anti smoking

THE EFFECTIVENESS OF ANTI-SMOKINGLEGISLATION: A REVIEW

Rajeev K. Goel

Department of Economics, Illinois State University

Michael A. Nelson

Department of Economics, University of Akron

Abstract. This survey focuses on government efforts to curb the use of undesir-able goods, notably tobacco products. We synthesize the economics literature andexamine the effectiveness of government curbs on tobacco consumption throughnon-price controls (such as bans on cigarette advertising, health warnings, andworkplace smoking bans) and price measures (or higher prices through highertaxes). This literature review is unique in that we do not merely aim to provide asummary of the literature. Rather, our main focus is to draw conclusions fromthe literature regarding the effectiveness of alternate policy measures acrosscountries in checking smoking and to provide directions/suggestions for extend-ing the scope of government intervention to other tobacco products.

Keywords. Advertising; Cigarettes; Elasticity; Tax; Tobacco

1. Introduction

In recent years, there has been heightened interest among the public and policy-makers regarding the costs of smoking, especially in light of the evidence on theeffects of second-hand smoke (Manning et al., 1989). Smoking-related prematuredeaths in the United States are estimated to be about 400,000 per year, and anadditional 3,000 premature deaths are due to second-hand smoke. It is estimatedthat by 2030 the worldwide death toll due to smoking will be around ten millionannually (Saffer and Chaloupka, 1999). Yet, despite such grim statistics, smokingacross the globe remains significant as 35% of men in developed countries and50% in developing countries smoke. For women, the corresponding figures are22% (developed countries) and 9% (developing countries) [World HealthOrganisation (WHO), 2002].

In the last four decades, governments across the world have tried to controlcigarette consumption (smoking) using various measures. Initially, these policieswere driven from concerns regarding the health of smokers, whereas morerecently the health of non-smokers (dangers of second-hand smoke) has also

0950-0804/06/03 0325–31 JOURNAL OF ECONOMIC SURVEYS Vol. 20, No. 3# 2006 The AuthorsJournal compilation # 2006 Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UKand 350 Main Street, Malden, MA 02148, USA.

been a concern. Roemer (1993, p. xi) reports that at the beginning of the 1990s,more than ninety countries and territories had national anti-smoking legislation.1

Beyond this, the World Health Assembly has adopted nearly twenty resolutionsto control tobacco consumption since 1970,2 and the World Health Organizationrecently negotiated an international treaty to impose worldwide restrictions ontobacco marketing, consumption, and smuggling.3

Policy makers have used both price and non-price measures to combat smoking(see Figure 1).4 The price measures are primarily based on reducing smokingusing higher cigarette prices driven by higher taxes. Governments everywherehave imposed taxes on cigarettes. The effectiveness of price measures, however,maybe limited by the habit-forming nature of cigarettes and their low priceelasticity of demand. A low price elasticity implies demand insensitivity, therebylimiting the opportunities for consumption reduction via higher taxes. Beyond taxhikes, governments have turned to the courts to seek restitution for their costs oftreating smoking-related illnesses. For example, the Master Settlement Agreement(MSA) signed in the US in 1998 resulted in an increase in the price of cigarettes of45 cents per pack.5 Governments elsewhere, including Guatemala, Nicaragua,Venezuela, Bolivia, the Ukraine, and several Canadian provinces have alsosought to recover health care costs associated with tobacco consumption throughsimilar means.6

Non-price smoking control measures include numerous initiatives including,but not limited to, cigarette-advertising bans, health warnings on cigarettepackages, and territorial restrictions (such as workplace bans, restrictions onsales of tobacco products to minors, bans on smoking in public places).7

Territorial restrictions are driven primarily from concerns about the effects ofsecond-hand smoke. A key input into the effectiveness of advertising relatedpolicy measures is the advertising elasticity of cigarette demand. A low advertis-ing elasticity would undermine the effectiveness of any policies that seek to limitadvertising.

Most of the research that has been conducted on tobacco control has focusedon developed countries (especially the United States) and on cigarettes [seeChaloupka and Warner (2000) and Centers for Disease Control (2000)].Cigarettes appear to be economically more important than other tobacco pro-ducts (including chewing tobacco, cigars, pipe tobacco, and snuff), and our focus

Price-based policies Non-price-based policies

Taxincreases

Other Advertisingrestrictions

Health warnings

Territorial restrictions

Figure 1. Smoking Control Policies.

326 GOEL AND NELSON

# 2006 The AuthorsJournal compilation # 2006 Blackwell Publishing Ltd

will also mainly be on this form of tobacco consumption.8 While most of theextant literature has focused on tobacco control in developed countries, it isnoteworthy that it is in developing countries where tobacco use has been increas-ing in recent years. The developing countries’ share of world tobacco consump-tion rose from 49% in 1974–1976 to 61% in 1984–1986 and was expected toincrease to 71% by 2010 (Baris et al., 2000; Food and Agriculture Organization ofthe United Nations, 2003).

Controls on cigarette consumption may be voluntary or involuntary. Voluntaryrestrictions are generally agreed upon between regulators and the cigarette indus-try to thwart potential regulation (e.g. agreements in certain countries to cut thetar content of cigarettes), whereas involuntary or legislative controls are man-dated by law and include indoor air quality regulations, bans on advertising incertain media, etc. In a study that compares the effectiveness of regulatory andvoluntary smoking controls across OECD countries, Cox and Smith (1984) findthat regulatory controls were more effective in curbing smoking than voluntarycontrols.

Most governments continue to use combinations of price and non-pricemeasures to combat smoking. The effectiveness of any one measure or any setof measures is not clear. Are non-price measures more effective than pricemeasures in curbing smoking? Under what circumstances should a given measure(or set of measures) be used? What are the effects of government intervention incigarette markets on related markets? Should policies be different for populationsubgroups (teens vs. adults)? A comparison of price and non-price measures is notonly important for policy purposes but it also has implications for the direction offuture research on tobacco control. For instance, can the effects of price measureson tobacco control be determined, as much of the literature does, without con-trolling for non-price measures?

In this paper, we review the economics literature on the effectiveness of priceand non-price policy initiatives in combating smoking.9 While a majority of theliterature deals with the United States, we have incorporated comparable studies,when available, from other countries as well. This international comparison isinteresting because some countries (notably Canada and New Zealand) havemore restrictive smoking control policies than the United States. Our review isdistinctive in that we do not merely aim to provide a summary of the literature.Rather, our main focus is to draw conclusions from the literature regarding theeffectiveness of alternate policy measures in checking smoking and to providedirections/suggestions for extending the scope of government intervention toother tobacco products.

2. Estimating the Demand for Cigarettes

The theoretical underpinnings of the cigarette demand models found in empiricalliterature vary from ‘traditional’ demand models that do not account for thespecial addictive nature of tobacco consumption to addiction models that aredynamic in nature and attempt to incorporate this characteristic into the model.

THE EFFECTIVENESS OF ANTI-SMOKING LEGISLATION 327


2.1. Traditional Demand Models

The traditional models of estimating cigarette demand follow the framework ofestimation used for other products [see Houthakker and Taylor (1970)] and donot make allowance for the addictive attributes of tobacco products, whereconsumption in a period affects consumption in subsequent periods. Hence,these models are essentially static models where there are cross-period linkagesamong the variables.

The general format of such estimation takes the form:

QðtÞ ¼ f�PðtÞ;XðtÞ;RðtÞ

�ð1Þ

Here Q is the per-capita cigarette consumption in period t, P is the per-unit price,X is a vector of shift variables including income, related prices, advertising etc.,and R is a vector of regulatory variables. The price elasticities can then be derivedfrom the coefficient on the P variable, income and cross-price elasticities followfrom X, and the effectiveness of regulatory measures can be gauged from R.

Strictly speaking, single equation models of demand assume unidirectionalinfluences and do not focus on the supply side. The models can provide consistentestimates when the supply curve is horizontal (perfectly elastic). However, moregenerally, the demand and supply equations should be estimated simultaneously[see Bishop and Yoo (1985)]. The demand equation would be of the form outlinedin equation (1), while the supply equation would include the quantity supplied asa function of price, input prices, and other factors affecting supply and relatedtaxes. The system of demand and supply equations may be estimated usingvarious econometric techniques, including two-stage least squares (2SLS) and3SLS [see Kmenta (1971)].

2.2. Irrational (Myopic) Addiction Models

The addiction models are dynamic models where, given the habit-persistence natureof tobacco products, consumption in previous periods affects current consumption.Another way to envision these models is to think of the dynamic stock-adjustmentprocess associated with durable goods. We refer to a subclass of addiction models asmyopic or irrational because, while the past consumption influences current con-sumption, no consideration is given to the future consequences of current consump-tion. This can be crucial in the case of tobacco products, where current consumptioncould have harmful health effects in the future.

Using the notation from above, the dynamic aspects of a myopic additionmodel may be captured as [see Houthakker and Taylor (1970)]

QðtÞ ¼ f�PðtÞ;XðtÞ;RðtÞ; SðtÞ

�: ð2Þ

Here the new term S(t) represents the stock of habits and may be decomposed as

SðtÞ ¼ Qðt� 1Þ þ ð1� aÞSðt� 1Þ; ð3Þ

328 GOEL AND NELSON


where a is the depreciation rate of habits. It is because of these dynamic effects thatspecial attention needs to be devoted to teenage smoking. If effective anti-smokingpolicies are able to dissuade young potential smokers, the more difficult task ofhaving older smokers (who have formed smoking habits) quit can be avoided.

2.3. Rational Addiction Models

The rational addiction models are a more general class of dynamic models thatimprove upon the myopic models by having current consumption not only affectedby past consumption but also influencing future consumption. The idea is thatrational utility-maximizing individuals incorporate the interdependence betweentheir past, present, and future consumption in their decision-making process. Thesemodels follow from the seminal work of Becker andMurphy (1988) and Becker et al.(1994). Here, the estimated demand equation might take the following general form:

QðtÞ ¼ f�PðtÞ;Qðt� 1Þ;Qðtþ 1Þ;XðtÞ;RðtÞ; SðtÞ

�ð4Þ

As evident from equation (4), in ration addiction models, current consumption isaffected by both past and future consumption. In other words, in this case, thediscount rate for future consumption is less than infinity.

A refinement over the rational addiction models has been proposed by (Gruberand Koszegi, 2001). They build upon the Becker–Murphy methodology by allow-ing smoking choices to be time-consistent. Then time-consistent rational smokersnot only impose externalities on others (e.g. health costs of second-hand smoke),they also impose ‘internalities’ on themselves (e.g. smokers themselves suffer illhealth in the long term). An important policy implication of this new approachis that higher cigarette taxes not only improve social welfare by internalizingexternalities, but they might also (contrary to conventional wisdom) be beneficialto smokers themselves.

3. Effectiveness of Non-price Measures

There are a number of non-price options available to policymakers to combatsmoking. We consider the three most prominent ones: effectiveness of cigarette-advertising bans, effectiveness of health warnings, and effectiveness of workplacebans.

3.1. Effectiveness of Advertising Bans

Tobacco advertising has tried to increase tobacco consumption for more than acentury. Over time, this advertising has become more specialized by targetingspecial media and population groups.10 Various tactics are used by cigarettecompanies to advertise their products, including media, in-store displays,coupons, sports promotion etc. Cigarette advertising expands the demand forcigarettes by bringing in new smokers (or inducing existing smokers to smokemore often). This is the traditional justification for advertising by cigarette firms.



However, market structure considerations might force firms to advertise evenwhen the total market is not expanding. In such cases, firms advertise to maintainor gain market share, i.e. advertising is ‘cannibalistic’. In practice, advertisingmight be expanding and reshuffling the market at the same time.11

Governments have tried to combat tobacco advertising to curb smoking. Forinstance, in 1998, the European Union adopted a resolution to ban most tobaccoadvertising by 2006. Boddewyn (1994) cites three primary reasons for justifyingadvertising bans on tobacco products. (1) Restrictions on cigarette advertisingwill help reduce smoking to some target level; (2) tobacco manufacturers shouldnot be allowed to undermine anti-smoking efforts by glorifying smoking inadvertisements; and (3) spending on tobacco advertising substantially exceedsspending on anti-smoking advertising. Various entities all over the world areinvolved in banning of cigarette advertising. These include governments (e.g.Canada, France, New Zealand, US), international organizations (e.g. TheWorld Health Organization), supranational institutions (e.g. The EuropeanUnion), and various health associations (Boddewyn, 1994, p. 311). Goel andNelson (2004) report that tobacco-advertising restrictions are now pervasiveexcept for low-income developing countries. Countries like Iceland, Finland,Norway, and Portugal were the early leaders in banning tobacco advertising inall media (Roemer, 1993, p. 291). In contrast, one high-income country (Japan)had no advertising restrictions imposed by the central government on tobacco atthe turn of the century.

In 1971, the United States banned the advertising of cigarettes in all broadcastmedia. This broadcast-advertising ban was preceded from 1968 to 1970 by theFairness Doctrine that subsidized anti-smoking messages. Cigarette advertisingspending in the US surrounding the period of these policy initiatives is shown inFigure 2. It is clear from these data that the cigarette industry made up for theloss in broadcast advertising by increasing advertising in other media after theban, most notably in-store advertising and in sports arenas.

Researchers have used data at different levels of aggregation, examined differ-ent time periods, and employed various estimation techniques for determining theeffectiveness of policy measures in reducing cigarette consumption. A detailedreview of some of the literature follows.

Numerous studies use state-level US data pooled over time to study cigarettedemand. These studies are summarized in Table 1. Part of the reason for theproliferation of studies at this level of aggregation is the easy availability ofcigarette demand data from The Tobacco Institute. Using state-level US data,Baltagi and Levin (1986) found mild support for the effectiveness of the subsi-dized anti-smoking messages under the Fairness Doctrine in curbing cigaretteconsumption. In another US study, Goel and Morey (1995) focused on thesimultaneity between cigarette and liquor consumption (i.e. the cross-price elas-ticities between cigarettes and liquor). Contrary to most of the literature, theyfound the broadcast advertising ban to be effective in reducing cigarette con-sumption.12 Similar to Goel and Morey’s (1995) findings, Keeler et al. (1993) alsofound smoking restrictions in California to be effective in reducing smoking. In

330 GOEL AND NELSON


another panel data US study, Keeler et al. (1996) provided a new ‘twist’ in thatthey examine whether cigarette firms price-discriminate across states by lookingat the effects on the retail price of cigarettes. The authors find that cigarette firmsdo price discriminate across states and that the regulatory anti-smoking effortsare countered by cigarette producers by lowering cigarette prices.

Advertising restrictions on cigarettes are prevalent in other countries besidesthe United States. In fact, restrictions on smoking in some countries are morestringent than those in the United States. Australia, for example, has quiteextensive curbs on cigarette advertising, including restriction on advertising inthe electronic media. Using Australian data, Bardsley and Olekalns (1999) foundthat whereas workplace smoking bans and health warnings were effective inreducing smoking, anti-smoking advertisements and bans on electronic mediaadvertising were ineffective. The ineffectiveness of electronic media bans wassupported in the case of the UK (Townsend, 1987). However, an earlier studyhad found Australian electronic media bans to be effective (Clements et al., 1985).The effects of smoking regulations (measured by an index that included a com-ponent on advertising policy) and cigarette taxation in Canada were examined byLanoie and Leclair (1998). Two alternate forms of the dependent variables wereused in this study, namely, cigarette consumption and the percentage of smokersin the population. Smoking regulations were found to be ineffective. The inter-esting result in this study is that these findings are reversed when the dependentvariable is the proportion of smokers, and not total cigarette consumption. Whenpooled samples of OECD countries are studied to determine the effectiveness ofadvertising bans in controlling cigarette consumption, it appears that the evidenceis mixed [see Cameron (1998)]. While some studies have found such bans to beeffective in reducing smoking (Laugesen and Meads, 1991), others have foundthat these bans in fact lead to higher cigarette consumption (Stewart, 1993). Yet

0

2

4

6

8

1963 1970 1985 1988 1991 1994 1997 2000Year

Bill

ions

(19

96 $

)

0

50

100

Perc

ent

Domestic cigarette advertising and promotional expenditures (left axis)

TV and radio as percentage of total advertising (right axis)

Figure 2. US Cigarette Industry Advertising.

Source: Eckard (1991) and FTC (2004)



Table 1. Effectiveness of Cigarette Advertising Restrictions.

Study DataEffectiveness of Advertising Restrictions

in Reducing Smoking

US Studies: Using Aggregate Data

Baltagi and Levin (1986) US, annual state-level, 1963–1980 Broadcast ban ineffectiveDoroodian and Seldon (1991) US, annual aggregate, 1952–1984 Restrictions effectiveFarr et al. (2001) US, annual aggregate, 1955–1994 Broadcast ban ineffectiveGallet (2003) US, annual state-level, 1961–1997 Broadcast ban increased consumptionGoel and Morey (1995) US, annual state-level, 1959–1982 Broadcast ban effectiveHu et al. (1995) California, monthly, 1980–1993 Both tax and media restrictions had negative

impact on consumptionSeldon and Doroodian (1989) US, annual aggregate, 1952–1984 Media policy coefficients negative but

statistically insignificantSung et al. (1994) US, annual state-level, 1967–1990 Broadcast ban ineffective

US Studies: Using Micro Data

Czart et al. (2001) US, college students, 1997 Cigarette billboard advertising bans effective;student newspaper bans ineffective

International Studies: Using Aggregate Data

Bardsley and Olekalns (1999) Australia, annual, 1962/1963–1995/1996 Antismoking advertising and electronicmedian bans had little effect

Cameron (1997) Greece Effect of TV-advertising ban insignificantClements et al. (1981) Australia, annual, 1953/1954–1981/1982 Electronic media ban reduced tobacco

consumptionCox and Smith (1984) 15 OECD countries, annual, 1962–1980 Legislative restrictions more effective

than voluntary restrictionsJohnson (1986) Australia, annual, 1961–1962 to 1982–1983 Effect of electronic media advertising ban

insignificantJohnson (1986) Australia, annual, 1961–1962 to 1982–1983 Effect of electronic media advertising ban

insignificant

(continued)

332

GOELAND

NELSON

#2006TheAuthors

Journalcompila

tion#

2006Blackwell

Publish

ingLtd

Lanoie and Leclair (1998) Canada, provincial, 1980–1995 Effects of anti-smoking regulation mixeda

Laugesen and Meads (1991) 22 OECD countries, annual, 1960–1986 Advertising restrictions effective inreducing smoking

Nelson (2003) 20 OECD countries, annual, 1970–1995 Advertising bans have no effect oncigarette consumption

Roemer (1993) 24 countries, 1970–1986 Comprehensive advertising bans mosteffective in reducing smoking

Saffer and Chaloupka (1999) 22 OECD countries, annual, 1970–1992 Comprehensive bans effective in reducingsmoking; partial bans have little or no effect

Stewart (1993) 22 OECD countries, annual, 1964–1990 Advertising bans led to a small, statisticallyinsignificant increase in average smoking

International Studies: Using Micro Data

Townsend (1987) UK, annual, adults, 1961–1977 Electronic media ban ineffective inreducing consumption

Notes: adenotes that the study uses an index of regulation.Some studies appear in multiple tables due to their broad scope.

THEEFFECTIV

ENESSOFANTI-S

MOKIN

GLEGISLATIO

N333

#2006TheAuthors

Journalcompila

tion#

2006Blackwell

Publish

ingLtd

others have found that the bans on cigarette advertising had no effect on cigaretteconsumption (Nelson, 2003).

While some of the data might be dated, Roemer (1993, p. 291) providesexcellent international comparisons of the effectiveness of tobacco-advertisingbans. Countries were placed in four groups based on the extent of tobacco-advertising restrictions: (i) countries enforcing complete tobacco-advertisingbans (e.g. Iceland, Finland, Norway, Portugal);13 (ii) countries allowing tobaccoadvertising in few media (e.g. Belgium, France, Italy, New Zealand, Singapore,Sweden); (iii) countries allowing tobacco advertising in most media (e.g.Australia, Austria, Canada, Denmark, France, Federal Republic of Germany,Ireland, the Netherlands, the Switzerland, the United Kingdom, the UnitedStates);14 and (iv) countries allowing tobacco promotion in all media (e.g.Greece, Japan, and Spain). Total advertising bans [group (i)] were most successfulin reducing smoking [also see Saffer and Chaloupka (1999)]. Interestingly,Roemer finds no difference in the smoking-reducing efficacy of policies thatallow advertising in most media and cases where there are no restrictions ontobacco advertising [i.e. between groups (iii) and (iv)].

A large part of the ambiguity with regard to the effect of advertising restrictionpolicies on cigarette demand appears to be because the underlying relationbetween advertising and cigarette demand has not been well understood. Moststudies model advertising restrictions as a 0–1 dichotomous variable that takes ona value of 1 for the period after the policy is implemented (or vice versa). Somestudies have found the effects of contemporaneous advertising on smoking to benegative (Baltagi and Levin, 1986; Goel and Morey, 1995), while others havefound a positive effect with a lag (Goel and Morey, 1995). Doroodian and Seldon(1991) use an extensive lag structure on advertising and report a short-run USadvertising elasticity (the responsiveness of advertising expenditure and cigarettesales) of 0.06 and the corresponding long-term elasticity of 0.20. This is consistentwith what Gallet and List (2003) find in their recent ‘meta analysis’ of advertisingelasticities. In light of this, perhaps a lag structure on the effects of regulation isalso called for.

There are a few studies that focus on regional cigarette demand, most notablyCalifornia. In an interesting study that examines regulation at the state-level, Huet al. (1995) examine the effects of California Tobacco Tax and Health PromotionAct. This Act not only raised cigarette taxes but also earmarked some of the taxrevenues for health education programs to reduce cigarette consumption. Whilethe authors found both positive and negative cigarette advertisements to beeffective, they concluded that advertising was not the most cost-effective way toalter consumer preferences.

It is important to study separately the effects of cigarette advertising on teen-agers because young people might respond differently to advertising messages.Furthermore, very few people begin smoking after their teenage years, hence, thisis an especially important group in any long-run strategy to control tobaccoconsumption.15 The success of teenage anti-smoking policies to date has beencalled in question by the international trend in the decline in smoking initiation

334 GOEL AND NELSON


age.16 In particular, a WHO (1996) report estimates that in many countries in theearly 1990s the median smoking initiation age was under 15 years.

As far as we have been able to determine, only one study has addressed theeffectiveness of advertising bans and anti-smoking messages on teenage tobaccouse. Pechmann and Ratneshwar (1994) exposed about 300 12- to 13-year oldCalifornians who were non-smokers, to magazine advertising to determine theirperceptions about peers who smoked. Subjects who saw anti-smoking advertise-ments ‘judged the smoker to be relatively immature and unglamourous’. In lightof this, special advertising campaigns are being targeted to prevent teen smoking.Such advertisements stress situations related to smoking that teens perceive as‘uncool’. There is an obvious need for more research on the responses of teenagersto cigarette advertising.

The overall impact of advertising restrictions on cigarette consumption variesacross studies, driven primarily by the nature of data and the estimation techni-ques (see Table 1). Some studies have used aggregated data sets, whereas othershave been able to use data at lower levels of aggregation. The time periodscovered in the analyses of cigarette demand have varied, although almost allstudies focus on the latter half of the twentieth century. In the case of theUnited States, most studies have focused on the effect of the broadcast advertis-ing ban imposed in 1971. This measure banned all cigarette advertising on radioand television.

The evidence in the literature regarding the effectiveness of this ban seems to bemixed.17 A part of this ambiguity seems due to the inability to adequatelyquantify advertising, its effects, and consequently, the effects of advertisingbans. More generally, the scope of advertising also matters, and some recentevidence points to the conclusion that relatively comprehensive bans on advertis-ing are an effective strategy to reducing smoking.

In sum, Table 1 summarizes that the overall evidence on the effectiveness ofmedia bans is mixed. Interestingly, nearly half of the US studies found the mediabans to be effective, whereas such evidence is less compelling among non-USstudies. While our understanding of the behavior of smokers and their responsesto advertising has improved in recent times, there is a need for more researchusing micro-level data including more research focusing on population subgroupssuch as teenagers and religious and ethnic groups. There also appears to be a needfor better modeling of effects of regulation besides using standard dummy vari-ables and for the evaluation of such regulations within a broader context of thefull array of public policies to reduce tobacco consumption. A cross-countrysurvey including consistent data for various countries would facilitate interna-tional comparisons.

It is also noteworthy that the recent advances in telecommunications haveopened up new avenues for advertisers. The most notable development in thisregard is the explosive growth of the Internet. Now smokers in a particularjurisdiction are susceptible to advertising from other (unregulated) jurisdictions.This would further undermine the ability of authorities to institute effective policymeasures.



3.2. Effectiveness of Health Warnings

Smoking can impose substantial health costs on smokers. Max’s (2001) survey ofthe literature finds that annual costs related to smoking constitute 6–14% ofpersonal health expenditures. Health warnings related to smoking include man-dated warning labels on cigarette packages and the distribution of other healthmaterials. The United States was at the forefront of legislating smoking-relatedhealth warnings with the passage of the Federal Cigarette Labeling andAdvertising Act in 1965. This law was revised and strengthened in 1970 and1984. The 1984 legislation required a set of four rotating health warnings oncigarette packages in the United States [for details, see Viscusi (1994, p. 262)].Internationally, the European Economic Community adopted the TobaccoLabeling Directive in 1989 to warn consumers of smoking-related health hazards(Cooper, 1990). This Directive was similar to the Federal Cigarette Labeling andAdvertising Act enacted in the United States in 1965. Globally, 90 countries (of atotal of 135 countries where data are available) mandated by the year 2000 someform of specific requirement as to how the warning label is to appear on tobaccopackaging (Goel and Nelson, 2004). In several of these countries, the require-ments are more restrictive or stronger than what is currently mandated in theUnited States.18

In 1964, the US Surgeon General issued a report warning about the ill healtheffects of smoking. This policy was followed by the Fairness Doctrine from 1968–1970 that subsidized anti-smoking messages. Baltagi and Levin (1992, p. 330)report that ‘The value of these (Fairness Doctrine) anti-smoking messages wereestimated at $75 million in 1970. This is roughly one-third the industry’s adver-tising expenditure on TV and radio for that year’. The broadcast ban on cigaretteadvertising replaced the Fairness Doctrine in 1971.

Most studies for the United States have examined the effectiveness of theFairness Doctrine in reducing cigarette consumption, while some studies dealwith the effectiveness of health warning labels on packages and other health-related information. These policies are classified under the broad heading ofhealth warnings in Table 2, and the evidence regarding the effectiveness of thesepolicies is mixed. One of the earliest studies to examine the effects of healthwarnings on smoking was by Hamilton (1972) who found them to be effectivein reducing US cigarette consumption. These findings were later supported byBaltagi and Levin (1986) and Seldon and Doroodian (1989). A study of the effectsof health warnings on smoking by youth found that the Fairness Doctrine ‘had asubstantial negative impact on teenage smoking participation rates’ (Lewit et al.,1981, p. 569).

Gallet and Agarwal (1999) used a switching regression model, instead ofdummy variables, to determine the effects of health warnings on smoking in theUnited States and found that the health warnings were gradually effective inchecking cigarette consumption. This finding is at odds with the earlier resultsfor the UK, where effects of health warnings were found to be short term(Atkinson and Skeggs, 1973). Farr et al. (2001) estimated both cigarette demand

336 GOEL AND NELSON


Table 2. Effectiveness of Health Warnings.

Study DataEffectiveness of Health Warnings

in Reducing Smoking

US Studies: Using Aggregate DataBaltagi and Levin (1986) US, annual state-level, 1963–1980 Fairness Doctrine effective in reducing smokingBaltagi and Levin (1992) US, annual state-level, 1963–1988 Effects of Fairness Doctrine mixedBishop and Yoo (1985) US, annual aggregate, 1954–1980 Health warnings had little impact on cigarette

demandBlaine and Reed (1994) US, annual aggregate, 1946–1992 Health warnings significant in reducing smokingFarr et al. (2001) US, annual aggregate, 1955–1994 Fairness Doctrine ineffectiveGallet (2003) US, annual state-level, 1961–1997 Fairness Doctrine effectiveGallet and Agarwal (1999) US, annual aggregate, 1955–1990 Health warnings gradually reduce smokingGoel and Morey (1995) US, annual state-level, 1959–1982 Fairness Doctrine increased smokingHamilton (1972) US, annual aggregate, 1925–1970 Health warnings effective in reducing smokingSeldon and Doroodian (1989) US, annual aggregate, 1952–1984 Health warnings reduced cigarette demandYurekli and Zhang (2000) US, state-level, 1970–1995 Anti-smoking laws significant in reducing smokinga


Lewit et al. (1981) US, youth, 1966–1970 Fairness Doctrine effective in reducing smokingparticipation rates


Atkinson and Skeggs (1973) UK, annual, 1951–1970 Health warnings only temporarily effective inreducing smoking

Bardsley and Olekalns (1999) Australia, annual, 1962/1963–1995/1996 Health warnings had minor effectCameron (1997) Greece Anti-smoking campaign effective

(continued)

THEEFFECTIV

ENESSOFANTI-S

MOKIN

GLEGISLATIO

N337

#2006TheAuthors

Journalcompila

tion#

2006Blackwell

Publish

ingLtd

Table 2. Continued.

Study DataEffectiveness of Health Warnings

in Reducing Smoking

Conniffe (1995) Ireland, annual, 1960–1990 Proportion of smokers negatively affectedby health warnings; smokingunaffected by health warnings

Leu (1984) Switzerland, annual, 1954–1981 Media publicity reduces cigarette consumptionPerkurinen (1989) Finland, annual, 1960–1981 Effects of anti-smoking advertising ambiguousRadfar (1985) UK, quarterly, 1965–1980 Health warnings effective in reducing smokingStavrinos (1987) Greece, annual, 1960–1982 Health warnings effective in reducing smokingTansel (1993) Turkey, annual, 1960–1988 Health warnings and anti-smoking campaigns

effective in reducing smokingWilcox et al. (1994) South Korea, monthly, 1988–1992 Health warnings not significant in reducing

smoking; however, consumption of foreigncigarettes reduced

Witt and Pass (1981) UK, annual, 1955–1975 Health warnings effective in reducing smoking


Lewit et al. (1997) US, Canada, school students, 1990, 1992 Anti-tobacco advertising correlated with greaterlikelihood of smoking

Zanias (1987) Greece, 1974 household survey (also usestime-series data)

Anti-smoking campaign significantlyreduced smoking


338

GOELAND

NELSON

#2006TheAuthors

Journalcompila

tion#

2006Blackwell

Publish

ingLtd

and supply for the US and found that while the Fairness Doctrine coefficient wasstatistically insignificant in the demand equation, it was positive and significant inthe supply equation. Time-series analyses of individual states, however, showedthe effect of the Fairness Doctrine to be mixed across states (Baltagi and Levin,1992). Goel and Morey (1995), on the contrary, found that cigarette consumptionincreased due to the Fairness Doctrine. Broadly speaking, an earlier review of theliterature by Cameron (1998) found effects of health warnings to be generallynegative.

Quantitative comparisons of the effects of health warnings across differentstudies are problematic because some studies use dummy variables to takeaccount of regulatory phases [e.g. Baltagi and Levin (1986, 1992); Goel andMorey (1995); Hamilton (1972)], whereas others use some indices of regulation[e.g. Lanoie and Leclair (1998), Laugesen and Meads (1991), Yurekli and Zhang(2000)]. The results of these and other studies discussed in this section aresummarized in Table 2.

In sum, the overall evidence on the effectiveness of cigarette health warnings ismixed. This ambiguity was probably the main impetus behind the short life of thisDoctrine. The main problem in ascertaining the impact of health warnings oncigarette consumption appears to be the difficulty in getting a handle on thequalitative differences among various health warnings and the related institu-tional details. How is the nature of certain health warnings different from othersand how can this difference be adequately measured?

3.3. Effectiveness of Workplace and Public Smoking Bans

Another strategy to reduce smoking deals with restrictions on smoking in publicplaces, including common work areas. While initially workplace restrictions wereput in place to address concerns about fire safety and food contamination, in theseventies indoor air quality became a concern (Chaloupka and Saffer, 1988). Thispolicy arose primarily from knowledge about the adverse effects of second-handsmoke (i.e. effects on the non-smoker). Another thrust behind the prevalence ofworkplace bans has been the economic incentives from insurance companies.Insurance companies, due to fire hazards of smoking and the costs on non-smokers, now routinely charge higher premiums to establishments without adesignated smoking area. These restrictions protect other consumers (e.g. desig-nated non-smoking areas in restaurants) and workers (e.g. smoking bans onairplanes). By imposing restrictions on where the smokers can smoke, policymakers are implicitly trying to raise the costs of smoking.

Significant work on the effectiveness of workplace smoking restrictions in theUnited States is due to Chaloupka and associates. Chaloupka and Saffer (1988)used data on US states to determine the effect of workplace smoking restrictionson cigarette consumption. They find that cigarette demand in a state had anegative impact on the state’s ability to legislate clean air restrictions. However,even in states that enacted such laws, they were found not to have a significantnegative impact on cigarette demand. In contrast, a study of the effects of



regulation on US college students by Chaloupka and Wechsler (1995) found thatstrong restrictions on smoking in public places (e.g. restrictions on smoking inrestaurants) significantly reduced smoking participation rates among teenagers.

Internationally, we also found evidence pointing to the effectiveness of work-place smoking bans, especially in the case of Australia and Germany. For exam-ple, Bardsley and Olekalns (1999), Borland et al. (1991), and Wakefield et al.(1992) found workplace smoking bans to be effective deterrents to smoking in thecase of Australia. Brenner and Mielck (1992) surveyed individuals in the FederalRepublic of Germany during 1987 and found that workplace smoking bans wereparticularly effective in reducing smoking among German females.

Some researchers have examined the effectiveness of territorial smoking bansusing composite indices of regulation (Lanoie and Leclair, 1998; Laugesen andMeads, 1991). In such cases, it is not possible to discern the effectiveness of anyone type of measure. Because workplace restrictions are particularly sensitive tothe location, one would need data at the micro level to discern the effectiveness ofthese restrictions. For example, potential smokers have less freedom in avoidingsome no-smoking areas than others. A smoker can avoid eating at a ‘no smoking’restaurant but has less freedom in avoiding his/her ‘no smoking’ workplace. Aresmoking restrictions imposed in restaurants more effective than those imposed inoffices? These kinds of differences are difficult to discern with aggregated datasets.

A summary of studies we could find on the effectiveness of territorial restric-tions in reducing smoking can be found in Table 3. The preponderance of thisevidence points to such restrictions as a promising strategy to control tobaccoconsumption. In fact, Table 3 summarizes that the territorial restrictions wereeffective in reducing smoking in most cases, and in only one case did we find theterritorial restrictions to be ineffective.

4. Effectiveness of Price Measures

Under the price measures to control smoking, our focus is tobacco taxation.Figure 3 plots the real weighted (by sales) average state tax and the Federal taxper pack of cigarettes for the United States since 1955. At the Federal level,nominal tax rates remained unchanged until 1983, resulting in a steady decline inreal rates. Despite a series of nominal rate hikes since then, the real Federal ratestands more than 40% below the level of 1955. At the state level, real ratesremained relatively constant until around the time of the 1964 SurgeonGeneral’s report on smoking and health. Rates then spiked dramatically upwarduntil the early 1970s, at which point state legislatures became concerned overrevenue losses from cross-state bootlegging (Advisory Commission onIntergovernmental Relations, 1985). It was not until the mid-eighties that statesagain started to raise rates in significant numbers and amounts. Today, the meanreal tax rate for the 50 states stands only slightly below the all-time high of themid-seventies. It is important to point out, however, that cigarette tax ratescontinue to vary substantially among the individual states. As of January 1,

340 GOEL AND NELSON


Table 3. Effectiveness of Territorial and Other Smoking Restrictions.

Study Data Effectiveness of Restrictions in Reducing Smoking

US Studies: Using Aggregate Data

Chaloupka and Saffer (1988) US, annual state-level, 1975–1985 Clean air laws ineffective in reducingcigarette demand

Farr et al. (2001) US, annual aggregate, 1955–1994 Indoor clean air laws effectiveSung et al. (1994) US, annual state-level, 1967–1990 Local regulations effective in reducing smokinga


Chaloupka and Wechsler (1995) US, survey of college students Bans on smoking in public places effectivein reducing teenage smoking

Czart et al. (2001) US, college students, 1997 Comprehensive geographic restrictions reducesmoking; bans on cigarette sales increase smoking


Bardsley and Olekalns (1999) Australia, annual, 1962/1963–1995/1996 Workplace smoking bans reduce consumption


Borland et al. (1991) Australia, Telecom Australia employees Workplace smoking bans effective inreducing smoking

Brenner and Mielck (1992) Federal Republic of Germany, individuals, 1987 Workplace smoking bans effective in reducingsmoking, especially among women

Lewit et al. (1997) US, Canada, school students, 1990, 1992 Policies limiting minors’ access to tobaccoand tobacco education reduce smoking; effectof geographic smoking restrictions insignificant

Wakefield et al. (1992) Australia, individuals, 1989 Workplace smoking bans effectivein reducing smoking


THEEFFECTIV

ENESSOFANTI-S

MOKIN

GLEGISLATIO

N341

#2006TheAuthors

Journalcompila

tion#

2006Blackwell

Publish

ingLtd

2004, New Jersey levied the highest rate ($2.05 per pack) while Virginia had thelowest rate (2.5 cents per pack).

Internationally, the US ranks around the middle with respect to the total taxesit imposes on tobacco. For example, the Economic Intelligence Unit reports thatthe price of a pack of Marlboro cigarettes in 2001 (US dollars) ranged from a highof $6.48 in Norway to a low of $0.62 in Indonesia.19 The US ranked slightlyabove the average for all high-income countries during the period (Goel andNelson, 2004).

4.1. Tax Policy and the Demand for Cigarettes

Will tax-induced price increases for tobacco result in a significant reduction inthe use of tobacco? Looking first at some raw data, Figure 4 plots malesmoking prevalence (percent of males aged 15 years and older who smoke) inthe early 1990s with the average retail price of a pack of cigarettes as of 1996 fortwenty-three (primarily developed) countries. The data do not reveal a strongnegative relationship between these two variables, although the simple correlation

0

5

10

15

20

25

30

35

1955 1960 1965 1970 1975 1980 1985 1990 1995 2000

Cen

ts

State Tax Federal Tax

Figure 3. US Real Average State and Federal Tax (cents per pack in 1982–1984$).

0.0

10.0

20.0

30.0

40.0

50.0

60.0

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

Retail price (US $/pack)

Mal

e sm

okin

g pr

eval

ence

Figure 4. Smoking Prevalence vs Retail Price of Cigarettes in Selected Countries.

342 GOEL AND NELSON


is �0.63. Norway, for example, has the highest retail price for cigarettes, yet malesmoking prevalence is nearly average for the sample.

Of course, such simple analysis does not control for other factors that influencetobacco consumption such as socioeconomic considerations (e.g. income andemployment status, education, religious beliefs) and the non-price public policiesdesigned to reduce tobacco consumption discussed in earlier sections of thispaper. Since the 1970s, numerous econometric and other statistical studies havebeen done, in varying degrees, to control for such factors and provide estimates ofthe price elasticity of cigarette demand. Because of space constraints and the factthat these studies have been adequately summarized elsewhere, they will not beindividually reviewed here.20 Instead, the cumulative findings/conclusions of themajor studies undertaken in the last 20 years will be summarized.

Most empirical studies have focused on the demand for cigarettes in the USand have used data on aggregate consumption, either at the state level or time-series national data. These studies also vary widely in (1) the level of dataaggregation, (2) underlying theoretical mode (e.g. traditional demand, myopic,or rational addiction demand specification), (3) estimation strategy, and (4)whether or not they control for non-price anti-smoking polices and otherdemand-related factors in their empirical setup. With the exception of Baltagiand Goel (1987), these studies have based the consumption response on priceelasticities rather than tax elasticities.21 Furthermore, the effects of both supplyand demand influences on cigarette demand have also received little attention [seeGallet (2004) for a notable exception]. Using an interesting new angle, Goel andRam (2004) provide quantile regression estimates of cigarette demand elasticitiesfor the United States over 1993–1999. They find that the price elasticity ofcigarette demand shows a significant variation across the high- and low-quantityquartiles. This finding puts in perspective the limitations of constant-elasticitymodels.

Given these methodological differences, it is perhaps not surprising that there isa considerable range in the elasticity estimates in the extant literature. The firstrow in Table 4 reports the lowest, mean, and highest elasticity estimate fortwenty-four studies that have been done since 1985 using aggregate US data.When these studies are viewed as a whole, the price elasticity estimates range from�0.14 to �1.12 with a mean of �0.40. Despite the wide range, over half of thestudies (14 of 24) report an elasticity estimate within the range from �0.3 and�0.5. Viewed collectively, these studies suggest that tax policy is a viable strategy toreduce cigarette consumption. That being said, there is some evidence that cigarettedemand is becoming more price inelastic over time. For example, two studies(Baltagi and Goel, 1987; Tegene, 1991) estimate price/tax elasticity estimates fordifferent points in time and report lower estimates for recent time periods. Asimilar conclusion is reached by Gallet and List (2003) in their meta-analysis ofcigarette demand price elasticities.

The second row in the table summarizes the results for six studies of developedcountries outside the United States. The mean cigarette elasticity estimate issimilar in magnitude to the United States. The third row pertains to studies of



developing countries. The mean price elasticity estimate for the nine studiessummarized here is nearly �0.54, well above the estimate for developed countries,and consistent with economic theory (Warner, 1990).

Several studies, especially more recent ones, have estimated price elasticity ofcigarette demand using data on individuals. Individual-level data sets have severaladvantages over aggregated data sets, including (1) measurement problems account-ing for cross-border sales and smuggling are no longer relevant, (2) the econometricproblems associated with possible two-way causality between the consumption andprice variables do not have to be addressed, and (3) the researcher can obtainelasticity estimates for specific population subgroups (e.g. youth, low-income, ethnicgroups) and estimate the effect of prices on the decision of whether or not to consumetobacco (participation decision) as well as the quantity consumed. However, the useof individual-level data introduces new methodological problems, including the factthe survey respondents typically under report tobacco consumption.

In Table 4, the fourth row reports the mean and range of price-elasticityestimates of cigarette demand of seven recent studies using US individual-leveldata. These studies, which vary in data set, time period studied, and estimationmethodology, focus on adult consumption. Despite these differences, the range inestimates is reasonably narrow, ranging from a low of �0.25 to a high of �0.47.Importantly, the mean estimate is �0.37 and is quite similar to the studies usingaggregate data for the United States.

The fifth row of Table 4 summarizes the results of seventeen studies on theinfluence of price on the decision to smoke by youth and young adults (smokingprevalence). All but one study report that price has a significant effect on thisdecision, with the mean elasticity of all studies estimated at �0.59. Even the onestudy by Douglas (1998) that reports a zero elasticity is only with regard to the

Table 4. Price Elasticity of Demand for Cigarettes.

LowestEstimate

MeanEstimate

HighestEstimate

Aggregate consumption data (24 US studies) �0.14 �0.398 �1.12Aggregate consumption data (10 internationalstudies: Developed Countries)

0 �0.425 �1.0

Developing countries (nine studies) 0 �0.540 �1.0Individual consumption data on adults(seven US studies)

�0.25 �0.374 �0.47

Youth prevalence (17 US studies) 0 �0.587 �1.21Youth consumption (12 US studies) 0 �0.608 �1.44

Source: Authors’ calculations based on Tables 6.7, 6.8, and 6.9 in the Centers for Disease Control(2000), DeCicca et al. (2002), Farr et al. (2001), and Gallet (2003) for the US studies. Internationalcalculations for the developed countries are based on Atkinson and Skeggs (1973), Goel (2004),Gruber et al. (2003), Lanoie and Leclair (1998), Laugesen and Meads (1991), Leu (1984), Reinhardtand Giles (2001), Stavrinos (1987), Valdes (1993), and Witt and Pass (1981).Developing countries estimates are taken from Chaloupka and Warner (2000), Jha and Chaloupka(2000) and Lance et al. (2004).

344 GOEL AND NELSON


effect of price on the decision by youth to start smoking. That same studyconcludes that price can significantly influence the decision to quit smoking; inparticular, a 10% increase in price is associated with a reduction in the durationof smoking by a similar percentage.

The last row of Table 4 considers that total effect of price on cigarette con-sumption by youth, including both the decision to smoke (smoking participation)and the quantity smoked. The mean elasticity estimate, based upon twelve studiesundertaken since 1980, is �0.61. This is substantially above the estimates foradults and the estimates based on aggregate data for developed countries that arereported at the top of the table. The findings confirm the widely held view thatteenagers are more responsive to the price of tobacco than adults. In short, withan increase in price, fewer youths choose to smoke, and those that do, smoke less.

Having said this, it should be pointed out that a few studies, notablyChaloupka (1991) and Wasserman et al. (1991), find that price has an insignif-icant effect on youth consumption. The findings by Wasserman et al., in parti-cular, are instructive because they include an index of non-price-smokingrestrictions in their model. When this index is excluded, the price-elasticity coeffi-cient becomes statistically significant. The authors suggest that price may serve asa proxy for a variety of anti-smoking policies and that the estimated coefficient onthe price variable is subject to upward bias when these other tobacco controlpolicies are not properly accounted for in the empirical set up.

In summary, it is apparent that tax policy can be used as an effective policy toreduce tobacco consumption, particularly among the youth and in developingcountries. The conclusion holds across a wide variety of demand specifications,estimation methods, and data sets (Gallet and List, 2003). Nevertheless, somecaution regarding this assessment is warranted because many of the empiricalstudies summarized above do not control for non-price approaches to tobaccocontrol, and this may lead to an assessment that overstates the relative impor-tance of tobacco taxation strategies. This is of special concern given that suchlevies are regressive, with the burden falling disproportionately on the poor.

5. Relationship of Cigarettes with other Products

The relationship of cigarettes with other products, including other tobacco pro-ducts and alcohol, has important implications for government policies trying toreduce smoking. For example, if cigarettes are substitutes for other goods, thenany decrease in smoking would be reflected as an increase in the consumption ofother goods.

5.1. Cigarettes and Other Tobacco Products

There is need for research on the relationship between cigarette demand and thedemand for other tobacco products, including cigars, smokeless tobacco, pipesmoking, snuff, chewing tobacco, etc. If it turns out that government restrictions



on smoking are merely shifting demand to other (less regulated) tobacco pro-ducts, then the social effectiveness of any government action will be undermined.

Currently, there are relatively few federal public health laws or regulations thatare applicable in the case of cigars, pipe tobacco, or fine cut-tobacco in the UnitedStates (US Department of Health and Human Services, 2000). The US federalgovernment passed the Comprehensive Smokeless Tobacco Health Education Actof 1986 to place warning labels on packages of smokeless tobacco. The Center forDisease Control reports that US taxes on smokeless tobacco are less than taxes oncigarettes (US Department of Health and Human Services, 2000). Similar tocigarettes, all advertising of smokeless tobacco in the United States is banned inelectronic media.

There has been some economic research on smokeless tobacco use. An inter-esting study by Ohsfeldt et al. (1997) examines the effects of tobacco policies oncigarette consumption and smokeless tobacco use (snuff and chewing tobacco) forthe United States.22 They find that higher taxes on smokeless tobacco decrease itsconsumption, whereas higher cigarette taxes increase the probability of smokelesstobacco consumption. The tax elasticity of smokeless tobacco in the United Stateswas estimated to be �0.55 and the cross-tax elasticity of smokeless tobacco withcigarettes to be around 0.4 (Ohsfeldt and Boyle, 1994). The cross-tax elasticitysuggests the substitution between cigarettes and smokeless tobacco (includingsnuff and chewing tobacco) and reinforces the case for a comprehensive tax policycovering all related tobacco products. In contrast to tax policy, non-price anti-smoking restrictions at the state level were found to have no effect on smokelesstobacco use. More recently, Goel and Nelson (2005) find that in the case ofUnites States, a percentage increase in cigarette taxes has a greater (negative)impact on smoking prevalence than a similar increase in smokeless taxes has onthe use of smokeless tobacco.

5.2. Cigarettes and Alcohol

Empirical evidence also needs to be brought to bear upon the relationshipbetween smoking and drinking. While casual empiricism might suggest theseproducts to be complements, this has not yet been conclusively demonstrated.Furthermore, the social attitudes toward smoking and drinking have changedover time. Over time, it appears that drinking in moderation is more sociallyacceptable than smoking. Cross-price effects between alcohol and tobacco haveprofound implications for tax policy coordination [see Grossman et al. (1993)].

Goel and Morey’s (1995) study of US states found cigarettes and liquor tosubstitutes implying that higher liquor taxes would result in greater cigaretteconsumption. Using survey data on teenagers, Dee (1999) found that forteenagers, drinking and smoking are ‘highly complementary’ (p. 791). Dee’sfindings were supported in the case of Canada (Gruber et al., 2003). Moreattention needs to be paid towards understanding the relation of cigarettes withother products.

346 GOEL AND NELSON


6. Conclusions

The global cigarette consumption per adult remained steady from the early 1970sto the early 1990s (WHO, 1996). However, the composition of this consumptionhas been changing dramatically, calling for new approaches and ideas. Tobaccouse in developed nations seems to have past the peak, while developing countriesare still experiencing increases in tobacco consumption. Smoking prevalenceamong males is slowly declining. In contrast, for females it is rising, especiallyin developing countries (WHO, 2002). According to the WHO (1996), ‘There hasbeen a dramatic shift in the group of countries with the highest rates of per adulttobacco consumption from the early 1970s to the early 1990s. In the early 1970s,consumption was highest in Canada, Switzerland, Australia, and the UK,whereas in the early 1990s, it was highest in Poland, Greece, Hungary, Japan,and the Republic of Korea’. Much of the future increase in tobacco consumptionduring first decade of the twenty-first century is expected to occur in China andother countries in the Far East (United Nations, 2003).

Our survey shows that nations across the world have been actively trying tocombat smoking through various policy measures, and the effectiveness of thesemeasures has been the subject of active economics research. The evidence on theeffectiveness of non-price measures is mixed. Non-price policy initiatives directedat reducing smoking seem to work in some cases, whereas other studies find theeffects of the same measures to be insignificant. No one measure appears superiorto others, although there is some evidence that workplace restriction is able toreduce smoking. Part of the ambiguity is due to the difficulty of quantifying theinstitutional regulatory details. A further practical problem in determining theeffectiveness of any smoking control measure is that these measures are seldomenacted in isolation. Rather, many measures are passed simultaneously, with theresult that the effectiveness of any single measure is nearly impossible to deter-mine.23 Better modeling techniques, especially those that allow for multiplepolicies under the same framework, are needed. Sensitivity analysis of the findingswith comparable analysis will aid in determining the robustness of some of thefindings. In comparing the price and non-price measures, whereas the dynamicsof price elasticities (i.e. short-run versus long-run elasticities) are well understood,we still do not have a good understanding of the short-run and long-run effects ofnon-price measures. For example, how effective health warnings are in controllingsmoking in the short term versus the long term is still not well understood.

Some of the problems of ambiguous results will be addressed as better data,especially disaggregated data, become available. Micro level data will enablefocus on population subgroups, including race, gender, age, ethnicity, educa-tional, and religious backgrounds, and also allow for other qualitative differences.For instance, we found little evidence on the effectiveness of health warningsdirected at youth.

In addition to those discussed above, we list below some directions for futureresearch that would enable better understanding of cigarette demand and helppolicy makers in framing more effective policies to reduce smoking.



. In thinking about the effectiveness of smoking control policies, one must alsokeep in mind the two-way causality between the demand and supply of thesepolicies. In particular, the lack of comprehensive anti-smoking legislation in acountry might be understood in the context of its low smoking rates. Countrieslike Norway, Singapore, Iceland, Finland, Australia, New Zealand and Swedenhave low smoking prevalence [roughly are in the bottom third of the sample ofeighty seven countries reported by WHO (1996)] and also have strong anti-smoking laws. So, either the low smoking rates in these countries might be theresult of these laws, or these could be due to other socio economic factors. Onthe contrary, countries like the United States do not have very strong anti-smoking policies. These policies might be unwarranted in light of the relativelylow smoking rates in the United States. Less-developed countries have quitehigh smoking rates, and few are at the forefront of regulating tobacco use. Infact, only one developing country (Thailand) among the top twenty-five inhighest male smoking rates has enacted comprehensive anti-smoking legislation.This lack of response to the smoking epidemic among developing countriesmight not be due to an absence of desire but due to economic and politicalconstraints. It is politically easier, for example, to legislate tough anti-smokingregulations in an environment where smokers comprise a minority of the elec-torate. Obviously, this is an important issue that warrants further investigation.

. A caveat about the literature on cigarette demand and advertising is in order.Cigarette-advertising data have different components, and their availabilityand aggregation is not consistent over time or across countries. For example,even after the broadcast ban, firms continue to advertise in various forms:print, billboards, promotions, in-store displays, sports promotions, etc. untilthe 1998 MSA in the case of the United States. It is not clear how qualita-tively different these advertising forms are and what, if any, is the differencein their lag structure. Furthermore, cigarette advertising ‘‘may simulta-neously persuade, inform, and create desirable product images’’ (Tremblayand Tremblay, 1995). We need a better understanding of the qualitativedifferences between the various forms of advertising.

. We need a better understanding of the relation of cigarettes with otherproducts including alcohol and other types of tobacco products (i.e. cross-price elasticities). A well-defined cross-price elasticity will have implicationsfor spillovers from the cigarette market to related markets.

. There is a need to evaluate tobacco control policies from a holistic per-spective, considering the whole of price and non-price policies imposed bygovernments, when assessing the merits of any given strategy. While thissignificantly increases the task faced by the researcher, an assessment of theeffectiveness of any given policy may be biased if other tobacco controlstrategies are ignored in the empirical investigation.

. We found a lack of research focusing on smoking behavior in developingnations where tobacco consumption is relatively high and growing. This isespecially relevant because there are a number of unregulated tobacco

348 GOEL AND NELSON


products, other than cigarettes, in developing countries. For instance,cigarettes form only about 65–85% of all tobacco consumption (WHO, 1996).

In conclusion, the spread of the Internet has added an entirely new dimensionto cigarette advertising (Cohen et al., 2001; General Accounting Office, 2002).Advertising can now freely move across national boundaries, creating difficultiesfor enforcement and regulation. Policy makers will have to rethink the regulationof cigarette advertising, especially when the honeymoon with the Internet ends. Inaddition, cigarette advertising seems to be shifting to other arenas, most notablyin-store displays and promotions. Whether the reach of regulators will extend thatfar remains to be seen.

Acknowledgements

A version of this paper was presented at the meetings of the Midwest EconomicsAssociation. We thank John Bridges, Michael Brun, Frank Chaloupka, HassanMohammadi, and Rati Ram for useful comments and discussions. Comments of tworeferees are also appreciated. R. Connelly, C. Duan, M. Mustaffa, and J. Wu provideduseful research assistance. All errors are our own.

Notes

1. For an up-to-date review of international tobacco control policies, see Goel and

Nelson (2004).2. World Health Organization: http://www.who.int/features/2003/08/en/3. http://www.who.int/mediacentre/releases/2003/prwha1/en/print.html

4. A third set of measures, not dealt with in this survey, includes programs aimed atsubsidizing smoking cessation such as nicotine replacement therapy (NRT) products(World Bank, 1999, p. 53).

5. The MSA resolved lawsuits filed by the Attorneys General against the tobacco indus-

try and provided the states funding intended for tobacco prevention and control. Theagreement required tobacco companies to take down all billboard advertising andadvertising in sports arenas, to stop using cartoon characters to sell cigarettes and to

make many of their internal documents available to the public. The tobacco companiesalso agreed not to market or promote their products to young people. The impact ofMSA will only be determined in due course (Capehart, 2001; Viscusi, 2002).

6. http://www.chadbourne.com/practice/sub_litigation.html?s_litigation_product.html�mainFrame

7. An excellent earlier discussion of policies in numerous OECD countries is found

in Marks (1982). While acknowledging that the policies varied considerably acrosscountries, Marks found three policies to be common across various countries: televi-sion and radio advertising bans, publication of health warnings, and restrictions on the‘creative content’ of cigarette advertising.

8. For instance, over the 1921–1999 period, about 98% of tobacco tax revenues generatedin the United States came from cigarettes (Orzechowski and Walker, 2000).

9. We term a measure effective in its ability to control smoking and do not focus on

cost-effectiveness. Ranson et al. (2000) provide an interesting comparison of thecost-effectiveness of various measures.



10. www.cdc.gov/tobacco/sgr/sgr_forwomen/factsheet_marketing.htm11. Duffy (1996b), however, found no evidence in UK data that advertising expands the

total demand for cigarettes.12. Both Baltagi and Levin (1986) and Goel and Morey (1995) used dummy variables to

capture the effect of the advertising ban.13. The WHO (1996) reports that in the early 1990s, 27 countries had total, or almost

total bans, on cigarette advertising. This number had, however, decreased to 18 by

mid-1990s.14. Countries appearing in more than one group are those where there are multiple studies

covering different smoking bans. Some countries might have further changed their

policies since these studies were conducted.15. Evidence of this is found in studies of all consumers where the habit-persistence effect,

measured by including lagged cigarette consumption as an additional regressor in the

estimated equation of cigarette demand, is found to be significant [see, for exampleBaltagi and Levin (1986) and Goel and Morey (1995)].

16. Examples of policies targeted specifically to teenage smoking include restriction on saleof tobacco products to teenagers, restrictions on advertising in the vicinity of schools,

etc. In the early 1990s, about 25 countries had regulations prohibiting the sale ofcigarettes to minors (WHO, 1996).

17. Duffy (1996a, p. 19) concludes in his survey of the literature on advertising restrictions,

‘There are results pointing in both directions in these studies with respect to the impactupon demand of advertising bans. On the basis of this review, however, we are left withthe definite impression that the weight of the evidence in these studies does not give

much support, if any, to those who believe that advertising bans are an effective meansof reducing consumption.’ Also, see Saffer and Chaloupka (1999) for an updated andcomprehensive review of the related literature.

18. For instance, Canada passed the Tobacco Products Control Act in 1988. This Act

requires warnings on cigarette packages sold in Canada that are more prominent thanthose in the US (Mahood, 1995). Mahood (1995) reports that Australia, Singapore,and Thailand are also at the forefront of health warnings on cigarette packages.

19. Source: Economic Intelligence Unit as reported in Guindon et al. (2002).20. One recent review that focuses primarily on the United States is the Centers for Disease

Control (2000, Chapter 6). The individual studies summarized in that report form the

basis for some of the summary data on price elasticity presented below. See also Galletand List (2003).

21. The two elasticities would be the same when the price responsiveness to the tax is unity.

This would happen when the supply is completely elastic.22. We were unable to find related studies dealing with other countries in this area.23. Mahood (1995, p. 14) provides some evidence from Canada in this regard, ‘‘Like many

tobacco control measures, the effectiveness of the new warnings may be difficult to

measure. They entered the market in the same year that Canada experienced severetobacco tax rollbacks which cut the price of tobacco by almost half in manyprovinces’’.

References

Advisory Commission on Intergovernmental Relations (ACIR) (1985). Cigarette TaxEvasion: A Second Look, Washington, DC: U.S. Advisory Commission onIntergovernmental Relations.

350 GOEL AND NELSON


Atkinson, A. B. and Skeggs, J. L. (1973). Anti-smoking publicity and the demandfor tobacco in the U.K. Manchester School of Economic and Social Studies 41:265–282.

Baltagi, B. H. and Goel, R. K. (1987). Quasi-experimental price elasticities of cigarettedemand and the bootlegging effect. American Journal of Agricultural Economics 69:750–754.

Baltagi, B. H. and Levin, D. (1986). Estimating dynamic demand for cigarettes using paneldata: the effects of bootlegging, taxation and advertising reconsidered. Review ofEconomics and Statistics 68: 148–155.

Baltagi, B. H. and Levin, D. (1992). Cigarette taxation: raising revenues and reducingconsumption. Structural Change and Economic Dynamics 3: 321–335.

Bardsley, P. and Olekalns, N. (1999). Cigarette and tobacco consumption: Have anti-smoking policies made a difference? Economic Record 75: 225–240.

Baris, E., Brigden, L. W., Prindville, J., Silva, V., Chitanondh, H. and Chandiwana, S.(2000). Research priorities for tobacco control in developing countries: a regionalapproach to global consultative process. Tobacco Control 9: 217–223.

Becker, G., Grossman, M. and Murphy, K. (1994). An empirical analysis of cigaretteaddiction. American Economic Review 84: 396–418.

Becker, G. and Murphy, K. (1988). A theory of rational addiction. Journal of PoliticalEconomy 96: 675–700.

Bishop, J. A. and Yoo, J. H. (1985). ‘‘Health Scare,’’ excise taxes and advertising ban in thecigarette demand and supply. Southern Economic Journal 52: 402–411.

Blaine, T. W. and Reed, M. R. (1994). U.S. cigarette smoking and health warnings: newevidence from post World War II data. Journal of Agricultural and Applied Economics26: 535–544.

Boddewyn, J. (1994). Cigarette advertising bans and smoking: the flawed policy connec-tion. International Journal of Advertising 13: 311–332.

Borland, R., Owen, N. and Hocking, B. (1991). Changes in smoking behaviour after a totalworkplace smoking ban. Australian Journal of Public Health 15: 130–134.

Brenner, H. and Mielck, A. (1992). Smoking prohibition in the workplace and smokingcessation in the Federal Republic of Germany. Preventive Medicine 21: 252–261.

Cameron, S. (1997). Are Greek smokers rational addicts? Applied Economics Letters 4:401–402.

Cameron, S. (1998). Estimation of the demand for cigarettes: a review of the literature.Economic Issues 3: 51–72.

Capehart, T., Jr. (2001). Trends in the cigarette industry after the Master SettlementAgreement. United States Department of Agriculture, TBS-250–01, October.Available at www.ers.usda.gov

Centers for Disease Control (2000). CDC Report. see U.S. Department of Health andHuman Services.

Chaloupka, F. J. (1991). Rational addictive behavior and cigarette smoking. Journal ofPolitical Economy 99: 722–742.

Chaloupka, F. J. and Saffer, H. (1988). The demand for cigarettes and restrictions onsmoking in the workplace. National Bureau of Economic Research, Working paperNo. 2663.

Chaloupka, F. J. and Warner, K. E. (2000). The economics of smoking. In A. J. Culyerand J. P. Newhouse (eds), Handbook of Health Economics, Vol. 1, Amsterdam:North-Holland, pp. 1539–1627.

Chaloupka, F. J. and Wechsler, H. (1995). Price, tobacco control policies and smokingcontrol among young adults. National Bureau of Economic Research, Working paperNo. 5012.

Clements, K. W., McLeod, P. B. and Selvanathan, E. A. (1985). Does advertising affectdrinking and smoking? Working Paper, The University of Western Australia.



Cohen, J. E., Sarabia, V. and Ashley, M. J. (2001). Tobacco commerce on the Internet:a threat to comprehensive tobacco control. Tobacco Control 10: 364–367.

Conniffe, D. (1995). Models of Irish tobacco consumption. The Economic and SocialReview 26: 331–347.

Cooper, Heather (1990). Symposium notes: tobacco litigation: a comparative analysis ofthe United States and European Community approaches to combating the hazardsassociated with tobacco products. Brooklyn Journal International Law XVI: 275–303.

Cox, H. and Smith, R. (1984). Political approaches to smoking control: a comparativeanalysis. Applied Economics 16: 569–582.

Czart, C., Pacula, R. L., Chaloupka, F. J. and Wechsler, H. (2001). The impact of pricesand control policies on cigarette smoking among college students. ContemporaryEconomic Policy 19: 135–149.

DeCicca, P., Kenkel, D. and Mathios, A. (2002). Putting out the fires: Will higher taxesreduce the onset of youth smoking? Journal of Political Economy 110: 144–169.

Dee, T. (1999). The complementarity of teen smoking and drinking. Journal of HealthEconomics 18: 769–793.

Doroodian, K. and Seldon, B. J. (1991). Advertising and cigarette consumption. EasternEconomic Journal 17: 359–366.

Douglas, S. (1998). The duration of the smoking habit. Economic Inquiry 36: 49–64.Duffy, M. (1996a). Econometric studies of advertising, advertising restrictions and

cigarette demand: a survey. International Journal of Advertising 15: 1–23.Duffy, M. (1996b). An econometric study of advertising and cigarette demand in the

United Kingdom. International Journal of Advertising 15: 262–284.Eckard, E. (1991). Competition and the cigarette TV advertising ban. Economic Inquiry 29:

119–133.Farr, S. J., Tremblay, C. H. and Tremblay, V. J. (2001). The welfare effect of advertising

restrictions in the U.S. cigarette industry. Review of Industrial Organization 18:147–160.

Food and Agriculture Organization of the United Nations (UNFAO) (2003). Projections ofTobacco Production, Consumption and Trade to the Year 2010. Rome: FAO.

Gallet, C. (2003). Advertising and restrictions in the cigarette industry: evidence ofstate-by-state variation. Contemporary Economic Policy 21: 338–348.

Gallet, C. (2004). The efficacy of state-level anti-smoking laws: demand and supplyconsiderations. Journal of Economics and Finance 28: 404–412.

Gallet, C. and Agarwal, R. (1999). The gradual response of cigarette demand to healthinformation. Bulletin of Economic Research 51: 259–265.

Gallet, C. and List, J. (2003). Cigarette demand: a meta-analysis of elasticities. HealthEconomics 12: 821–835.

General Accounting Office (GAO) (2002). Internet Cigarette Sales: Giving ATFInvestigative Authority May Improve Reporting and Enforcement, GAO-02–743,August, Washington, DC.

Goel, R. K. (2004). Cigarette demand in Canada and the U.S.-Canadian cigarette smug-gling. Applied Economics Letters 11: 537–540.

Goel, R. K. and Morey, M. J. (1995). The interdependence of cigarette and liquor demand.Southern Economic Journal 62: 451–459.

Goel, R. K. and Nelson, M. A. (2004). International patterns of cigarette smoking andglobal anti-smoking policies. Journal of Economics and Finance 28: 382–394.

Goel, R. K. and Nelson, M. A. (2005). Tobacco policy and tobacco use: differences acrosstobacco types, gender and age. Applied Economics 37: 765–771.

Goel, R. K. and Ram, R. (2004). Quantile-regression estimates of cigarette demandelasticities for the United States. Journal of Economics and Finance 28: 413–421.

Grossman, M., Sindelar, J., Mullahy, J. and Anderson, R. (1993). Policy watch: alcoholand cigarette taxes. Journal of Economic Perspectives 7: 211–222.

352 GOEL AND NELSON


Gruber, J. and Koszegi, B. (2001). Is addiction ‘rational’? theory and evidence. QuarterlyJournal of Economics 116: 1261–1303.

Gruber, J., Sen, A. and Stabile, M. (2003). Estimating price elasticities when there issmuggling: the sensitivity of smoking to price in Canada. Journal of HealthEconomics 22: 821–842.

Guindon, G. E., Tobin, S. and Yach, D. (2002). Trends and affordability of cigarette prices:ample room for tax increases and related health gains. Tobacco Control 11: 35–43.

Hamilton, J. L. (1972). The demand for cigarettes: advertising, the health scare, and theadvertising ban. Review of Economics and Statistics 54: 401–411.

Houthakker, H. S. and Taylor, L. D. (1970). Consumer Demand in the United States1929–1970: Analyses and Projections. Cambridge: Harvard University Press.

Hu, T., Sung, H. and Keeler, T. (1995). The state antismoking campaign and the industryresponse: the effects of advertising on cigarette consumption in California. AmericanEconomic Review 85: 85–90.

Jha, P. and Chaloupka, F. (eds) (2000). Tobacco Control in Developing Countries, Oxford,UK: Oxford University Press.

Johnson, L. W. (1986). Advertising expenditure and aggregate demand for cigarettes inAustralia. International Journal of Advertising 5: 45–58.

Keeler, T., Hu, T., Barnett, P. and Manning, W. (1993). Taxation, regulation, andaddiction: a demand function for cigarettes based on time-series evidence. Journal ofHealth Economics 12: 1–18.

Keeler, T., Hu, T., Barnett, P., Manning, W. and Sung, H. (1996). Do cigarette producersprice-discriminate by state? An empirical analysis of local cigarette pricing andtaxation. Journal of Health Economics 15: 499–512.

Kmenta, J. (1971). Elements of Econometrics. New York: Macmillan.Lance, P. M., Akin, J. S., Dow, W. H. and Loh, C. P. (2004). Is cigarette smoking in

poorer nations highly sensitive to price? Evidence from Russia and China. Journal ofHealth Economics 23: 173–189.

Lanoie, P. and Leclair, P. (1998). Taxation or regulation: looking for a good anti-smokingpolicy. Economics Letters 58: 85–89.

Laugesen, M. and Meads, C. (1991). Tobacco advertising restrictions, price, income, andtobacco consumption in OECD countries. 1960–86. British Journal of Addiction 86:1343–1354.

Leu, R. E. (1984). Anti-smoking publicity, taxation, and the demand for cigarettes. Journalof Health Economics 3: 101–116.

Lewit, E. M., Coate, D. and Grossman, M. (1981). The effects of government regulationon teenage smoking. Journal of Law and Economics 24: 545–569.

Mahood, G. (1995). Canadian tobacco package warning system. Tobacco Control 4: 10–14.Manning, W., Keeler, E., Newhouse, J., Sloss, E. and Wasserman, J. (1989). The taxes of

sin: Do smokers and drinkers pay their way? Journal of American Medical Association261: 1604–1609.

Marks, L. (1982). Policies and postures in smoking control. British Medical Journal 284:391–395.

Max, W. (2001). The financial impact of smoking on health-related costs: a review of theliterature. American Journal of Health Promotion 15: 321–331.

Nelson, J. P. (2003). Cigarette demand, structural change, and advertising bans:International evidence: 1970–1995. Contributions to Economic Analysis and Policy 2:1–27.

Ohsfeldt, R. L. and Boyle, R. G. (1994). Tobacco excise taxes and rates of smokelesstobacco use in the U.S.: an exploratory ecological analysis. Tobacco Control 3:316–323.

Ohsfeldt, R., Boyle, R. and Capilouto, E. (1997). Effects of tobacco excise taxes on theuse of smokeless tobacco products in the USA. Health Economics 6: 525–531.



Orzechowski and Walker (2000). The Tax Burden on Tobacco: Historical Compilation 1999.Arlington, VA: Orzechowski and Walker.

Pechmann, C. and Ratneshwar, S. (1994). The effects of antismoking and cigarette adver-tising on young adolescents’ perceptions of peers who smoke. Journal of ConsumerResearch 21: 236–251.

Perkurinen, M. (1989). The demand for tobacco products in Finland. British Journal ofAddiction 84: 1183–1192.

Radfar, M. (1985). The effect of advertising on total consumption of cigarette in the U.K.European Economic Review 29: 225–231.

Reinhardt, F. S. and Giles, D. E. A. (2001). Are cigarette bans really good economicpolicy? Applied Economics 33: 1365–1368.

Roemer, R. (1993). Legislative Action to Combat the World Tobacco Epidemic, 2nd edn.Geneva: World Health Organization.

Saffer, H. and Chaloupka, F. (1999). Tobacco advertising: economic theory andinternational evidence. NBER Working Paper 6958, February.

Schmalensee, R. (1972). The Economics of Advertising, Amsterdam: North Holland.Seldon, B. and Doroodian, K. (1989). A simultaneous model of cigarette advertising:

effects on demand and industry response to public policy. Review of Economics andStatistics 71: 673–677.

Stavrinos, V. G. (1987). The effects of an anti-smoking campaign on cigarette consump-tion: empirical evidence from Greece. Applied Economics 19: 323–329.

Stewart, M. J. (1993). The effect on tobacco consumption of advertising bans in OECDcountries. International Journal of Advertising 12: 155–180.

Sung, H. Y., Hu, T. W. and Keeler, T. E. (1994). Cigarette taxation and demand: anempirical model. Contemporary Economic Policy 12: 91–100.

Tansel, A. (1993). Cigarette demand, health scares and education in Turkey. AppliedEconomics 25: 521–529.

Tegene, A. (1991). Kalman filter and the demand for cigarettes. Applied Economics 23:1175–1182.

Townsend, J. L. (1987). Cigarette tax, economic welfare, and social class patterns ofsmoking. Applied Economics 19: 355–365.

Tremblay, C. and Tremblay, V. (1995). The impact of cigarette advertising on consumersurplus, profit, and social welfare. Contemporary Economic Policy 13: 113–124.

United Nations (2003). Projections of Tobacco Production, Consumption and Trade to theYear 2010. Food & Agriculture Organization of the United Nations.

US Department of Health and Human Services (2000). Reducing Tobacco Use: A Report ofthe Surgeon General. Atlanta, GA: Centers for Disease Control and Prevention.

Valdes, B. (1993). Cigarette consumption in Spain: empirical evidence and implications forpublic health policy. Applied Economics 25: 149–156.

Viscusi, W. K. (1994). Cigarette warnings: the perils of the Cipollone decision. SupremeCourt Economic Review 2: 239–275.

Viscusi, W. K. (1995). Cigarette taxation and the social consequences of smoking. InJ. Poterba (ed.) Tax Policy and the Economy, Vol. 9. Cambridge, MA: MIT Press,pp. 51–101.

Viscusi, W. K. (2002). Smoke-Filled Rooms: A Postmortem on the Tobacco Deal. Chicago:University of Chicago Press.

Wakefield, M. A., Wilson, D., Owen, N., Esterman, A. and Roberts, L. (1992). Workplacesmoking restrictions, occupational status, and reduced cigarette consumption. Journalof Occupational Medicine 34: 693–697.

Warner, K. E. (1990). Tobacco taxation as health policy in the Third World. AmericanJournal of Public Health 80: 529–531.

Wasserman, J., Manning, W., Newhouse, J. and Winkler, J. (1991). The effects of excisetaxes and regulations on cigarette smoking. Journal of Health Economics 10: 43–64.

354 GOEL AND NELSON


Wilcox, G. B., Tharp, M. and Yang, K. T. (1994). Cigarette advertising and consumptionin South Korea, 1988–1992. International Journal of Advertising 13: 333–346.

Witt, S. F. and Pass, C. L. (1981). The effects of health warnings and advertising on thedemand for cigarettes. Scottish Journal of Political Economy 28: 86–91.

World Bank (1999). Curbing the Epidemic: Governments and the Economics of TobaccoControl. Available at www.worldbank.org/tobacco/reports_pdf.htm

World Health Organisation (WHO) (1996). The Tobacco Epidemic: A Global Public HealthEmergency, Fact Sheet N118, May. Available at www.who.int/inf-fs/en/fact118.html;www.who.int/archives/tohalert/apr96/intro.htm

World Health Organisation (WHO) (2002). The Tobacco Atlas. London: The HanwayPress.

Yurekli, A. A. and Zhang, P. (2000). The impact of Clean Indoor-Air Laws and cigarettesmuggling on demand for cigarettes: an empirical model. Health Economics 9:159–170.

Zanias, G. P. (1987). The demand for cigarettes: habit formation and health scare. GreekEconomic Review 7: 248–262.



anti smoking

Documents

nonprice measures

checking smoking

costs of smoking

low price elasticity

developing countries

various measures

andworkplace smoking

world health assembly