Article Text

other Versions


The ethics of a smoking licence
  1. Daniel Halliday
  1. Correspondence to Dr Daniel Halliday, Department of Philosophy, University of Melbourne, Melbourne, VIC 3010, Australia; daniel.halliday{at}

Statistics from


Many governments implement policies that officially aim to discourage tobacco use among their citizens. Much of this effort is realised through attempts to better educate people about the effects that tobacco has on health. Most familiar are the increasingly graphic warnings displayed on cigarette packets. Other policies combat the activities of the tobacco industry, for example, restrictions on advertising. It is widely accepted, however, that it is not enough to simply educate users and coerce industry. Some amount of coercion must also be imposed on peoplei to make them smoke less or not start. So, governments often attach substantial consumption taxes to tobacco products. Almost always, this is implemented as sales tax, or VAT: Each time someone buys a tobacco product, they are required to pay a certain extra amount. The idea is that people will respond by smoking less, quitting smoking or not starting in the first place. But sales tax is not the only consumption tax available. The licence approach, as I shall call it, is a plausible alternative. Its main distinguishing feature is that users are forced to pay a relatively large amount of tax before being allowed to make any tobacco purchases.

In this paper, I am going to explore some of the moral considerations relating to smoking licences. And I shall offer a limited defence of licences as a replacement for sales tax on tobacco products. This defence will include some moral arguments in favour of one particular licence design over others.

What follows is, in at least some respects, an attempt at doing non-ideal theory. Very roughly, ideal theory attempts to give an account of what an ideally just society would look like, with relatively little attention paid to contingent facts about society. Non-ideal theory seeks to develop proposals about how to cope with problems that ideal theory sometimes assumes away in the pursuit of a more general and abstract treatment of what justice and related concepts are ultimately about.ii From the perspective of ideal theory, smoking licences might not be a very interesting topic. There exist different conceptions of what an ideally just society would look like. Some of these might ban tobacco altogether, others might not restrict it at all and some might restrict it in ways not involving tax policy. However, the fact is that consumption taxes on tobacco are a fairly entrenched part of regulation in many actual societies. The use of such taxes could, from a moral point of view, probably be improved. It is part of non-ideal political theory to respond to such facts by exploring what sort of changes might be morally defensible and by drawing attention to which unanswered questions might be most urgent or morally significant.

Doing non-ideal theory requires one to take seriously the actual effects of any proposed policy change. I can only do this to a limited degree. Some considerations (to do with, say, enforcement) might take on a greater significance for licensing than for standard sales tax approaches. But I suspect that, like many distinctions in philosophy, ‘ideal versus non-ideal’ does not have a strongly binary form and can be viewed as a spectrum of positions between two extremes. It is perfectly possible to motivate the sort of theorising that might lie somewhere in the middle ground, even if one would like to get closer to some conception of the non-ideal extreme later on.

Here is how the paper proceeds. In the section ‘Background: taxes and justice’, I will make a few background points about taxation and justice. This will help me locate my proposal within a set of more abstract claims about taxes and justice, register a few assumptions and distinguish my proposals from other philosophically interesting positions on tobacco taxation. In the section ‘The general case for a smoking licence’, I will spell out more fully the general idea of a tobacco licence and a basic principle that supports it. The section ‘Refining the licence proposal: options and challenges’ compares different refinements of the licence approach. The section ‘Conclusion’ concludes.

Background: taxes and justice

Taxation can be used for various different purposes, each of which might require its own moral justification. A classic function of taxation is to ensure the supply of public goods like national defence and coastal flood barriers. Since these cannot be supplied by a free market, the state can permissibly force people to pay through taxation. A second function of taxation lies in its capacity to realise principles of fairness or distributive justice. Taxation provides a way of transferring goods from better-off to worse-off individuals, realising certain distributive outcomes. And a third function of taxation relies on the idea that it can be permissible for a state to coerce its citizens for their own good. By taxing undesirable activities, the state can discourage these activities. This sort of paternalism is most easily pursued through consumption taxes, which (unlike, say, income tax) can target quite specific sorts of behaviour.

In this paper, I am going to rely only on the assumption that tobacco tax policy is a permissible instance of the third (paternalistic) function. I will say more about this assumption below. I will first quickly distinguish my own approach from arguments about tobacco policy that invoke either of the first two types of justification. First, it is often said that clean air is a public good. The state can therefore permissibly coerce people to smoke less in order to maintain air quality. This argument may not actually support any claims about taxation, as opposed to other coercive policies, like banning smoking in public places. Second, there are arguments for tobacco taxes that invoke some principle of distributive justice. Consumption taxes can force consumers to internalise the costs of their consumption: A tax on tobacco purchases could fund healthcare for smoking-related diseases, reducing the need to use revenue from income tax paid by non-smokers. It might be considered fairer for smokers to pay more than non-smokers to fund the sort of healthcare they often need later in life. However, such claims are controversial, and the general principles behind them are often said to have implausible implications.iii I set these controversies aside by not locating any principles about fairness in the general justification for a smoking licence. However, considerations about fairness may still have a substantial role to play when making any proposal more specific and determinate. Given this, I will have something to say about how considerations of fairness bear on narrower questions about licence design.

One might doubt whether paternalistic justifications for consumption tax have any force in the first place. Let me note a few points in response to this. First, it is a standard view in the philosophy of healthcare ethics that mere education of individuals is not enough when pursuing better health outcomes. Philosophers, that is, accept that some provision of stronger incentives is necessary.iv Coercive taxation continues to be recommended by expert bodies on public health: In spite of the progress that has been made through non-coercive measures (like health warnings on cigarette packs), the World Health Organisation still maintains that “increasing taxes is the best way to reduce tobacco use”.v Whatever one feels about paternalistic justifications for taxation in general, tobacco may be a special case. First, much tobacco consumption occurs in persons under the age at which smoking is legally permitted (see below). Second, tobacco is profoundly addictive, and addictive behaviour can be hard to change through non-coercive means alone. Finally, there is the sheer scale of tobacco-related harm. It has been reported that cigarette use accounted for 100 million deaths during the 20th century and will cause many more deaths in the 21st century if current consumption trends The relationship between tobacco and cancer is familiar.vii Less well known is that tobacco consumption is correlated with various other burdensome conditions. These include peptic (ie, stomach) ulcers, erectile dysfunction, bronchitis, glaucoma and emphysema. A substantial reduction in tobacco use would be among the best health outcomes that we could secure.viii One could coherently support paternalistic policy about tobacco consumption, while remaining sceptical about paternalistic policy elsewhere.

The general case for a smoking licence

Licences, as consumption taxes, are rarely used by governments.ix There are few widespread examples besides annual road taxes and related vehicle registration taxes. One explanation might be that they carry bureaucratic operating costs that sales taxes do not. But the rarity of licences need not be evidence of any good moral argument for their being inferior to sales taxes. Indeed, there is a good positive case for using licences to combat at least some cases of harmful consumption. This begins with the following idea: The Temporal Distribution Principle When an individual's behaviour is subject to temporally delayed costs, a defensible means of discouraging it will (other things being equal) impose a bulk of new costs at or close to the time of the behaviour's onset.

This principle is adapted from some observations made by the economists Julian Le Grand and Divya Srivastava.13 They argue that when identifying ways to discourage costly behaviour, policy should take seriously the way that such behaviour's costs are distributed over time. Sales taxes make consumption more costly only by distributing extra cost across the various points in time at which units of consumption are purchased. Licences, by contrast, impose a heavy cost at the point where consumption begins and (depending on their design) at certain subsequent points where a decision is made to continue such consumption. So, licence policies mean tobacco users must absorb a large financial burden just to become consumers of tobacco. Sales taxes have no such advantage. In short, then, tobacco licences satisfy the temporal distribution principle while sales taxes do not.

The argument so far might look more like an appeal to effectiveness than like a more normative justification for licences. In response, one might say that the effectiveness of any paternalistic policy is not easily separable from its moral justification. Paternalistic action often involves imposing some sort of burden on its subject, the justification for which depends on the role this imposition plays in securing some appropriate large benefit to the subject. The point of the appeal to the effectiveness of licences is that they can be expected to make these interventions more proportionate. If sales taxes are not working as well as licences would, then they impose a burden that enjoys less moral justification. If licences do more to reduce tobacco-related harms than sales taxes do, then they may count as more proportionate in this sense.x In my view, however, the normative justification for licence proposals becomes most apparent when we examine which groups they aim to target.

Tobacco users often begin their consumption during adolescence. Paternalistic policies can often be evaluated partly as to whether they aim at this group of consumers. There are two reasons why tobacco licences make it especially difficult for consumption to begin at such a young age. First, the average minor has a poorer cash flow than the average adult. Adolescents find it relatively easy to absorb the costs of sales tax on cigarettes since they are often in a position to absorb small financial burdens frequently, but typically less able to absorb larger burdens less frequently. This fact has already been used to develop useful antitobacco policy, namely, legislation against the sale of cigarettes in very small quantities, which the tobacco industry formerly used to attract child customers (p. 78).10 Perhaps unsurprisingly, data suggest that price increases in tobacco tend to particularly lower consumption levels among the young.14–,16 As to why paternalism is easier to justify when it targets minors, various explanations might be given. The preferences of minors might, for example, be viewed as less authentic or authoritative than those of adults. More narrowly, studies of adolescent smoking suggest that the motivation to consume tobacco is enhanced by various peer group norms that might be less present once adolescents reach legal adulthood.16 More generally, appealing to the way in which licences target minors is to make an uncontroversial appeal to the general view that there are special reasons for using paternalism against this group.

Second, licences offer greater assurance that the prohibition of sales to minors will actually be respected. On the sales tax approach, society relies largely on sellers to ensure that the consumers of tobacco are of legal age. However, sellers naturally have an incentive to make sales and may be inclined to sell when there is some doubt as to whether a consumer is old enough. An advantage of the licence approach is that it relieves society of placing such heavy reliance on sellers. Whether a licence is issued to a particular applicant is decided by government officials who have no financial incentive to maintain tobacco sales. Moreover, the process of gaining a licence may be such that a greater burden of proof can be placed on an applicant who needs to maintain that they are of adult age. Licence applications may require copies of identification that it is impractical to expect sellers to ask for, even if tobacco purchasing requires some sort of proof of age. Any policy on licences may require that a consumer produces their licence when making a purchase. Given this, the existence of licences makes it harder for a seller, if caught supplying a minor, to rely on the defence that precise visual judgements of age are hard to make. Since licences can be designed such that they need to be electronically scanned as part of recording a sale (see below), sellers who approve sales without scanning a licence thus expose themselves to a further risk of being found out. The sales tax approach offers neither of these benefits. Indeed, sales taxes accommodate no interesting distinction between minors and adults.

This leaves the question of what to say about adult smokers. It would be fallacious to argue that since licences would reduce adolescent tobacco use, they would (therefore) comparably reduce adult use. The fact that most smokers actually began using during adolescence does not entail that, if prevented, they would not have still taken up tobacco consumption later on. But there is some reason to believe that being able to access tobacco during adolescence may be decisive as to whether one smokes as an adult. Barring access for just a few years may be enough to outlive much of the youthful motivation to begin using tobacco, particularly if adults are at least somewhat less exposed to (or disposed to act on) peer pressure. But even if a licence requirement would merely delay the onset of tobacco use by a few years or less, such a delay might still be valuable. For one thing, less time spent consuming tobacco, especially during a period of important education and physiological development, still has a positive impact on health. There is evidence that smoking during high school is correlated with dropping out14 (one must, however, be cautious about inferring anything about the direction of causation in such cases). Apart from anything else, starting tobacco use later in life may just mean that one spends less time having one's health negatively impacted by smoking.xi

The case for a smoking licence does not rely wholly on considerations about reducing adolescent uptake; it merely draws a large part of its strength from it. Of course, one might think this an advantage if one thinks that paternalism against adults is harder to defend. To the extent that paternalistic tobacco policies represent a defensible treatment of adult users, the temporal distribution principle can be applied to them, too: Adults still find it harder to absorb financial burdens in advance of some activity than to absorb these burdens gradually, over a period of time.

Refining the licence proposal: options and challenges

Various licence designs are broadly compatible with the general case already outlined. Many of these differences in licence design are particularly interesting from a moral point of view, and so too are some features that are common across designs. In this section, I will try to highlight, and offer some remarks about, what I take to be some of the most important questions and concerns.

Duration of consumption versus quantity of consumption

There are two ways of defining the lifespan of a smoking licence. One is to adopt the road tax model: Smokers buy a licence that allows them to smoke for a fixed period of time, perhaps a year. The price of the licence is the same, irrespective of how much tobacco a user will go and buy with it. The temporal model represents the received view about how to design a smoking licence, being the one proposed by Le Grand and Srivastava in drafts of their report on incentives and health policy, and more recently by Simon Chapman.17

Another model attaches a licence's life to some fixed amount of consumption. On this ‘absolute consumption’ model, a licence might allow its owner to buy, say, a thousand packets before expiring. On this design, the rate of consumption is more of a factor: Heavy smokers will have to buy licences more frequently, whereas occasional smokers will tend to have licences that last longer. The absolute consumption model, then, is essentially like an accelerated sales tax: How much one pays is directly proportional to how much one consumes, but the payment must still occur in bulk, prior to the amount of consumption in question.

The absolute consumption model may, I think, be morally superior to the temporal duration model. At least, it might be fairer: It preserves some link between how much tax a user pays and how much tobacco they consume. A temporal model is likely to distribute the consumption tax burden unevenly. Furthermore, the price of a licence might be less arbitrary on the absolute consumption model than on the temporal one. I make no claim about what the optimal licence price should be. This depends partly on empirical questions about how the deterrent effect varies with this amount. But one advantage of a licence based on absolute consumption is that its price could be fixed by making it the same as the current amount of sales tax levied on the relevant number of units. Even if the current level of sales tax is arbitrary, it might not be arbitrary to retain that level when transitioning to a licence, if only because this would guarantee that the licence does not smuggle in a higher tax burden. With a temporal licence, this sort of transition is not available because it is not known how many units a user will purchase. This is in addition to the fact that whatever price is chosen a temporal licence will inevitably make smoking cheaper for very heavy smokers and more expensive for occasional smokers.

The absolute consumption model might be more technologically demanding than the temporal alternative. Licences will need to be equipped with some device that can help record the number of sales and alert a seller when the limit has been reached. Such technology is nowadays not particularly expensive or difficult.xii Which model should be adopted ultimately depends, I think, on precisely what is said in response to some of the considerations and objections discussed in the remaining paragraphs.

Staggered licences

Another possibility is to ‘stagger’ the licence. In simple terms, the idea here is that a smoker's first licence will cost more than subsequent ones. There are a variety of ways in which staggering might be introduced. A ‘beginner’ smoker might have to pay a supplementary fee if they have not purchased a licence before. Alternatively, beginners may have to commit to buying a licence with a longer life (in terms of greater temporal duration or an entitlement to more units), and hence greater price.

A general reason for staggering is that it might enhance the sense that a licence targets potential smokers, and especially minors, to a greater extent than it targets existing adult smokers. Political realities are such that licences stand to be quite unpopular if not offered in ways that do something to placate existing smokers. This is why it might be important to propose licences as a complete replacement for sales tax rather than an additional burden to be added alongside sales taxes. People tend to react angrily (if fallaciously) to being taxed ‘twice’ even if they have not incurred a larger burden overall.xiii So, staggering may have a real political advantage, as well as providing a way of deterring potential smokers yet more.

Staggering could be made quite extreme: We might, for example, grant smokers a permanent licence once they have been purchasing licences for some number of years. Alternatively, persons over some age might be made exempt altogether. This possibility might draw its support from the fact that paternalistic policies can make people feel oppressed or belittled, especially when they feel they are being forced to ask permission to do something that they have been doing for years. The fact that someone would prefer not to be engaging in addictive behaviour may not change the fact that they feel hounded or stigmatised by licence-style coercion than by standard sales tax. If entrenched adult smokers are burdened less by the cost of a licence, then staggering might thereby further its feasibility.xiv

I am hesitant to take a definite stance on staggering. Certain designs may introduce certain problems, particularly if an older section of the population is able to obtain tobacco more cheaply than a younger section, due to obvious incentives for illicit selling (see further discussion below). But it should be emphasised that if there is any sort of advantage to be gained from any sort of staggering design, then this would add to the general case for licences over sales taxes since sales taxes present no opportunity for staggering of any sort.

The psychology of sunk costs

People's tendency to honour ‘sunk costs’ in practical reasoning indicates a problem with the licence approach: Once an individual has paid for a licence, they may believe that they have gained a new reason to make purchases so as to ‘get their money's worth’. Similarly, smokers who want to quit may feel like postponing any effort until their licence expires.

Clearly such implications are undesirable. I do not deny that the licence approach faces a problem here that sales taxes do not. The important question is how serious this problem is. It should be remembered that sunk costs are only incurred by those who actually buy a licence in the first place. Since a licence can be expected to reduce the number of smokers, the value of this overall reduction in smokers might outweigh the disadvantage of having a more highly motivated remaining group of smokers. Another point is that the licence approach does not impose a single sunk cost, but rather a series of such costs each time a licence needs to be renewed. While there might be a perverse incentive to continue to smoke once a licence has been purchased, a smoker will later face the decision of whether to incur another large cost once their licence expires. In other words, the cost of purchasing a licence is not wholly submerged: There is a sense in which the cost can be limited by not buying another licence after the expiration of an existing one. More precisely, the fact that licences expire means that there is necessarily a limit on the extent to which smokers can continue to purchase cigarettes as a way of honouring the cost of the licence.

None of the above points wholly dispense with the worry about sunk costs. I do not have any complete answer to it. The problem might prove to be useful in another way: It is possible that the facts about how people are motivated by sunk costs could prove valuable, or even decisive, in selecting between the time model and the duration model described above: One way of incurring a sunk cost is to have paid to make some consumption possible for a certain amount of time, another is to have paid to make a certain absolute quantity of consumption possible. If one of these sunk costs tends to have more motivational force than the other, then this might tell us which of the time or the consumption model has the greater tendency to motivate licence holders to purchase more than they would on the sales tax model. Since the temporal model means that a licence grants a user a limitless amount of purchasing, it may be more vulnerable to concerns about the psychology of sunk costs.

Cooperative purchasing

Another worry is that licences might be ‘shared’: Individuals could pool funds so that a single licence can be purchased at a lower cost per person. This reduces the capacity of licences to temporally redistribute the financial costs of tobacco use. The attractiveness of such cooperation depends partly on the possibility that a relatively large number of cigarette packs could be purchased per licence. This can be ruled out. Recall the idea, already mentioned, that licences may be designed such that they may be electronically scanned whenever tobacco is sold to the licence holder. This allows some record to be made of how many purchases have been made on that licence. Licences could be designed such that they could only be scanned, say, for a maximum of three or four packs of cigarettes per day, or even per week.xv

Reflection on cooperative purchasing also indicates a further advantage of the absolute consumption model over the temporal model. When the life of a licence is linked to how many cigarettes can be bought with it, it will tend to run out more quickly if used to supply multiple users. This means that if licences are linked to quantity of purchasing, then cooperators will not, overall, save any money by sharing a licence. They will pay less per licence but licences will last less long. What cooperation still does, however, is restrict the degree to which smoking is made subject to large infrequent financial burdens. In other words, cooperation limits the temporal redistribution of costs. This is still a problem that sales taxes do not have. My best response is merely that it is a cost that licence design can mitigate, and that the worry may have less force on the absolute consumption model than the temporal alternative.

Black markets

If licences make it genuinely harder for people to purchase cigarettes, then this creates greater demand for cheaper illegal selling. Black markets are bad because they threaten to keep tobacco consumption high, undoing the work that tobacco taxes aim to achieve. What's more, a growth in black market activity diverts revenue away from the state and often funds quite heinous sorts of criminality. It can also give rise to police corruption (p. 49).20 Black markets in cigarettes have in fact reversed attempts to raise cigarette taxes on certain occasions in the past. After the Canadian government began to raise cigarette taxes in the late 1980s, they were led to cut them in response to a growth in cigarette smuggling from the USA, where cigarettes could be bought more cheaply. If licences were to create a greater demand for illegal supply than sales taxes, then this might count against the licence proposal.

In response, some general points limit the case for worrying about black markets. First, the Canadian case is perhaps unusual. It involved a long, relatively porous border (at least at the time) that had a very large supply of cheap cigarettes on the other side of it. In other parts of the world, smugglers may not enjoy such favourable conditions. Moreover, black markets are typically subject to serious capacity limits: When someone attempts to sell something illegally, transportation and storage can be difficult to arrange. Indeed, studies of smoking rates following the Canadian tax cuts have suggested that the raised sales taxes were in fact working in spite of the substantial level of smuggling, and that smoking may have declined at a greater rate had they been retained.21 So, it is not obvious whether black markets are really able to sustain the same levels of supply and consumption as the ‘white’ markets that they aim to replace.xvi Arguably, it is not taxation as such that stimulates the largest black markets, but rather discrepancies between tax policies in different geographic regions.xvii So, the objection's force may be highly contingent upon which region is under question.


Paying a consumption tax typically means foregoing some other sort of purchases. The poorer a consumer is, the more important these foregone purchases are likely to be. In this way, consumption taxes tend to be regressive—they burden poor consumers more than rich ones, other things being equal. Since regressivity tends to be a feature of consumption taxes in general, it presents a problem for sales taxes and licences alike. But given that poor people suffer from problems of cash flow and not just a sheer lack of resources, licences may discriminate to some degree more than sales taxes do. Taxes aside, there is also evidence that poor people face various non-financial burdens in adopting a non-smoking lifestyle, at least in certain cultures. This counts as further evidence that antismoking policy risks being antiegalitarian.22

Granted, discrimination is objectionable, and society has too much of it already. One response is to dispute the claim that poor smokers will simply divert their money to pay for licences, rather than just respond in the desired way, by not smoking. Le Grand and Srivastava cite some data that they think supports this conclusion.13 Others make similar claims (p. 23)10 (p. 3)17. Even if this is view right, there may still be an objectionably large number of poor people who end up disadvantaged by committing funds to licence purchases. Less aggressively, Robert Goodin suggests that discrimination worries need not be met by an internal solution, that is, one within antismoking policy.xviii The problem of disadvantaging the poor instead needs to be met by other policies, such as a more progressive income tax, free education, guarantees of sick pay for unskilled jobs and so on. Granted, continued appeals to such ‘external’ solutions can just trivialise the proposal being defended because the proposal ends up not doing any work. And the force of such appeals relies heavily on the feasibility of other policies that will really address background inequalities rather than just the theoretical availability of such policies (which might be fairly uncontroversial). That said, there might be some scope for elements of a solution that do fall within policy on tobacco use. Proctor, for example, suggests that the state could subsidise cessation programmes, which might issue free nicotine patches (p. 552).10

Discrimination does not just occur when some group is economically disadvantaged. Imposing a burden sometimes conveys an evaluative message. For example, physical punishments (like imprisonment) arguably express a kind of disapproval that cash fines do not.23 There is good evidence to suggest that tobacco control has had a similar effect, particularly when it stigmatises smokers as a group in their own right. I take this concern seriously, but my sense is that licences do not express any new condemnation of smokers that is not carried by sales tax or other existing policy. After all, applying for a licence is basically an invisible action. This contrasts with policies that, for example, confine smokers to labelled areas in public places, such as glass cubicles in airports. These public ‘shamings’ do rather more to stigmatise smokers than any taxes do.xix Nevertheless, it would be naïve to think that a tobacco licence would not make some contribution to the tendency to stigmatise smokers in ways that are now objectionably overlapping with a sort of demonisation of working class people.


Applying the temporal distribution principle to tobacco tax need not mean that it has to be applied to all other consumption taxes. Many taxed purchases do not relate to behaviour that satisfies the description in the principle. There are, of course, other sorts of behaviour that carry temporally delayed costs, notably alcohol consumption and other features of having a poor diet, as well as activities with long-term health effects, like boxing and attending noisy music events. One might worry, once the temporal distribution principle is adopted, that there is no way of consistently supporting the case for tobacco while resisting the same conclusion about many other consumption taxes.

Nevertheless, it is possible to maintain a relevant distinction between tobacco and other ‘vices’. This draws on how paternalistic policies with respect to some activity can be informed by whether or not there are healthy alternatives to that activity.xx Fast food, for example, gets much of its appeal from the fact that it is not simply cheap, but cheaper than healthier food. If this disparity were ended, then fast food would sell less (ideally this would be done by making healthy food cheaper, but the point still stands about the possibility of relying on a sales tax increase). In other words, policy that aims to reduce consumption of fast food should aim, in large part, to make it easier for people to consume healthy food instead. Many forms of unhealthy behaviour still count as some instance of a general type of activity that people need to perform. The problem is that the more ‘virtuous’ behaviours are simply harder for many people to engage in. Purchasing video games and cable television (assuming these are bad) is just cheaper than a gym membership and the occasional hiking trip. Again, the reason for resisting licences for these activities is that it is relatively easy to make the better alternatives more accessible. Crucially, the same relation to superior alternatives is not shared by cigarettes. Tobacco is just bad. It is not simply the worse option from within some broader category that people have genuinely good reason or real need to consume, like food, news information or leisure activity. While many tobacco users report that smoking has stress-relieving properties, the sincerity of these reports does not entail that they have any accuracy as an indicator of a positive feature of tobacco consumption. This is because it is plausible to treat the stress levels of smokers as being partly determined by their background nicotine dependency.28 Stress levels quickly rise after the effects of a cigarette wear off, and there is evidence that smokers are in fact more stressed overall than the average non-smoker. I should say that there is some debate here as to exactly what is shown by the relevant data.29 I lack the expertise necessary for saying anything very authoritative about this. But setting aside exactly what should be said about tobacco's supposed use as an aid against stress, the fact remains that the problem is not an overall one of getting consumers to switch to some similar but healthier alternative.


The licence approach has several advantages over the more popular way of imposing some consumption tax on tobacco, namely, taxing at each discrete sale. It begins from a plausible general principle about the temporal distribution of costs and has an attractive strength against adolescent smoking. In light of these considerations and others, I have argued that the licence approach is likely to be a superior consumption tax to sales tax. If the licence approach is to be adopted, then certain choices will have to be made between different ways of completing its design. I have tried to identify elements of what needs to considered when disambiguating the general licence proposal. I have tried to be clear, as well, about the disadvantages of licences, some of which might be mitigated by certain designs. In this way, the aim of this paper has been mainly to map out the various moral considerations and questions that are relevant to what remains a relatively underexamined type of policy. The fight against the harm done by tobacco needs to consider every available resource. If designed in the right way, licences may yet be one of the very effective and morally defensible options currently available.


Earlier versions of this paper were presented to audiences at Melbourne, Princeton and Oxford Universities. I am grateful to all who attended for the helpful feedback provided. Special thanks are due to Kristin Voigt, who delivered written comments at the Princeton session.


View Abstract


  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed.

  • i I make the following assumption about coercion: Taxes are coercive insofar as the state threatens to punish people who attempt to avoid paying them while engaging in whatever activity the state aims to tax. Policies that aim to educate users are non-coercive insofar as those who choose to ignore such policies do not thereby commit a criminal offence.

  • ii These remarks are extremely rough. For a much fuller and useful treatment of the ideal/non-ideal distinction, see Valentini1.

  • iii For discussion of this sort of fairness argument in a healthcare context, see Wikler2 and Segall (ref. 3, pp. 45–48). Anderson4 is an influential critique of egalitarian ideas about allowing persons’ choices to determine their entitlements. Compare Stemplowska5.

  • iv See, for example, Daniels (ref. 6, p. 156). Goodin (ch. 2)7 and Chapman (ch. 1)8 contain good discussions of smoking, education and other issues. For a summary of some concerns about the use of non-coercive incentives in pursuit of health outcomes, see Schmidt et al9.

  • v WHO Report on the Global Tobacco Epidemic: Warning about the Dangers of Tobacco 2011 (p. 66) available online:

  • vi See Proctor (ref. 10, pp. 5, 50–51, 549). According to Proctor, the figure of 100 million will be surpassed merely by the number of Chinese who will die prematurely from tobacco in the 21st century.

  • vii The myriad health consequences of tobacco use are laid out in the US Surgeon General's report, 2004. Available at: (accessed 26 Oct 2012).

  • viii Measuring health outcomes is a complicated matter and does not depend simply on counting deaths or occurrences of disease. For a general philosophical discussion, see Nord11. For discussion of premature death and prolonging life in particular, see Broome12.

  • ix Licences are often used for purposes that do not really count as taxation: Driving licences and hunting licences ensure that people do not engage in certain activity without proper training. They are not typically aimed at discouraging that activity as such, and need to be kept separate from the discussion here.

  • x The idea that licences are more proportionate than sales taxes might be extended another way. One view, which influences Le Grand and Srivastava,13 is that the discounting of future costs is not endorsed by our ideal selves. Policy that satisfies the temporal distribution principle might this be considered empowering rather than paternalistic. I do not rely on these claims here.

  • xi The US Surgeon General notes that smoking attributed mortality is correlated with ‘long-term previous exposure’ (p. 859 – see note above).

  • xii See the discussion of cooperative purchasing, below.

  • xiii See Murphy and Nagel (ref. 18, pp. 142–44). The unfortunate tendency to fixate on the number of times taxation occurs rather than on persons’ overall tax burden might also count against hybrid schemes in which licences are used alongside sales taxes. I do not have any immediate theoretical objection to such schemes, however. In fact, the proposals made in this paper might aid their design as well as the design of a scheme in which licences are used in the absence of sales taxes.

  • xiv Another way of increasing the popularity of a tax is to invoke ‘hypothecation’. This is what occurs when revenue from a tax is committed to a specific form of expenditure. Le Grand proposes that revenues from smoking licences be allocated to state provision of healthcare. He also provides a more general discussion of hypothecation elsewhere (ch. 11)19.

  • xv Here I follow Chapman (ref. 17, p. 2), with whom this point originates.

  • xvi It is also worth mentioning that black markets tend to be more successful under cases of absolute prohibition, that is, when there is no white market with which they need to compete—a point made by Chapman17.

  • xvii Proctor (ref. 10, pp. 52–55). Proctor also claims that black markets do worse when they sell goods that are not genuinely recreational. For this reason, he rejects the common analogy between measures to curb cigarette consumption and the failed attempt to prohibit alcohol sales in the 1920s USA (p. 557).

  • xviii Goodin (ref. 7, pp. 107–112). Note that Goodin is here responding to the discrimination raised against sales tax increases, but the appeal to external solutions may also be made by a proponent of the licence approach.

  • xix I borrow the example of airport cubicles from Chapman (ref. 8, p. 35). For more discussion of stigmatisation, see Graham24, Kim and Shanahan25, Bell et al,26 and Voigt27. Interestingly, licensing does offer a certain opportunity for ‘shaming’ since it creates a record of licence holders. Before full prohibition was practiced in the USA, a number of more moderate proposals were raised, including the possibility that alcohol would be sold only to persons who had purchased a licence, subject to the added condition that the names of new purchasers be published in local newspapers (Pittsburgh Gazette Times, 26 August 1907).

  • xx For some other differences, see Proctor (ref. 10, pp. 3, 489–505).

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles