Trading one problem for two: The case against tobacco bans

Abstract

The past two decades have seen growing calls for the “tobacco endgame.” Its advocates are united by their commitment to two ideas. First, tobacco‐related harms represent a catastrophic health emergency, and second, current tobacco‐control approaches are an inadequate response to the scale of that emergency. To endgame advocates, tobacco policy should have more ambitious goals than merely “controlling” tobacco. Instead, it should aim to bring about a smoke‐free world. While a range of different policies are included under the umbrella of the “tobacco endgame,” the most radical proposal is for a complete ban on tobacco. Its advocates argue that in addition to improving global public health, an effective ban on tobacco would also promote overall autonomy and would have important egalitarian benefits. This article critically examines these arguments for a tobacco ban. I argue that they rely on idealizing assumptions about the likely effects of a ban. Because an effective ban would require robust enforcement to control the illegal market in tobacco, it would be more likely to undermine autonomy and equality than it would be to promote them. By relying on idealizing assumptions and ignoring the likely consequences of a tobacco ban, advocates of a ban obscure, rather than clarify, both the policy debate and the ethical stakes. I conclude by considering the ways that idealizing assumptions should—and should not—play a role in debates about ethical issues in public policy.

1. INTRODUCTION

Over the past two decades, several people working in the field of tobacco control have called for more planning for the “tobacco endgame.” Endgame advocates are united by their commitment to two ideas. First, tobacco‐related harms represent a catastrophic health emergency: smoking kills more than 7 million people every year.¹ Second, current tobacco‐control approaches are an inadequate response to the scale of that emergency: more than 1 billion people worldwide are current smokers, a number than has steadily increased over the past several decades.² To endgame advocates, tobacco policy should have more ambitious goals than ‘controlling’ tobacco. Instead, it should aim to bring about a smoke‐free world.

Beyond this broad agreement, endgame advocates differ on important details and propose a range of policies.³ Some target changes to available tobacco products.⁴ Others involve stricter economic regulation of the tobacco industry.⁵ Still others involve further restricting consumer access to cigarettes.⁶ The most radical endgame proposal of all, however, is a call for a complete ban on the sale of tobacco.⁷ Not all endgame advocates endorse such a ban; in fact, most do not. But a ban certainly represents the most obvious expression of the goal of the tobacco endgame.

This article critically examines the arguments for a tobacco ban. I argue that these arguments rely on idealizing assumptions; while advocates for a ban argue that it would promote autonomy and equality, in practice, it would undermine both. By relying on idealizing assumptions and ignoring the likely consequences of a tobacco ban, advocates of a ban obscure, rather than clarify, both the policy debate and the ethical stakes.

Section 2 outlines the main arguments in defense of a tobacco ban, focusing on those that highlight the autonomy benefits of tobacco prohibition. Section 3 argues that a ban would likely lead to an illegal market, and Section 4 considers some possible policy responses to that market. Section 5 highlights the costs—in both autonomy and equity terms—of enforcing a ban. Finally, Section 6 considers the role of idealizing assumptions in policy debates and sets out a vision of the connection between idealizing ethical arguments and empirically informed policy evaluation.

2. ARGUMENTS FOR A TOBACCO BAN

Advocates for a ban make two different arguments. The first emphasizes the public health benefits of an effective ban. The end of smoking could prevent hundreds of millions of premature deaths in the coming decades.⁸ Since a ban takes the hardest line on the availability of tobacco, it might seem the most direct route to the tobacco endgame. Even very restrictive policies seem justified if they are necessary to prevent hundreds of millions of premature deaths.

This approach implicitly accepts the all‐too‐common framing of public health as a site of conflict between the value of overall population‐level welfare and a commitment to individual freedom or autonomy. While tobacco bans might restrict the freedom of smokers to choose to smoke, that restriction on their freedom is outweighed by the enormous public health benefits it would bring about.⁹

Several recent defenses of a tobacco ban, however, have developed an alternative way of framing the debate. On this view, tobacco bans can enhance, rather than restrict, individual freedom.¹⁰ It's not just that the freedom of smokers to smoke conflict with the freedom of non‐smokers to enjoy smoke‐free environment, which is an argument that has long been used to justify smoking bans in workplaces and shared public spaces. More dramatically, the argument is that a tobacco ban promotes the freedom or autonomy of smokers themselves.

These two approaches need not conflict and can be complementary. But an advantage of the autonomy argument is that it has the potential to escape the “autonomy vs. public welfare“ conflict, and so to ground tobacco bans in values that are acceptable to all parties to that debate. The appeal of the autonomy‐based defense of tobacco bans is simple and powerful. It turns what might be the central objection to bans into a reason to endorse them. But just how would a ban on tobacco promote or enhance autonomy? After all, it would use the force of law to prohibit currently permitted activities, and while these restrictions might be justified—the value of freedom competes with other values—they are nevertheless normally understood as limiting freedom.

The central insight of the autonomy argument is the fact that cigarettes are highly addictive and that most people who smoke regret ever having started and wish that they could quit.¹¹ Since an effective ban would enable smokers to quit smoking and free themselves from their addiction, it would in fact increase their overall freedom. So even if we care about smokers' freedoms, the argument goes, we should still support a complete ban on cigarettes, because such a ban would in fact “liberate humanity from the world's deadliest consumer product.”¹²

Addiction can undermine freedom at least two ways. Schmidt's focus is on the freedom of choice that depends on external options.¹³ While tobacco bans limit external options by removing the choice to smoke, Schmidt argues that restricting a person's choices can sometimes enhance their overall freedom. Smoking reduces life expectancy, causes diseases that limit physical functioning, and is expensive, and all these limit the options of those who are addicted.

This form of freedom is external: smokers have a narrower range of choices than non‐smokers do. But there is also an important internal dimension to autonomy, and addiction has long figured in philosophical debates about the value and nature of this form of autonomy.¹⁴ To have an addiction, on this view, is to be powerless over one's choices; it is to fail to be self‐governing. Addiction not only limits a smoker's external choices, it also undermines their freedom from within. In what follows, I will not distinguish “external” and “internal” freedom, and will speak of “freedom” and “autonomy” interchangeably. After all, they are closely connected; a lack of volitional autonomy undermines freedom of choice. Both forms are valuable, and both could be enhanced by an effective ban.

We can summarize the argument in favor of tobacco bans as follows. Smoking carries both enormous health costs and robs smokers of their freedom. An effective tobacco ban would therefore prevent countless deaths while restoring autonomy to smokers enslaved by addiction. Given the state's interest in promoting both public health and the autonomy of its citizens, the state would be justified, even on autonomy grounds, in pursuing a complete ban on tobacco.

I take issue with this argument in two related ways, one empirical and one normative. First, we have good reason to think that a ban would not be fully effective, and that any policy that was intended to be even partly effective would carry significant autonomy costs. Second, if we take these costs into account, it is highly unlikely that a tobacco ban could be justified on autonomy grounds.

3. THE ILLEGAL MARKET

Neither Schmidt nor Grill & Voigt claim that a ban would be effective. Instead, they argue that were it effective, a ban would be justified. In this, their arguments differ from Proctor, who claims that a tobacco ban would in fact be effective and that “enforcement … should be a trivial matter.”¹⁵ Still, to the extent that they are advocating for a ban to be adopted as policy, all three arguments depend on the assumption that a ban could be effective. So if we're considering whether a ban would be a justifiable public policy, it's worth asking whether a ban would be effective. The available evidence strongly suggests that it would not; as we'll see, that has important implications for the principled argument in defense of a ban.

Central to the argument that tobacco bans would enhance autonomy is the claim that most smokers want to quit and are prevented from doing so by the force of their addiction. But given addiction's force, it's hard to see how a legal ban on the sale and production of tobacco would bring about the tobacco endgame unless it also prevented access to illegal cigarettes. After all, if smokers are so enslaved by the power of their addiction that the prospect of an early death is not enough to get them to quit, many of them would not be deterred by a legal ban and would be willing to purchase illegal cigarettes were they available. Not all smokers, of course—no doubt some would be deterred by a respect for the law even if they were not by a concern for their health—but certainly a nontrivial number. So we should expect that a tobacco ban would generate an illegal market for cigarettes.

We can see evidence of this phenomenon in the only jurisdiction that has, thus far, implemented a tobacco ban. In 2004, the Kingdom of Bhutan banned the sale, manufacture, and distribution of tobacco. Bhutan is an ideal setting to implement a ban: it is small, mountainous, and landlocked with few border crossings. Despite these advantages, an illegal market in smuggled cigarettes arose, and smoking rates remained high 10 years after the ban was implemented.¹⁶ When the Bhutanese government closed the border with India in 2020 as a response to the COVID‐19 pandemic, the ban was lifted and tobacco was made available duty‐free as an “essential product” in an effort to reduce cross‐border smuggling.¹⁷

For more evidence that an illegal market would likely emerge, we can look to other potentially harmful and addictive drugs that governments have tried to control through prohibition. It's possible that prohibition of drugs such as opioids, cocaine, and methamphetamine reduces use of those drugs relative to what it would be were they legally available. What is clear, however, is that the global War on Drugs has conspicuously failed to achieve the drugs endgame. According to one estimate, the global illegal drugs market has an annual value between US$400 and $600 billion.¹⁸ The wholesale and retail prices for most illegal drugs have fallen steadily over the past several decades, and the United Nations Office on Drugs and Crime estimates that more than a quarter of a billion people worldwide use illegal drugs.¹⁹ The existing evidence, then, gives us reason to suspect that a tobacco ban would lead to an illegal market in tobacco.

4. POLICY RESPONSES TO THE ILLEGAL MARKET

Banning tobacco is a policy choice, and it would lead to another policy choice: What should the state do about the illegal market in cigarettes? Those who argue in favor of a ban need to answer that question. “Ban” is vague enough to be compatible with many approaches, but the most realistic scenarios involve prohibiting the commercial production, distribution, and sale of tobacco, and also taking steps to prevent the emergence of an unregulated black market. This would likely require the use of the criminal justice system. That's not to say that a ban would criminalize smoking, however. Instead, the most promising approaches would criminalize the production and distribution of tobacco, without targeting smoking or the possession of tobacco.²⁰

Illicit drug policy offers instructive examples. The Canadian province of British Columbia recently decriminalized possession of small amounts of opioids, cocaine, and methamphetamine. Those in possession of under 2.5 g no longer face arrest and do not have their drugs seized. Production and sale remain illegal and subject to criminal prosecution, as does possession of more than 2.5 g. In Portugal, drug possession is prohibited, but is treated as an administrative rather than a criminal offense. Portugal continues to target the production, distribution, and sale of illicit drugs with criminal prosecution.

It's worth noting that both British Columbia and Portugal are committed to Harm Reduction, a policy approach that acknowledges that drug use will occur and so aims to reduce the harms of drug use rather than the prevalence of drug use. So neither aims to achieve the “drugs endgame.” But short of threatening addicted smokers with jail time, these two jurisdictions represent the most plausible policy approaches to a tobacco ban.

Empirical speculation on the outcomes of specific policy approaches is premature. Still, I want to highlight two observations about the likely results of these scenarios. First, it's not clear that a tobacco ban would achieve the promised public health benefits imagined by endgame advocates of a ban; there's even a risk that a ban would make overall smoking rates worse. Second, making a ban even somewhat effective would require enforcement, and that would carry its own autonomy costs. Both are reasons to be skeptical about the putative justification of a ban.

Banning tobacco would mark a radical departure from tobacco control policies that have been remarkably effective in many countries. Global smoking rates have fallen dramatically in many countries over the past 50 years, and much of that decline has taken place in the past 20 years.²¹ Several countries have achieved very low smoking rates in absolute terms. According to data from the World Health Organization (WHO), the age‐adjusted smoking rate in Nigeria in 2020 was 3.7%, down from 9.3% in 2000. And even many countries that continue to have higher smoking rates in absolute terms have achieved drastic reductions in the past 20 years. The United Kingdom cut age‐adjusted smoking rates by more than half in the past two decades, from 37.9% in 2000 to 16.1% in 2020.²²

Tobacco control policies such as public education, advertising bans, bans on smoking in public places, and pricing controls have played an important role in these declines, leading Warner to suggest that “measured in terms of premature deaths averted, tobacco control is arguably the developed world's single greatest public health success story of the past half century.”²³ Endgame advocates argue that as effective as they have been, these policies are inadequate to bring about the true end of smoking, but it's worth considering whether abandoning such policies would in fact bring about the benefits that defenders of a ban suggest.

First, a ban would mean abandoning many existing and effective regulations, including mandatory warning labels, regulation on packaging, and limits on chemical contents. And several promising endgame proposals would go untried, including mandatory Very Low Nicotine Content cigarettes²⁴ and “sinking lid” production quotas.²⁵

Second, the most significant effect of shifting all tobacco to the illegal market would be the loss of the ability of the state to control the price of tobacco through taxes. This would mean abandoning what the WHO calls “the most effective and cost‐effective means to reduce tobacco use and encourage users to quit.”²⁶ There is extensive evidence that increases in the price of tobacco lead to reductions in tobacco consumption.²⁷ The WHO estimates that a 10% increase in the retail price of tobacco typically leads to a 4% reduction in tobacco consumption in high‐income countries and a 5% reduction in low‐ and middle‐income countries.²⁸

In the absence of direct state control over tobacco prices, the price of cigarettes might fall. A tobacco ban that led to a drop in tobacco prices would therefore risk increasing smoking rates. It's therefore possible that a tobacco ban would not achieve significant public health benefits, because it would abandon policy interventions we know to be effective.

5. THE COSTS OF A TOBACCO BAN

At the core of the pro‐ban argument we've been considering is the idea that an effective ban would enhance the autonomy of smokers and would‐be smokers. Schmidt's argument is that the value of freedom from addiction swamps the putative value of the freedom to choose to smoke.

There are, however, three different forms of freedom in the mix, rather than two. The first is the freedom to choose to smoke. The second is the freedom Schmidt emphasizes: the increased choice and volitional autonomy that come with freedom from addiction. But there is a third form of autonomy that is relevant to this debate. Proponents of a ban are no doubt right that there are more freedoms at stake than the supposed freedom to smoke, but in focusing on freedom from addiction, they overlook ways in which enforcing a ban through the criminal law would almost certainly require significantly limiting the autonomy of many thousands of people involved in the tobacco trade.

Because a ban would be likely to generate an illegal market, enforcing a ban would require efforts to eliminate that market. Any state aiming to do so effectively enough to approach the tobacco endgame would almost certainly make use of the criminal law, and in particular criminal punishment, possibly including incarceration.

Again, the criminal prohibition of other addictive drugs offers an instructive example. In 2021, more than 238,000 people were serving a sentence for a drug crime in a US state or federal prison; people convicted of a drug crime make up 14% of inmates in US state prisons and 45% of inmates in US federal prisons.²⁹ While the incarceration rate for drug crime is highest in the United States, 18% of European inmates, representing almost 270,000 people, are serving a sentence for a drug crime, more than for any other category of crime.³⁰ Even Portugal, famous for decriminalizing drug possession, uses incarceration to deter drug use.³¹ In fact, Portugal incarcerates more people per capita for drug crimes than the European average, and its drug incarceration rate (17.7% of incarcerated people are serving a sentence for a drug crime) is the median rate in Europe.³²

Globally, then, hundreds of thousands of people are incarcerated for participating in the illegal drug trade. Despite decades of effort, prohibition has not achieved the endgame for drugs. It's possible that criminal prohibition has prevented enough addictive drug use to be justified on either public health or autonomy terms. But even if prohibition works, its success depends on incarcerating hundreds of thousands of people. The same point would likely apply were tobacco banned. Given the scale of the demand for tobacco—over 1 billion people smoke worldwide—and the degree of enforcement states would likely see as required to make a ban fully effective, it is plausible that a tobacco ban that aimed to bring about the tobacco endgame would lead to the incarceration of a great many people.

Incarceration is perhaps the most profound restriction on autonomy that the state can impose. Someone convicted of a crime and sentenced to a period of incarceration loses their freedom of movement and of association, the freedom to work, and in many cases the freedom to make even basic decisions about how their lives will be conducted. To be incarcerated is to experience a radical form of unfreedom.

Perhaps the concern about incarceration is overblow, however. After all, many laws—including many criminal laws—are enforced through sanctions that fall short of incarceration. It's possible that enforcement of a tobacco ban would not require imprisoning those involved in the illegal tobacco trade, but would work using alternatives to incarceration. Setting aside the question of whether such noncarceral enforcement would be effective, it's important to note that criminal sanctions short of incarceration can also limit a person's autonomy.³³ For instance, in many jurisdictions, those sentenced to probation can face restrictions on what they can do, where they can go, and with whom they can associate. Moreover, whatever freedoms they do have are conditional; if they violate the conditions of their probation, they can be incarcerated to complete their sentences. A criminal conviction reduces the convict's freedom of choice in Schmidt's sense, and such convictions can continue to constrain someone's autonomy well beyond the prison walls.³⁴

In assessing public policy proposals, it is important to adopt an equity lens, and to consider not only the overall costs and benefits of a policy but also the distribution of those costs and benefits. As Voigt argues, smoking raises equity‐focused questions of social justice.³⁵ Within many countries, smoking is more common among those of lower socio‐economic status.³⁶ Globally, smoking rates are generally lower and have declined more sharply in high‐income countries than in low‐ and middle‐income countries.³⁷

These facts lead some advocates of ban to emphasize its potential egalitarian benefits.³⁸ Because the tobacco endgame would disproportionately benefit less advantaged individuals and countries, the tobacco endgame would advance the cause of equality of both health and freedom.

Such egalitarian concerns should play a role in formulating tobacco policy. But just as the autonomy benefits of a ban need to be balanced against the autonomy costs of enforcement, the egalitarian benefits of the ban also need to be balanced against its corresponding costs.

There is a close connection between using drugs and selling drugs. Those who use drugs are significantly more likely to sell them, and vice versa.³⁹ If the illegal tobacco trade were anything like other illegal drug markets, smokers would be much more likely to sell illegal tobacco, and people involved in the illegal tobacco trade would be much more likely to smoke. These would be the people who would stand to lose the most, in autonomy terms, from the criminal enforcement of a tobacco ban. And since smoking is more common among people of lower socio‐economic status, this means that even if smoking itself was not criminalized and the state was scrupulously just in enforcing tobacco prohibition, the autonomy costs of enforcement would fall disproportionately on the already disadvantaged smokers the ban is meant to protect.

Of course, we know that in many states the criminal law is enforced in a highly discriminatory rather than a scrupulously just way. For instance, White and Black Americans both use and sell drugs at similar rates.⁴⁰ Despite these similarities, Black Americans are more than four times as likely to be incarcerated for drug crimes than White Americans.⁴¹ So while the tobacco endgame would indeed promote equality across a range of values, attempting to bring about the endgame by enforcing a tobacco ban would risk exacerbating many existing inequalities. A concern with equality, then, should make us very cautious about embracing the criminal law to advance even very valuable public health and autonomy goals. Even if the autonomy costs incarceration were outweighted and therefore justified by the significant health and autonomy benefits of an effective ban, the risk that the costs of enforcement would be unjustly and unequally shared is a distinct reason to reject a complete ban on tobacco.

The argument thus far is that a ban would lead to a black market for tobacco, and enforcing a ban would involve the criminal law. This would require criminally punishing people involved in the tobacco trade. This use of the criminal law in pursuit of the tobacco endgame would therefore involve significant restrictions on autonomy and would likely exacerbate existing socioeconomic inequalities.

Despite the significant concerns that I've raised about the potential costs of a tobacco ban, the argument I've made this far is not that a ban could not possibly be justified or that the autonomy restrictions of enforcement are necessarily unjust. I accept for the sake of the argument that even significant restrictions on individual autonomy can be justified on both autonomy and public health grounds.⁴² Criminal laws prohibiting assault, exploitation, and fraud all serve to protect autonomy, and all are enforced through criminal punishment. Vaccine mandates similarly restrict individual autonomy in order to protect public health. Both the criminal law and public health mandates can be justified even if their enforcement requires significant restrictions on autonomy, so long as their overall benefits outweigh the very real autonomy costs of enforcement.

What I have argued, however, is that the normative assessment of a tobacco ban needs to include the autonomy and equality costs of such a ban, and that those costs are not limited to the loss of the freedom to smoke. One important reason to acknowledge these costs is because they might outweigh the autonomy benefits of even a perfectly effective ban. If ending smoking requires incarcerating hundreds and thousands of people every year, this might be worth it, or it might be too high a cost. Deciding whether the benefit would be worth the cost requires, among other things, deciding how to compare the autonomy loss of incarceration with the autonomy loss of addiction and reduced freedom of choice, and that requires properly acknowledging the autonomy costs of enforcing a ban. Moreover, an ineffective ban would generate fewer autonomy benefits but the same autonomy costs. So the autonomy case for a tobacco ban further suffers when we abandon unrealistic best‐case scenarios and consider the much more likely scenarios in which the ban is at best only partially effective.

If a ban were partly effective—meaning that it reduced but did not come close to eliminating global tobacco use and its corresponding health costs—then the question of whether the ban is justified would be complex and would depend in part on how much it reduced tobacco use and how significant the autonomy costs of enforcement would be. Even partially effective policies are often preferable to no policies at all. Speed limits, for instance, can be justified on the grounds that they increase safety even though the limits are not always obeyed, and enforcement is uneven and carries real autonomy costs. Policy makers will rarely know in advance just how effective a given policy will be in achieving its aims, and so will need to make empirically informed predictions about the potential costs and benefits. The crucial point, however, is that such predictions would need to take into account the very real autonomy costs of enforcing the ban. The predictions would also need to consider how likely it is that a ban would succeed in being even partially effective at reducing tobacco use. I've argued above that because a ban would require abandoning several demonstrably effective policies, it's plausible that it would fail to reduce smoking rates at all. It would, of course, be a complete disaster if a tobacco ban failed to reduce—or worse, increased—global smoking rates while simultaneously throwing hundreds and thousands of people into prison. It is hard to know how likely this outcome is, but however likely it is, it is almost certainly more probable than the prospect of a complete ban being completely effective and achieving the true tobacco endgame.

6. PRINCIPLES AND POLICY

We can draw an important general lesson about the role of ethical arguments in the policy process from the very likely failure of a tobacco ban to achieve the tobacco endgame. Neither Schmidt nor Grill & Voigt are naive about the difficulty of enforcing a ban, nor do they claim that a ban would be effective. Both are explicit that their arguments depend on idealizing assumptions. But both claim that the evaluation of an effective ban can provide insight into the normative justification of such a ban. Grill and Voigt argue that “the necessary debate about different policy instruments in various contexts will be greatly facilitated by consideration of the principled argument for a perfectly effective ban,”⁴³ and both suggest that their arguments support ban on tobacco as a policy approach. Grill & Voigt's article is called “the case for banning cigarettes” and Schmidt is clear that his argument supports “radical endgame proposals, like a cigarette ban.”⁴⁴

This use of idealizing assumptions to defend specific policy proposals is unjustified. Because we can be confident that a ban would not successfully bring about the tobacco endgame, pretending it would work is not a helpful way of justifying such a ban. Policy proposals are not philosophical thought experiments, and policies that we know won't work as intended are not justified by arguments that assume that they do.

If nicotine addiction is so powerful that a ban would be autonomy‐enhancing, then it's surely powerful enough to create the kind of urgent and unmet demand for cigarettes that the black market would step in to fill. Idealizing assumptions about the effectiveness of bans are out of place in this context: we know that many smokers will do what they feel they must to satisfy their cravings. The autonomy‐undermining strength of the addiction is what sets up the problem and justifies the proposed solution, and so cannot be assumed away when considering the potential drawbacks of that very solution.

Schmidt acknowledges this point; he says that “concern with freedom of choice does not speak against even the most radical endgame proposals, like a cigarette ban (conditional on such proposals being effective).”⁴⁵ But since this condition is very unlikely to obtain, Schmidt is offering a tacit admission that the argument is not a justification of a tobacco ban at all.

Does this mean that Schmidt's autonomy argument has no value? Not at all. But rather than conclude that an effective ban would be justified—a conclusion that is irrelevant in a world where a ban would not be effective—we should draw a different conclusion. His central insight is that the freedom to smoke should carry no weight in our normative assessment of tobacco control policies. The desire of (some) smokers to have the freedom to smoke is irrelevant to tobacco policy, because such freedom is self‐undermining. So a concern for the autonomy of smokers is not a reason to oppose restrictive tobacco‐control policies and might in fact count among the reasons to endorse such policies.

This is an important insight. But notice: this conclusion is not a policy proposal. It's a claim about the normative significance of autonomy in assessing and setting such policies. Even if we are concerned that tobacco policy pays attention to autonomy, we do not need to give weight to the freedom of smokers to continue to smoke, since such a freedom is not worth protecting for its own sake. But it is a long way from that claim to the conclusion that a tobacco ban should be adopted, because as I've argued above, there are many more ways that a tobacco ban would affect individual autonomy. The freedom to smoke might not be valuable, but the freedom of movement and association certainly are. A recognition of the value of those freedoms is also central to our assessment of tobacco control policies, but that value is obscured by idealizing assumptions about the effectiveness of a potential ban. Such assumptions are out of place when discussing policies, as opposed to the goals those policies should aim to achieve.

My argument is not that contested ethical values have no place in policy‐making. Values are essential in setting out both policy goals—such as equity or public health—and constraints on the pursuit of those goals—such as individual autonomy. Idealizing assumptions that abstract away empirical complications can illuminate important ethical issues, and so help in setting policy goals. But idealizing assumptions that are appropriate at one level can be misplaced at another; they can help identify the relevant values but can interfere with the details of policy design. This is because making good policy involves choosing the best means to the goals set out by the ethical values, and doing so requires an understanding of the empirical details. The full assessment of public policy proposals requires both a value‐informed assessment of the goals it aims to bring about and an empirically informed assessment of the barriers and the prospects of success. Without both, we cannot engage in the evaluation of policy, rather than of thought experiments. Moral philosophers can offer valuable ethical insights that help shape public policy. They can even do so using the kind of idealizing philosophical arguments that abstract away empirical details. But they should not mistake the conclusions of such arguments for policy proposals, even tentative ones.

Tobacco is a public health disaster, and policies that could bring us closer to the tobacco endgame would be an enormous benefit in welfare, autonomy, and equality. But it is a mistake to conclude on this basis that a tobacco ban would yield such benefits. Such a ban would be unlikely to be effective, and its enforcement would impose significant autonomy and equality costs on the vulnerable people it was meant to help. Tobacco control policy should take the autonomy benefits of freedom from addiction seriously, but the best way to do so is to avoid idealizing assumptions. Instead, those concerned to use tobacco policy to advance autonomy should pay close attention to the empirical details of what does—and does not—work to reduce the catastrophic toll of smoking.

Biography

Mathieu Doucet is an associate professor in the Department of Philosophy at the University of Waterloo. His research is in ethics, with a focus on moral psychology. His current research project explores addiction and the ethics of drug policy. His work can be found at mathieudoucet.org

Doucet, M. (2025). Trading one problem for two: The case against tobacco bans. Bioethics, 39, 205–212. 10.1111/bioe.13348