Should Vaccination Be Mandatory in a Public Health Emergency?
A pathogen spreads. People die. The state compels vaccination. This sequence conceals the question that actually determines whether mandates succeed: not whether governments may compel vaccination, but whether compulsion produces sufficient genuine compliance to reach the herd immunity threshold the pathogen requires. A mandate that generates administrative records while leaving the epidemiological gap intact is not policy. It is performance. This essay argues that mandatory vaccination in a genuine public health emergency is both permissible and sometimes effective, but that effectiveness depends on conditions governments must demonstrate rather than assume. When those conditions are absent, the obligation to act does not disappear. It shifts.
I. Permissibility: The Harm Principle and Its Limits
John Stuart Mill's harm principle holds that the state may restrict individual liberty only to prevent harm to others.[1] Non-vaccination in a genuine public health emergency satisfies this criterion. The individual who declines vaccination reduces population immunity below the threshold required to protect those who cannot be vaccinated regardless of preference: the immunocompromised, neonates, and the medically contraindicated. The harm is real, attributable, and falls on third parties who bear no responsibility for their vulnerability.
The strongest objection comes from within liberal political philosophy itself. Robert Nozick treats bodily integrity as a side-constraint the state may not override even for aggregate welfare gains, and insists that the force of side-constraints is not diminished by the scale of the imposition: a small boundary violation is still a violation.[2] The objection is serious. But it does not survive scrutiny in the emergency context, for one structural reason.
Nozick's paradigm cases — forced organ donation, coerced military service — involve extracting a resource from the individual's body for others' benefit. A vaccination mandate does not extract a resource. It interrupts the individual's ongoing causal role in transmitting harm to others. Judith Jarvis Thomson's analysis of the limits of bodily autonomy is instructive here, though it must be applied carefully.[3] Thomson's violinist argument is most often read as establishing that bodily autonomy rights can override obligations to sustain another's life — a reading that might seem to cut against mandates. But the underlying principle Thomson isolates is more precise: strong rights over one's body do not extend to using that body as an active mechanism of harm to third parties who have not consented to bear that risk. The asymmetry matters. Thomson's argument concerns what you are not obligated to do for others; the vaccination case concerns what you may not do to others. Nozick's own framework confirms this distinction: side-constraints prohibit the use of persons as mere means, but they do not license the ongoing imposition of serious probabilistic harm on non-consenting third parties. The mandate addresses the latter. It falls outside the class of acts the side-constraint was designed to prohibit.
A second objection holds that statistical harms — where no individual non-vaccinator can be identified as the proximate cause of any specific infection — fall outside the harm principle's scope. This objection has force in ordinary circumstances. In a declared emergency, it fails. Epidemiological modelling establishes, with actuarial precision, the mortality consequence of coverage falling below threshold. This is the same structure of probabilistic harm that underlies criminal liability for drunk driving before any accident occurs. The harm is not speculative; its probability is quantified, and each non-vaccinator's causal contribution is statistically real and morally attributable.[4]
II. Effectiveness: Three Conditions, Rarely Met Together
Permissibility does not entail effectiveness. A mandate generates genuine compliance — compliance sufficient to achieve the pathogen-specific herd immunity threshold — only when three conditions obtain simultaneously. Each is mechanistically derived from how enforcement converts legal obligation into actual coverage.
First, the herd immunity threshold must be reachable under realistic resistance rates. Enforcement operates at a speed determined by state administrative capacity, and where resistance is high, each coerced vaccination carries greater friction cost — legal challenge, community conflict, personnel deployment — which reduces coverage velocity. If the threshold is high and resistance-adjusted delivery is slow, mandates do not outperform voluntary campaigns; they substitute friction for trust at equivalent or lower coverage. Second, resistance must be insufficient to collapse enforcement coverage below what well-resourced voluntary campaigns achieve in comparable populations. This is empirically testable. Third, functioning delivery infrastructure must already exist: supply chains, accessible administration sites, trained personnel. Coercion cannot substitute for absent logistics. It presupposes them.
The smallpox eradication campaign is the clearest historical confirmation that all three conditions can obtain simultaneously. Fenner et al.'s authoritative WHO account documents their joint presence in the final elimination phases across South Asia and the Horn of Africa.[5] Smallpox's basic reproduction number of five to seven meant the herd immunity threshold was achievable at 80 to 85% coverage — well below the 95%+ required for measles. Ring vaccination made contacts, not universal coverage, the operative target. State capacity was sufficient, and the campaign window was long enough for compulsion to compound. Crucially, smallpox remains the only case in modern public health history where all three conditions were jointly satisfied under a mandate. It is a demonstration of what is possible, not a template that transfers.
The 2018 Ebola outbreak in North Kivu and Ituri illustrates what happens when the second and third conditions fail. Armed resistance and access restrictions severely degraded enforcement coverage in mandate zones, while voluntary ring vaccination with community engagement, deployed in parallel zones with higher trust baselines, achieved substantially higher coverage. WHO field assessments and the outbreak correspondence record document this divergence consistently; Mbala-Kingebeni et al. confirm the pattern.[6] Coercion converted a public health intervention into a confrontation that communities had concrete historical reasons to resist. The mandate produced worse outcomes than the available alternative.
The France-Guadeloupe comparison during COVID-19 isolates the third condition precisely. France's passe sanitaire imposed identical legal obligations across metropolitan and overseas territories. In Guadeloupe, where coverage stood at roughly 13% in July 2021 — constrained by documented supply failures — the mandate produced coverage below 30% by October. In metropolitan France, coverage exceeded 75%.[7] The legal instrument was identical; the outcome diverged because the infrastructure that makes compliance possible was absent. Portugal, operating without a formal mandate but with structured voluntary engagement and nationally consistent supply chains, reached 86% by September 2021. The lesson is not that mandates are worse than voluntary campaigns in principle. It is that mandates without the third condition produce worse outcomes than voluntary campaigns with investment — and that investment, not compulsion, is the operative variable.
III. When Conditions Fail: Structural Responsibility
When the three conditions are absent, mandatory vaccination produces the appearance of policy action while leaving the structural gap intact. But the unavailability of a mandate does not end state obligation. It redirects it — toward the delivery infrastructure, community health networks, and equitable supply chains that genuine compliance actually requires.
More significantly, when genuine compliance is impossible — because supply chains are broken, centres are inaccessible, or populations carry rational grounds for institutional distrust — the question of moral responsibility for preventable deaths becomes urgent and must be answered precisely. Philip Pettit's republican framework identifies domination as the exercise of arbitrary power over others, including the power to expose them to serious harm through structural neglect.[8] On Pettit's account, actors who possess the capacity to build health infrastructure, choose not to, and thereby expose populations to preventable mortality are exercising domination by omission. Pettit develops this analysis primarily in the context of domestic democratic institutions, but the structural logic extends to the international setting through the same mechanism: domination requires only that an agent hold power over outcomes affecting others, and that it exercise that power arbitrarily — without accountability to those affected. Donor states and multilateral institutions that determine whether health infrastructure exists in low-capacity settings meet this description. Responsibility attaches not to the individuals who decline vaccination under genuinely constrained conditions — they are not the authors of those constraints — but to the governments and international actors whose choices determined whether capacity would exist at the moment it was needed.
An objection presses here: is liability without enforcement mechanism meaningfully different from no liability at all? The objection conflates two distinct functions. Enforcement compels action. Visibility makes consequences legible that policy has kept invisible. The history of international human rights law demonstrates that the articulation of state responsibility for torture and enforced disappearance preceded enforceable mechanisms by decades — and was not therefore meaningless. It created the normative foundation on which enforcement was subsequently built and raised the political cost of continued inaction. Naming structural responsibility for absent health infrastructure performs the same function: it converts the political invisibility of neglect into a documented, attributable, and contestable claim.
IV. Conclusion: The Harder Question
Mandatory vaccination in a genuine public health emergency is permissible: the harm principle establishes this, the Nozickian side-constraint objection fails on structural grounds, and the statistical harm objection dissolves when probabilistic mortality is actuarially quantified. It is also sometimes effective — when the herd immunity threshold is achievable under realistic resistance rates, when resistance is insufficient to collapse enforcement below voluntary campaign benchmarks, and when delivery infrastructure already exists. Governments bear the burden of demonstrating these conditions before mandating. The cases of the DRC and Portugal demonstrate, at the level of outcomes rather than theory, that when those conditions are absent, voluntary engagement with genuine investment outperforms coercion in coverage achieved and lives saved.
The strongest objection to this conclusion is that structural conditions change slowly and emergencies move fast. If the conditions for mandate effectiveness are absent at the moment of crisis, it may be too late to build them. This is correct — and it is the essay's deepest charge against the existing framework. The conditions that make mandates either unnecessary or effective are precisely those that governments in low-capacity contexts have systematically failed to build. Their absence during an emergency is not a contingent misfortune. It is the accumulated consequence of choices made long before the pathogen arrived, by actors with the capacity to have chosen otherwise.
That consequence has a political address. The mandate is not the hardest question in a public health emergency. The hardest question is why the conditions that would make it either unnecessary or effective were never built — and who is accountable for that absence. Current global health governance is not designed to enforce that accountability. The vaccination debate, at its most honest, is not a debate about compulsion. It is, at bottom, a question about institutional will — whether the actors with the capacity to make compulsion unnecessary will choose to exercise it.
Endnotes
[1] John Stuart Mill, On Liberty (London: Parker and Son, 1859), p. 22.
[2] Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), pp. 28–33, 48–51.
[3] Judith Jarvis Thomson, 'A Defense of Abortion,' Philosophy and Public Affairs 1, no. 1 (1971), pp. 47–66. The application here is deliberately asymmetric: Thomson's argument concerns obligations to sustain life, whereas the vaccination case concerns the imposition of probabilistic harm on non-consenting third parties. The underlying principle — that bodily autonomy does not license active harm to others — is Thomson's; the direction of application is the essay's own.
[4] Mark Navin, Values and Vaccine Refusal (New York: Routledge, 2016), Chapter 3, pp. 44–67, addresses the probabilistic harm objection directly and concludes that statistical causal contribution satisfies the harm principle's threshold.
[5] F. Fenner et al., Smallpox and Its Eradication (Geneva: World Health Organization, 1988), pp. 1023–1068.
[6] P. Mbala-Kingebeni et al., 'Ebola virus disease in the Democratic Republic of the Congo,' The Lancet Infectious Diseases 19, no. 11 (2019), pp. 1170–1171. Coverage divergence between mandate and voluntary zones is drawn from WHO Health Emergency situation reports for North Kivu and Ituri, August–November 2018.
[7] Santé Publique France, Bulletin de Santé Publique: Guadeloupe, August 2021; European Centre for Disease Prevention and Control, COVID-19 Vaccine Tracker, September 2021.
[8] Philip Pettit, Republicanism: A Theory of Freedom and Government (Oxford: Oxford University Press, 1997), pp. 52–79; On the People's Terms (Cambridge: Cambridge University Press, 2012), pp. 83–91. The extension of Pettit's domestic domination analysis to international actors is the essay's own application; the structural logic — that domination requires power over outcomes exercised without accountability to those affected — applies wherever that power-without-accountability relation obtains, regardless of whether the agent is a state or a transnational institution.