com/ The industry certainly benefits from increased cern to tobacco control advocates and sales associated with smuggling. Worldwide, legislators who genuinely want to do nearly a third of legally identified exports find something to decrease the burden of smoking. their way into the contraband market. As Fortunately, the argument is not necessarily Joossens and Raw have shown,46 48 however, true. Even if it is true, the feared regressivity of the determinants of smuggling are far more a tax increase will be far smaller than that fore- complicated than one might expect. They cast by budget analysts, with tools available to include countries’ general tolerance of corrup- legislators to minimise and even eliminate the tion and specific failure to police smuggling. In feared inequity. Europe, smuggling problems appear to be A given tobacco tax typically will be distrib- more serious in the low price countries; at least uted regressively. A tax increase, however, may the authors find no positive correlation not be regressive. This can result because the on September 30, 2021 by guest. Protected copyright. between price and the size of a country’s poor are typically considerably more smuggling problem. This does not negate the responsive to price changes than are the aZuent. In the United Kingdom, for example, importance of inter-jurisdictional price Townsend and colleagues found that the diVerences, but it does emphasise that many demand for cigarettes within social class I (the factors are at work, with price diVerences not most aZuent) was almost entirely unrespon- necessarily the dominant influence. sive to price increases, while it was quite Smuggling can be successfully combatted responsive within social class V (the poorest).49 through better and more complete record Thus, as the tax rate rises in the UK, few aZu- keeping, the use of prominent tax stamps, ent smokers are deterred from smoking and increased penalties for violation of the law, vig- hence their tax burden rises roughly orous enforcement of the law, the banning of proportionately with the tax rate. Among in-transit trade, and other supply smokers in social class V, in contrast, a tax rate 46 48 restrictions. More conscientious self increase is met with a decline in the number of policing by the cigarette manufacturers could cigarettes smoked of nearly the same substantially reduce smuggling. proportion. For the class as a whole, the tax The essential observation is that the threat of burden rises modestly, if at all. The burden of smuggling is systematically exaggerated by the the tax increase is thus experienced dispropor- tobacco industry to combat increased taxes tionately by the social class most able to aVord that will discourage purchase of its products. it. The author is aware of no documented In the US, Farrelly and colleagues found instances of tax revenues declining when tax that lower income persons had a price respon- rates were increased. siveness 70% greater than those with higher 84 Warner

incomes.50 Chaloupka found a similar relation- given their poverty, more such individuals Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from ship between education, which is correlated would be induced to avoid (or minimise) with income, and price response—the least smoking in the face of a sizable tax increase. educated were the most price responsive.51 Lacking more evidence, we cannot conclude with certainty that this relationship holds in Myth 6 (TC) general. Further, if in a given country the At the same time that health ministries urge their diVerences in price responsiveness across the citizens not to smoke, governments in many coun- income spectrum are smaller than those found tries subsidise tobacco growing. This is hypocritical, in the UK, the burden of a tax increase might and damaging to the health of the nation. By be distributed regressively; but again, the encouraging tobacco growing, the subsidy burden of the increase will be less regressive encourages smoking. than is the burden of the existing tax. REALITY Legislators can reduce concerns about ineq- This myth is likely predominantly true, uity by dedicating some portion of the although the reason diVers dramatically from revenues from the increased tax to assist low the direct link that tobacco control activists income smokers to quit. A sizable majority of perceive, at least in a country in which the low income smokers in developed countries “subsidy” system is similar to that in the US. report that they would like to quit. Given their Many American activists believe that financial circumstances, however, they may government support of tobacco growing have limited access to professional help and increases the supply of tobacco products. In pharmaceutical quitting aids. The mix of point of fact, the direct impact of the US greater price responsiveness among the poor system is exactly the opposite. and allocation of revenues toward smoking Given that diVerent countries’ approaches cessation in this group can make the net impact require diVerent interpretations, the issue must of the tax increase measure progressive. be addressed in the context of a specific In short, the potential regressivity of a approach to “subsidising” tobacco growing. To tobacco tax increase represents much less of a illustrate the analytical thinking one must concern than one might expect. That its use by devote to this matter, this discussion focuses the tobacco industry to fight such increases is exclusively on the system known best to the disingenuous is demonstrated by the fact that author, the US tobacco agriculture support the industry never expresses the same concern system. when raising its own wholesale prices. The The US system is a complicated mix of price economic burden on the poor who smoke is supports and output restrictions.19 52 Its two

identical. principal components are a system of http://tobaccocontrol.bmj.com/ Two additional issues warrant mention in allotments used to permit the growing of considering the eVects of tax increases. One is tobacco and to limit annual output to that the theory underlying optimal taxation predetermined quantities and price supports calls for a proportional or progressive tax that set minimum prices farmers receive for system. However, every individual component their crops. The allotments are licenses to grow of an overall tax system need not be tobacco, similar to liquor licenses required to proportional or progressive. Thus, even if sell alcoholic beverages to the public. a cigarette tax increase were regressive, it Developed in the 1930s during the Great might well be justiﬁed on other grounds, Depression, the allotments were originally dis- with other components of a country’s tax tributed to then existing tobacco farmers. system determining the overall degree of Without an allotment, owned or rented, a 40

progressivity. farmer cannot legally grow tobacco. This in on September 30, 2021 by guest. Protected copyright. One such ground constitutes the second itself restricts production. Production is issue warranting attention here: precisely further restricted by the US Department of because poorer smokers are more price Agriculture setting the maximum output per responsive than more aZuent smokers, a tax type of tobacco each year, the quantity increase will cause more poorer smokers to depending on anticipated purchases by the quit smoking. Especially given that in industr- major cigarette producers and stocks ialised nations poorer smokers suVer remaining from the previous year. disproportionately from the diseases caused by The price support system establishes a mini- smoking, a larger proportion of the eventual mum price per pound. If the price at auction health beneﬁts of quitting will accrue to the falls below this minimum, a cooperative buys low income population. In this regard, increas- unsold tobacco and holds it for sale in a subse- ing the cigarette tax is clearly a “progressive” quent year. public health policy. Rather than encourage cigarette sales and The issue of equity raises concerns in the consumption, the direct eVect of this system is world’s poorest countries even if the burden of to discourage smoking, by increasing the price a cigarette tax is distributed progressively (that of cigarettes. Supply restrictions increase the is, the rich pay proportionately more of it, price of tobacco, as do minimum support compared to income, than do the poor). In prices. Were tobacco grown and sold in a free these countries, although poor smokers market, supplies would be larger and prices consume fewer and less expensive cigarettes lower. As such, to the extent that cigarette than more aZuent smokers, the paucity of manufacturers use domestically produced their incomes makes any additional tax, regres- tobacco, the cost of the major raw ingredient in sive or otherwise, a genuine burden. Of course, their cigarettes will be higher than it would be Economics of tobacco: myths and realities 85

otherwise. All other things being equal, this increasingly aggressive invasion of poor countries Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from increases the price of the finished product. by the multinational companies. As it turns out, the impact is very small. REALITY Tobacco itself constitutes only 3% of the retail There is no doubt that the multinational price of cigarettes, and the allotment/price tobacco companies have moved aggressively support system is estimated to increase the into the world’s poorer nations in recent years, price of domestically grown tobacco by about often attempting to develop an indigenous 18–23%.53 Since the manufacturers of tobacco industry (tobacco farming and/or cigarettes in the US buy approximately a quar- finished product manufacturing), always trying ter to a third of their tobacco abroad, the net to expand tobacco product consumption. In impact of the tobacco “subsidy” system is to countries with a longstanding tradition of raise the price of cigarettes by no more than 1 smoking, the industry employs sophisticated cent per pack.52 Given an overall price elasticity marketing techniques to compete with domes- of demand for cigarettes in the neighbourhood tic brands of cigarettes, increase daily of −0.4,19 the direct eVect of the tobacco agri- consumption among existing smokers, and culture support system is to reduce the number encourage traditionally low smoking groups to of cigarettes smoked in the US by an estimated “modernise” by becoming smokers (for exam- 0.23%.52 ple, young women in many Asian societies).56 57 This outcome runs contrary to the According to the one empirical study that has conventional wisdom that the US tobacco sup- examined its impact, the introduction of mod- port system increases smoking. When one fac- ern western advertising in Asia increased ciga- tors in political considerations, however, it is rette consumption by about 10%.58 entirely plausible, even likely, that that is the Many tobacco control advocates believe that net impact of the existence of the system. One the multinationals’ move into developing of the principal consequences of the tobacco countries is a direct result of declining markets support system in the US has been to create a at home. Cigarette consumption is falling strong political constituency supporting gradually in the most industrialised nations tobacco. The political power of the tobacco (about 1.5% per year); it is increasing in low bloc in the US Congress is legendary, with leg- and middle income countries (about 2.1% per islators from the tobacco growing region in the year).56 The logic seems almost inescapable: as southeastern US fiercely devoted to protecting the profit balloon is squeezed in the developed the economic interests of tobacco farmers, the nations, the industry’s activity necessarily allotment owners (the true beneficiaries of the bulges out into the less exploited and hence tobacco agriculture regulatory system54), and, more promising developing country markets.

to a lesser extent, those of the cigarette manu- The corollary is that success in controlling http://tobaccocontrol.bmj.com/ facturers. An almost certain consequence of tobacco use in the First World exacerbates the this concentrated political power has been con- tobacco epidemic in Second and Third World gressional inaction on tobacco control policies. nations. Thus, policy and other tobacco In the absence of the tobacco bloc, it is quite control victories in western countries are possible that Congress would have adopted blemished by their inevitable negative much higher cigarette taxes, nationwide repercussions for the rest of the world. A cause restrictions on smoking in public places, and otherwise to be celebrated is converted, at least stronger restrictions on tobacco advertising in part, into a reason to feel guilt. This was a and promotion. Thus, the tobacco protective primary concern in the international political constituency created by the system community during the debate in the US on the almost certainly has had the net eVect of proposed comprehensive settlement of lawsuits 57

encouraging smoking, simply by severely limit- against the tobacco industry. on September 30, 2021 by guest. Protected copyright. ing tobacco control policy making.52 55 As such, Although the argument sounds compelling, in the context of the US system the tobacco there is no evidence to support it. Tobacco control community’s conclusion about the profits are not a zero sum game. Rather, they impact of the “subsidy” system is likely correct, are whatever the tobacco industry can make although the logic leading to that conclusion them, subject to the laws of supply and has been fallacious. demand and interventions of governments, It is imperative to reiterate that diVerent helpful or otherwise. No rational profit maxim- countries’ systems diVer suYciently from one ising firm would await diminished profit at another and that the analysis of their impacts home before venturing abroad into lands will vary, potentially substantially. Some promising new opportunities for profit. To the countries truly do subsidise the growth of contrary, all profit oriented firms will seek out tobacco or tobacco product manufacturing, a profits wherever and whenever they are policy that, other things being equal, does available. encourage smoking by increasing supply and Although there is a correlation between decreasing price. declining tobacco consumption in the west and industry expansion into the east, no evidence points to causation. A better explanation of the Myth 7 (TC) multinationals’ contemporaneous move into The tobacco companies have moved into other countries is the development of a fortui- developing countries in recent years in an eVort to tous set of economic circumstances: the compensate for the decline in markets in the world’s general easing of trade barriers for all interna- most aZuent nations. Ongoing tobacco control vic- tional commerce59 (a function of technological tories in rich countries will be paid for by and economic improvements in product distri- 86 Warner

bution and marketing, as well as the easing of Myth 8b (TC) Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from legal impediments); the emergence of a level of Cigarette advertising and promotion constitute one consumer income in such countries adequate of the principal determinants of smoking, especially to support consumption of western cigarettes; initiation of smoking by children. and the bulging treasuries of multinational tobacco companies that permit expansion REALITY overseas. The battle over cigarette marketing certainly Lost in the conventional wisdom is the fact stands as one of the major fronts in the tobacco that declining consumption in western nations wars. At issue are such diverse concerns as the is not always accompanied by declining profits. propriety of the advertising of an addictive In the US, over the past quarter century during substance to which children are inevitably which per capita consumption has fallen by exposed and constitutional protections of commercial free speech. At stake are more than a third,60 Philip Morris and RJ Rey- principles, the public’s health, and substantial nolds have frequently reported new record economic interests, in the advertising industry annual profits. Every time the companies as well as within the tobacco industry. increase their wholesale prices, they decrease The principal issue is whether or not adver- their sales but increase their immediate profits, tising and other forms of promotion aVect a reflection of the fact that the demand for smoking decisions: whether to try smoking; cigarettes in the US is inelastic. Thus, Philip how much to smoke; whether to quit; whether Morris and RJ Reynolds can raise their prices to remain abstinent once one has quit smoking. and increase profits at the same time that they Some of the more subtle mechanisms by which bemoan decreases in cigarettes smoked. (One advertising may increase smoking receive little might wonder why, if this is true, the public attention—for example, the impact of companies do not increase their prices all the constant visual cues to smoke (such as way to the point that profits are maximised. billboards) on the number of cigarettes existing The answer likely lies in the distinction smokers consume, or the eVect of the media’s between short run and long run profit maximi- dependence on cigarette advertising revenues sation for an addictive substance. Prices that on coverage of the hazards of smoking,62 and maximise profits in the short run may drive the impact, in turn, of reduced coverage on the enough new and potential customers out of the public’s appraisal of the severity of the risk market so as to decrease future (long run) posed by smoking.63 Rather, attention focuses profits. Interestingly, as tobacco control eVorts primarily (although not exclusively) on the succeed in reducing the prospects for the mar- earliest of smoking decisions: whether to begin.

keting of cigarettes in the future, the cigarette The tobacco industry and the tobacco http://tobaccocontrol.bmj.com/ companies may choose to raise prices more in control community take diametrically opposite the pursuit of short run profit maximisation.61) positions on this critical matter. The industry To put the general issue into perspective, the insists that advertising and promotion have no United States—one of the principal villains in impact on children’s decision to start smoking, this piece—includes about 4% of the world’s nor on anyone else’s for that matter. Many 1.1 billion tobacco consumers. A small decline tobacco control advocates believe fervently in US consumption, even were it to reduce that the seductive imagery of tobacco ads and profits, would have a barely discernible impact promotions plays a significant role in initiating on the number of tobacco consumers what are often life long nicotine addictions. worldwide (for example, a 10% decline in the One need dig only slightly below the surface number of US smokers would reduce the of the industry’s claim to conclude that it is world’s tobacco consuming population by likely quite disingenuous. Particularly in coun- on September 30, 2021 by guest. Protected copyright. tries in which only one or two companies con- about 0.4%). Even if tobacco consumption trol the vast majority of the market, advertising were expanding in the US, the modest would appear to be a futile gesture if its sole potential eVect on global industry profits function was to vie for slices of a pie of fixed would still mean that companies seeking future size.64 In the US, for example, in which Philip profit centres would necessarily look overseas. Morris controls half of the market (and RJ In a profit driven global marketplace, the mul- Reynolds half of the remainder), if Philip Mor- tinational tobacco companies intensify their ris succeeds in getting an existing smoker to pursuit of foreign markets independent of their switch brands, the smoker as likely as not successes or failures in First World countries. switches from one of the company’s own Tobacco control successes in developed brands to another. The profit potential in such nations thus are unadulterated successes. They an enterprise is obviously quite limited. One do not have negative ramifications for tobacco must wonder, therefore, why a company like use in other societies, and in fact are likely to Philip Morris would not jump at the opportu- have precisely the opposite eVect, by serving as nity to have cigarette advertising banned, models for national tobacco control. thereby saving billions of dollars annually without (according to the industry’s argument) losing sales. Indeed, the major companies likely Myth 8a (TI) would be the principal beneficiaries of an ad Cigarette advertising and promotion have no eVect ban, since established brands fare best when on the amount of smoking.Their only function,and advertising is not permitted. Yet Philip Morris, impact, is to distribute the number of cigarettes sold RJ Reynolds, and the rest of the American among the rival companies. industry have vigorously opposed multiple Economics of tobacco: myths and realities 87

opportunities to ban cigarette advertising “substantial, contributing, and therefore mate- Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from legislatively.64 rial factor” in children deciding to smoke. The The logic underlying the proposition that agency acknowledged, however, “that advertis- advertising and promotion play a significant ing may not be the most important factor.”74 role in kids’ decisions to start smoking is Advertising is an attractive target for tobacco substantial,65 enhanced by seemingly compel- control activists in part because of its visibility ling research.66 The empirical evidence relating and, of course, its symbolism. It is also marketing to smoking decisions is less substan- attractive because, unlike peer and parental tial and less compelling. The literature is behaviour, it is amenable to control through replete with analyses that show that children public policy. Expectations for the potential attracted to cigarette ads or promotional mate- impact of greatly reducing advertising and pro- rials are more likely to express interest in motion, or even eliminating them altogether, smoking or actually begin smoking must be kept in perspective, however. The col- subsequently.67–69 But no study has as yet dealt lective evidence suggests that an advertising adequately with the possibility that interest in and promotion ban would reduce smoking, smoking attracts children to cigarette ads, likely including both the initiation of smoking rather than the reverse. by children and the maintenance of smoking The econometric literature on the eVects of by adults, albeit modestly. A ban would be no advertising has focused on the relation between tobacco control panacea, however. Rather, it spending on advertising and promotion and would represent but a single piece in the jigsaw overall (hence primarily adult) cigarette puzzle of eVective tobacco control, one we consumption, rather than smoking by children continue to struggle to put together. per se. One group of studies finds no evidence of an impact of advertising on cigarette Conclusion consumption; a roughly equal number of stud- Interest in, even obsession with, the economics ies finds a significant, although quite small, of tobacco reflects the old adage that “money impact. Econometric approaches have distinct talks”. Focusing on the economic implications limitations in evaluating the fundamental of tobacco cultivation, product sales, tax 19 question, however. revenues and the like permits the tobacco Another approach to assessing the impact of industry to deflect attention from the domain advertising on consumption is to examine the in which it inevitably faces a humiliating, and outcomes of advertising and promotion bans. of course costly, defeat, namely public health. If bans reduce smoking, one can infer that It shifts tobacco control from the arena of advertising increases it. Scores of countries health policy to that of fiscal policy. That the have adopted partial or complete bans, permit- industry finds a receptive ear for its economic http://tobaccocontrol.bmj.com/ ting analysis of the association between arguments merely supports the wisdom of the adoption of restrictions and subsequent old adage. This bait-and-switch strategy has changes in the prevalence of smoking. Several forced the public health community to fight the years ago, a prominent review of the industry on the industry’s chosen battle- international evidence concluded that bans did ground. 70 reduce smoking. Industry supported critiques Both sides have honed seemingly compelling 71 of the analysis found fault with it, however. cases. Yet each side’s arguments survive only 72 Recently, SaVer and Chaloupka have because of their self evident grains of truth. developed compelling theoretical and empiri- Missing from the public debate, and from the cal evidence that complete bans can have a sig- knowledge base of those who make public nificant impact on smoking, while partial policy, is the fallacy that underlies each

restrictions have little impact, primarily argument, the reason that, as used, it is on September 30, 2021 by guest. Protected copyright. because the industry can find alternative misleading. Ironically, from the perspective of means of promoting their product that are the group of analysts who care most about nearly as eVective as those prohibited. Accord- monetary matters—economists—the eco- ing to these authors, compared to unrestricted nomic issues in tobacco are interesting but not advertising and promotion, a complete ban fundamentally important. The fundamentally would decrease cigarette consumption by important issue is the four million people killed approximately 6%. by tobacco each year, and the 10 million who In 1989, the US surgeon general concluded will follow annually three decades hence.1 If that the weight of the evidence strongly the tobacco control community can develop a favoured the proposition that advertising does sophisticated appreciation of the essence of 73 influence smoking. The surgeon general also tobacco economics, and convey that under- concluded, however, that the extent of standing to public decision makers, perhaps we advertising’s influence on smoking, and can force the issue of tobacco back where it especially on the initiation of nicotine properly belongs, in the domain of public addiction by children, was unknown and possi- health. bly unknowable. A wealth of evidence indicts other factors, including peer, parental, and role This work was prepared with the assistance of grant #026421 model behaviour, as potentially more from the Robert Wood Johnson Foundation of Princeton, New 67 Jersey. I am indebted to Frank Chaloupka, two anonymous important. In its unprecedented decision to referees, and journal senior editor Gary Giovino for helpful regulate the sale and distribution of cigarettes comments on a previous draft. and smokeless tobacco products to protect 1 World Bank. Curbing the epidemic: governments and the children, published in 1996, the US Food and economics of tobacco control. Washington, DC: The World Drug Administration indicted advertising as a Bank, 1999. 88 Warner Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from 2 Agro-Economic Services Ltd. and Tabacosmos Ltd. The 33 Warner KE, Hodgson TA, Carroll CE. The medical costs of employment, tax revenue and wealth that the tobacco industry smoking in the United States: estimates, their validity, and creates. London: Agro-Economic Services, 1987. their implications. Tobacco Control 1999;8:290–300. 3 Chapman S, Wong WL. Tobacco control in the third world: a 34 Viscusi WK. Cigarette taxation and the social consequences resource atlas. Penang, Malaysia: International Organiza- of smoking. In: Poterba JM, ed. Tax policy and the economy. tion of Consumers Unions, 1990. Cambridge, Massachusetts: Massachusetts Institute of 4 Price Waterhouse. The economic impact of the tobacco industry Technology Press, 1995:51–101. on the United States economy. Arlington, Virginia: Price 35 Barendregt JJ, Bonneux L, van der Maas PJ. The health care Waterhouse, 1992. costs of smoking. N Engl J Med 1997;337:1052–7. 5 Gale F. Tobacco dollars and jobs. Tobacco Situation & Out- 36 Leu RE, Schaub T. Does smoking increase medical look. US Department of Agriculture, Economic Research expenditures? Soc Sci Med 1983;17:1907–14. Service, TBS-239; September 1997:37–43. 37 Hodgson TA. The health care costs of smoking. N Engl J 6 Gale F. Economic structure of tobacco-growing regions. Med 1998;338:470. Tobacco Situation & Outlook. US Department of Agricul- 38 Hodgson TA. Cigarette smoking and lifetime medical ture, Economic Research Service, TBS-241; April expenditures. Milbank Q 1992;70:81–125. 1998:40–7. 39 Manning WG, Keeler EB, Newhouse JP, et al. The taxes of 7 Irvine IJ, Sims WA. Tobacco control legislation and resource sin: do smokers and drinkers pay their way? JAMA 1989; allocation eVects. Canadian Public Policy 1997;23:259–73. 261:1604–9. 8 Warner KE, Fulton GA. The economic implications of 40 Warner KE, Chaloupka FJ, Cook PJ, et al. Criteria for deter- tobacco product sales in a nontobacco state. JAMA 1994; mining an optimal cigarette tax. Tobacco Control 1995; 271:771–6. 4:380–6. 9 Warner KE, Fulton GA, Nicolas P, et al. Employment 41 Cook PJ. The social costs of drinking. In: Norwegian Minis- implications of declining tobacco product sales for the try of Health and Social AVairs. The negative social regional economies of the United States. JAMA 1996; consequences of alcohol use. Oslo: Norwegian Ministry of 275:1241–6. Health and Social AVairs, 1991:49–81. 10 Buck DC, Godfrey C, Raw M, et al. Tobacco and jobs. York, 42 Hodgson TA, Meiners MR. Cost-of-illness methodology: a UK: Society for the Study of Addiction and the Centre for guide to current practices and procedures. Milbank Memo- Health Economics, University of York, 1995. rial Fund Quarterly 1982;60:429–62. 11 McNicoll IH, Boyle S. Regional economic impact of a 43 Shoven JB, Sundberg JO, Bunker JP. The social security cost reduction of resident expenditure on cigarettes: a case of smoking. In: Wise DA, ed. The economics of ging. study of Glasgow. Applied Economics 1992;24:291–6. Chicago: University of Chicago Press, 1989:231–54. 12 van der Merwe R. Employment and output eVects for Bang- 44 Atkinson AB, Skegg JL. Anti-smoking publicity and the ladesh following a decline in tobacco consumption. Wash- demand for tobacco in the UK. The Manchester School of ington, DC: Population, Health and Nutrition Depart- Economic and Social Studies 1973;41:265–82. ment, The World Bank, 1998. 45 Sweanor DT. The smuggling of tobacco products. Ottawa, 13 van der Merwe R. An empirical analysis of the output eVects Ontario: Smoking and Health Foundation, 1997. of cigarette taxes in South Africa and the regional impact. 46 Joossens L. Impact of smuggling and cross border shopping on In: Lu R, Mackay J, Niu S, Peto R, eds. The growing world wide consumption. Washington, DC: Population, epidemic. Singapore: Springer-Verlag. In press. Health and Nutrition Department, The World Bank, 14 City Newspaper, Rochester, New York, October 16, 1986, p January 1999. 1, 6–8. 47 Dow Jones Newswires. RJR aYliate to pay $15M for acting as 15 Chase Econometrics. The economic impact of the tobacco smuggling front. December 22, 1998. industry on the United States economy in 1983. Bala Cynwyd, 48 Joossens L, Raw M. Cigarette smuggling in Europe: who Pennsylvania: Chase Econometrics, 1985. really benefits? Tobacco Control 1998;7:66–71. 16 American Economics Group. The US tobacco industry in 49 Townsend JL, Roderick P, Cooper J. Cigarette smoking by 1994: its economic impact on the states. Washington, DC: socioeconomic group, sex, and age: eVects of price, American Economics Group, March 1996. income, and health publicity. BMJ 1994;309:923–6. 17 Schelling TC. Economics and cigarettes. Prev Med 1986; 50 Farrelly MC, Bray JW, OYce on Smoking and Health. 15:549–60. Response to increases in cigarette prices by race/ethnicity, 18 Warner KE, Fulton GA. Importance of tobacco to a income, and age groups—United States, 1976–1993.

country’s economy: an appraisal of the tobacco industry’s MMWR Morb Mortal Wkly Rep 1998;47:605–9. http://tobaccocontrol.bmj.com/ economic argument. Tobacco Control 1995;4:180–3. 51 Chaloupka FJ. Rational addictive behavior and cigarette 19 Chaloupka FJ, Warner KE. The economics of smoking. In: smoking. Journal of Political Economy 1991;99:722–42. Newhouse J, Culyer A. Handbook of health economics. 52 Zhang P, Husten C. Impact of the tobacco price support Amsterdam: Elsevier. In press. program on tobacco control in the United States. Tobacco 20 Joossens L. Employment in the tobacco sector in the EEC. Control 1998;7:176–82. Europe for Smoking Prevention. Newsletter 6, p 6. Brussels, 53 Zhang P, Husten C, Giovino G. The impact of the price Belgium: Belgian Association for Smoking Prevention, support program on cigarette consumption in the United States. August/September 1989. Atlanta, Georgia: OYce on Smoking and Health, Centers 21 US Department of Health and Human Services. The health for Disease Control and Prevention, 1997. consequences of involuntary smoking. A report of the Surgeon 54 Babcock BA, Foster WE. Economic rents under supply General, 1986. Rockville, Maryland: Public Health Service, controls with marketable quota. American Journal of Agri- Centers for Disease Control, Center for Health Promotion cultural Economics 1992;74:630–7. and Education, OYce on Smoking and Health. Washing- 55 Warner KE. The tobacco subsidy: does it matter? JNatl ton DC: US Government Printing OYce, 1986. (DHHS Cancer Inst 1988;80:81–3. Publication No (CDC) 87–8398.) 56 Connolly GN. Worldwide expansion of transnational 22 US Environmental Protection Agency. Respiratory health tobacco industry. Journal of the National Cancer Institute eVects of passive smoking: lung cancer and other disorders. Monographs No 12. International Conference on Cancer

Report no. EPA/600/6–90/006F. Washington, DC: EPA, Prevention, 1992:29–35. on September 30, 2021 by guest. Protected copyright. OYce of Health and Environmental Assessment, OYce of 57 Hammond R. Addicted to profit: big tobacco’s expanding global Research and Development, 1992. reach. Washington, DC: Essential Action, 1998. 23 Glantz SA, Parmley WW. Passive smoking and 58 Chaloupka FJ, Laixuthai A. US trade policy and cigarette heart disease: mechanisms and risk. JAMA 1995; smoking in Asia. NBER Working Paper No 5543. 273:1046–53. Cambridge, Massachusetts: National Bureau of Economic 24 US Environmental Protection Agency. The costs and benefits Research, 1996. of smoking restrictions. An assessment of the smoke-free 59 Chaloupka F, Corbett M. Trade policy and tobacco: environment act of 1993 (HR. 3434). Washington, DC: towards an optimal policy mix. In: Abedian I, van der EPA, OYce of Air Radiation, Indoor Air Division, April Merwe R, Wilkins N, Jha P, eds. The economics of tobacco 1994. control: towards an optimal policy mix. Cape Town, South 25 Public interest pretenders. Consumer Reports May 1994; Africa: Applied Fiscal Research Centre, University of 316–20. Cape Town, 1998:129–45. 26 Traynor MP, Begay ME, Glantz SA. New tobacco industry 60 Tobacco Institute. The tax burden on tobacco. Historical com- strategy to prevent local tobacco control. JAMA 1993; pilation 1995. Vol 30. Washington: Tobacco Institute, 270:479–86. 1998. 27 Glantz SA, Smith LRA. The eVect of ordinances requiring 61 Becker GS, Grossman M, Murphy KM. An empirical smoke-free restaurants on restaurant sales. Am J Public analysis of cigarette addiction. American Economic Review Health 1994;84:1081–5. 1994;84:396–418. 28 Glantz SA, Smith LRA. The eVect of ordinances requiring 62 Warner KE, Goldenhar LM, McLaughlin CG. Cigarette smoke-free restaurants and bars on revenues: a follow-up. advertising and magazine coverage of the hazards of smok- Am J Public Health 1997;87:1687–93. ing: a statistical analysis. N Engl J Med 1992;326:305–9. 29 Glantz SA, Charlesworth A. Tourism and hotel revenues 63 Marsh A, Matheson J. Smoking attitudes and behavior. An before and after passage of smoke-free restaurant ordi- enquiry carried out on behalf of the Department of Health nances. JAMA 1999;281:1911–18. and Social Security, OYce of Population Census and Sur- 30 Alberta Tobacco Control Centre. The economic impact of veys. London: HMSO, 1983. smoke-free restaurant bylaws. Alberta, Canada, June 22, 64 Tye JB, Warner KE, Glantz SA. Tobacco advertising and 1999. consumption: evidence of a causal relationship. J Public 31 Coopers & Lybrand. A study of the economic impact of a ban Health Policy 1987;8:492–508. on cigarette advertising in Hong Kong. Hong Kong, Coopers 65 Warner KE. Selling smoke: cigarette advertising and public & Lybrand, 1996. health. Washington: American Public Health Association, 32 American Economics Group. Economic impact in the states of 1986. proposed FDA regulations regarding the advertising, labeling 66 Fischer PM, Schwartz MP, Richards JW Jr, et al. Brand logo and sale of tobacco products. Washington, DC: American recognition by children aged 3 to 6 years: Mickey Mouse Economics Group, March 1996. and Old Joe the camel. JAMA 1991;266:3145–8. Economics of tobacco: myths and realities 89 Tob Control: first published as 10.1136/tc.9.1.78 on 1 March 2000. Downloaded from 67 US Department of Health and Human Services. Preventing 71 Stewart MJ. Tobacco consumption and advertising restric- tobacco use among young people: a report of the Surgeon Gen- tions: a critique of Laugesen and Meads (1991). eral. US Department of Health and Human Services, Pub- International Journal of Advertising 1992;11:97–118. lic Health Service, Centers for Disease Control, National 72 SaVer H, Chaloupka FJ. Tobacco advertising: economic Center for Chronic Disease Prevention and Health theory and international evidence. Cambridge, Massachu- Promotion, OYce on Smoking and Health, Washington. setts: National Bureau of Economic Research, Working (US Government Printing OYce Publication No S/N Paper No 6958, February 1999. 017–001–00491–0.) 73 US Department of Health and Human Services. Reducing 68 Pierce JP, Choi WS, Gilpin EA, et al. Tobacco industry pro- the health consequences of smoking: 25 years of progress. A motion of cigarettes and adolescent smoking. JAMA report of the Surgeon General. Atlanta: US Department of 1998;279:511–5. Health and Human Services, Public Health Service, Cent- 69 Siegel M, Hingson R, Biener L. The influence of tobacco ers for Disease Control, National Center for Chronic Dis- marketing and counteradvertising on smoking initiation among ease Prevention and Health Promotion, OYce of Smoking youth. Presented at the annual meeting of the Robert Wood and Health, 1989. (DHHS Publication No (CDC) Johnson Foundation Substance Abuse Policy Research Pro- 89–8411.) gram. Charleston, South Carolina, December 10, 1998. 74 US Food and Drug Administration. Regulations restricting the 70 Laugesen M, Meads C. Tobacco advertising restrictions, sale and distribution of cigarettes and smokeless tobacco to chil- price, income, and tobacco consumption in OECD coun- dren and adolescents: final rule. Federal Register August 28, tries 1960–1986. Br J Addict 1991;86:1343–54. 1996;61:44476.

Reviewers for Tobacco Control

In 1999, 140 people assisted in providing reviews for original articles. The editors would like to thank these people most sincerely for their eVorts. Without them, peer review would not be possible.

Michael Alavanja James Davis Deborah Hennrikus Mimi Nichter Seth Emont Eva Alberman Richard Daynard David Hill Jennifer O’Loughlin Herb Severson Harmony Allison Joe DiFranza Bert Hirschhorn Tracy Orleans Don Sharp http://tobaccocontrol.bmj.com/ Robert Anderson Annette Dobson Andrew Hyland Rosalie Pacula Julia Shelley JeVrey Arnett Rob Donovan Maritta Jaakkola John Pauly Donald Shopland David Ashley Douglas Eadie Peter Jacobson Andrew Penman Michael Siegel Edith Balbach Arthur Farkas Konrad Jamrozik Cheryl Perry John Slade Scott Ballin Matthew Farrelly Martin Jarvis John Pierce Fran Stillman Clive Bates Ellen Feighery Leonard Jason Stephen Platt Theresa Stockwell Teti Adrian Bauman JeVrey Fellows Luk Joossens Rudolf Poledne JeVrey Stoddard Michael Begay Roberta Ferrence Murray Kaiserman Richard Pollay Daniel Stokols Neal Benowitz Brian Flay Lynn Kozlowski Linda Pucci Victor Strecher Lisa Bero Brian Flynn Tim Lancaster Thomas Radecki Stephen Sugarman Anthony Biglan Jean Forster Anne Landman Dorothy Rice Charyn Sutton Richard Bonnie Jonathan Foulds Harry Lando Bill Rickert David Sweanor

Ron Borland Don Garner Nadine-Rae Leavell Arja Rimpela Michael Thun on September 30, 2021 by guest. Protected copyright. David Buck Elizabeth Gilpin Scott Leischow Chris Roberts Scott Tomar David Buller Afaf Girgis Scott Lennox Lyn Roberts Joy Townsend John Cairns Joe Gitchell David Levy Leslie Robinson Wayne Velicer Frank Chaloupka Stanton Glantz Edward Lichtenstein Robert Robinson Jeanette Ward Laurie Chassin Tom Glynn Jim Lightwood Todd Rogers Ken Warner Won Choi Christine Godfrey Willard Manning Laurie Ruggiero Cassandra Welch Joanna Cohen Margo Goodin Danny McGoldrick Henry SaVer Andrew Wilson Suzanne Colby Neil Grunberg William McIntosh Jon Samet Susan WoodruV Greg Connolly Prakash Gupta David Mendez James Sargent Alistair Woodward Rob Cunningham JeVrey Harris Jane Moore Penny Schoﬁeld Trevor Woollery Candace Currie Gerard Hastings Joel Moskowitz Caroline Schooler Robert Wyckham Daniel McGoldrick Jack Henningﬁeld Cheryl Moyer Michelle Scollo Shu-Hong Zhu