There are two main types of argument supporting an obligation for the state to compensate for losses created by public health interventions. The first type are pragmatic arguments and the second ethical arguments, but as we shall see this is one of the cases where ‘all roads lead to Rome’ in the sense that we get strong support for an obligation to compensate from both pragmatic and ethical arguments, and from ethical arguments of several different types [a similar portmanteau approach to justification has been taken by Mello in her analysis of vaccine injury compensation (Mello 2008)].
If these arguments go through they also provide strong support for a societal obligation to put into place legislation and social policies aimed at preventing some of the costs and burdens from occurring, e.g. by protecting the employment rights of people in isolation, quarantine or complying with recommend social distancing as well as the rights of their relatives, and by working to limit the stigmatising effects of being an asymptomatic carrier.
The pragmatic arguments only need to be stated briefly here, since they have already been analysed in detail by Rothstein and Talbott (2007b) and Ly et al. (2009). They are based on the assumption that if people are threatened with big losses such as loss of livelihood or significant risks to health if they comply with public health interventions, then compliance will be undermined and it will become more likely that they will try to evade the public health intervention. This is in essence an argument based on the incentive structure created by, for instance home quarantine or self-isolation without protection of employment status and compensation for lost wages. If compliance is a desirable societal goal it can then be achieved by changing the incentive structure in either of two ways. We can either increase the penalties for non-compliance or decrease the costs of compliance, or some combination of the two. Unless non-compliance is easily detectable and a penalty approach therefore likely to work, decreasing the costs of compliance, for instance through compensation schemes becomes the pragmatically preferable policy option. Elements of the Chinese response to the Covid-19 outbreak illustrates that effective surveillance and punishment regimes may be possible (Culbertson 2020), but probably only in societies that already have a high level of state surveillance.
There are also several ethical lines of argument supporting an obligation to compensate. Apart from consequentialist arguments which will broadly track the pragmatic considerations outlined above and support compensation because it leads to more effective public health interventions and better aggregate outcomes, ethical arguments can be based onFootnote 6:
A perfect duty of non-maleficence
An imperfect duty of beneficence
Reciprocity (in itself or as part of a solidarity based consideration)
Let us analyse these in turn. It is plausible that individuals have a perfectprima facie duty not to harm others and that the state has a similar perfect prima facie duty not to cause harm. The assumption that the public health interventions we discuss here are legitimate entail that this perfect prima facie duty has been considered as part of the decisions to implement the interventions and that it has been found that it is outweighed by one or more other prima facie duties that are more important in the context of the specific decision, e.g. the state’s duty to protect the life and welfare of its citizens. But, the fact that the harms and losses caused by the intervention are justified in the sense that they are not sufficiently severe to show that we should not implement the intervention does not in itself show that it would be justified to ignore these harms and losses and the effect they have on those people to whom they fall. The duty not to harm come with a correlative duty to make good any harm that we do cause, for instance by compensation.Footnote 7 That duty is also a prima facie duty and can be outweighed by other duties. However, when analysing whether the duty to compensate for harms done is outweighed the question is no longer whether the public health intervention, with the known imposition of harm is justified, but which of two possible interventions is most justified, i.e. the intervention with or the intervention without a compensation scheme. In that analysis and weighing the state’s duty to protect the life and welfare of its citizens will in many cases have much less weight, simply because a compensation scheme does not impede the state’s ability to discharge that duty.
The general imperfect duty of beneficence also supports providing compensation in this context, at least in cases where the harms and costs lead to a significant reduction in welfare. Here the consideration would not be the causal connection between the public health intervention and the harms and cost, but simply the claim to help generated by a loss of that magnitude.
What about considerations of justice? Decisions about when and how to implement public health interventions raise basic issues of justice, including fundamental issues about distributive social justice. The issue of whether justice arguments support compensation is, however a more restricted question about whether those whose positions are changed in a negative way by the intervention should be compensated as a matter of justice. Or to put it differently, do the harms, costs and burdens caused by public health interventions count as unjust changes that have to be rectified? This focus on negative change significantly limits the range of approaches to justice we need to consider. One way of approaching this question is through the lens of luck egalitarianism (Arneson 2004; Lippert-Rasmussen 2015). Are these harms, costs and burdens the result of brute or option luck? Becoming the target of a public health intervention is in most cases the result of a complex causal network where the individual in question has little or no control over many of the causal nodes. For instance, most MRSA is either picked up in hospitals or other health care settings, or from animals in farming; and it is eminently possible to pick up MRSA and become a carrier even when taking all the recommended precautions. Similarly complex causal networks lies behind individual cases of flu during a flu epidemic etc. In all of these causal networks there will be some nodes that reflect personal choices, some that reflect the choices of others, some that reflect structural background conditions, and some that simply reflect bad luck, and many nodes will even in a single node reflect a complex combination of all four. E.g. Let us imagine that ending up sitting next to an infectious flu sufferer on the bus going home from work at Manchester University is the causal node which constitutes the Mackiean INUSFootnote 8 factor that determines that a specific person becomes infected. This may be the result of a choice to use public transport, the choice of someone else to come to work even when ill, the bus always being very full due to public transport funding issues, and making an unlucky choice between the last two empty seats. If this analysis of the causal situation is correct in most cases it means that acquiring the infection, or acquiring the possibility of having an infection which makes you a target for public health intervention is a question of brute luck and not a question of option luck, even though the causal network contains some nodes of personal choice. And, as such the costs that follow from this bad brute luck ought to be compensated.
Another justice based argument which should appeal to consequentialists and welfarists in general considers the public health intervention through the lens of Pareto optimality. Pareto optimality is the standard approach in economics to evaluating the acceptability of a social change. I have made this argument in an earlier paper, and cannot state it more clearly than I did then:
If detention in the public health context succeeds in limiting the spread of the disease it constitutes a potential Pareto optimal social change compared to the situation where the disease outbreak is allowed to run its course unchecked. Analysis of the Toronto SARS outbreak does for instance seem to indicate that despite the very large costs of quarantine the economic benefits outweigh the costs (Gupta et al. 2005). A potential Pareto optimal social change is by definition a change where some lose and some win but where there is a net social welfare gain such that the winners could compensate the losers for their loss. A potential Pareto optimal social change can therefore, again by definition, be converted into an actual Pareto optimal change, i.e. a change where some gain welfare but no one loses welfare, if the winners compensate the losers for their loss.
Let us assume (1) that the number of detained persons is small compared to the general population and (2) that the principle of diminishing marginal utility of resources holds (at least approximately). Then it follows straightforwardly that the social state after compensation has been provided will have higher aggregate welfare than the state before compensation. The former losers will gain more welfare from the redistribution inherent in compensation than the former winners will lose, and the net loss to each former winner will be small.
There is thus a strong welfarist argument for providing compensation. This argument is even stronger if we accept a form of prioritarianism because this will direct us to give special weight to welfare improvements for those who are worst off in welfare terms (Rabinowicz 2001). (Holm 2009, p. 201).
Reciprocity is, finally the basic idea that there is an obligation to respond in a proportionate and fitting manner for the good or ill we receive from others (Becker 1986, 2005). We can think of reciprocity as a free standing obligation or as a part of distributive justice or a larger framework of social solidarity. Within a social solidarity framework enacting reciprocity will be one of the solidaristic practices that sustain the overall solidarity within the relevant group or society.Footnote 9 In the present context not much hinges on precisely where we locate reciprocity. The good involved in the present context is willingly complying with public health interventions that are instituted to achieve a wider public good. The proportionate and fitting response is gratitude towards those fellow citizens who comply (Silva and Viens 2015), and if compliance generates significant personal costs for them a proportionate and fitting response would be to help to defray those costs. While we may not call our response to the costs of our fellow citizens ‘compensation’ when contemplating it from the perspective of reciprocity, the end result is essentially the same. They have freely provided an important good for society and letting them carry the full costs themselves would not be a fitting and proportionate response.
Slightly more formally we can combine the pragmatic and the ethical lines of argument in the following way (where PHI = Public Health Intervention, PEmp is an empirical premise, PEth an ethical premise, PPrag a pragmatic premise and C a conclusion).
PHI carried out in order to ensure public benefit
The positive effects of the PHI are widely shared (and may in the technical sense constitute a public good)
The harms/burdens fall to a small number of people
There is no particular justification for harming/burdening these specific people (e.g. they have not explicitly or implicitly waived their rights not to be harmed, they are not personally responsible for being in the state which necessitates the PHI etc.)
The people who are harmed/burdened do not have any specific duty to accept this harm
Harms and burdens caused by actions such as PHI should normally be compensated (follows from a number of independent ethical considerations enumerated above)
There is an ethical case for compensation.
The existence of a compensation system is likely to increase adherence to the PHI
C. The people who are harmed and/or burdened by the PHI should be compensated for the harm and/or burden.
This argument is valid, but it is obviously only sound in those cases where all the premises are true.