Version classiqueVersion mobile
OpenEdition Books

Arguing about justice

Axel Gosseries
Philippe Vanderborght

Philosophers and taboo trade-offs in health care

Erik Schokkaert


Psychologen leren ons dat mensen het moeilijk hebben om afwegingen te maken tussen enerzijds waarden als leven en gezondheid en anderzijds materiele belangen ze zullen naar manieren zoeken om die vraagstelling niet expliciet onder ogen te moeten zien. Prioriteitenstelling in de gezondheidszorg vereist typisch dit soort van afweging. Ik suggereer dat de procedurale benadering van Norman Daniels en de hypothetische benadering van Ronald Dworkin zeer goed gelijken op de vluchtwegen die mensen voor zichzelf zoeken als ze met die moeilijke keuzen geconfronteerd worden. Dit inzicht is op zichzelf geen argument om die theorieen te verwerpen. Maar het kan ons wel helpen om hun tekortkomingen duidelijker onder ogen te zien.

Texte intégral

  • 1 I borrow from the title of the second chapter ('Reapprendre a mourir') in Van Parijs (1996).

1Some of the hardest choices societies have to make are situated in the domain of health care. If health insurance is financed collectively (through different arrangements ranging from tax financing to social insurance), a collective decision is needed about the size of the health budget and about the priorities within that budget. This raises questions concerning the reimbursement of end-of-life treatment: perhaps we have to learn again "how to die".1 Even more challenging are the problems raised by the treatment of rare diseases and the reimbursement decisions with respect to orphan drugs. What to do if these become extremely expensive? We all die, but by definition we do not all get a rare disease.

2I will briefly introduce two popular and coherent frameworks that have been proposed by philosophers to think about this issue: the procedural approach of Norman Daniels and the hypothetical ex ante-approach of Ronald Dworkin. I will argue that, despite their popularity, both have crucial flaws. Then why did they become so popular? I suggest that this is partly because they respond to deep psychological needs. This raises questions concerning the kind of arguments that can and should be used in discussions on ethical values. More specifically: can one question the validity of an argument by pointing to (some of) the psychological motivations behind it?

3I will for the sake of the argument focus on the exemplary case of a genetic disease X that hits only a small fraction of the population (say, less than 1 person in 2000, which is the ”official” European threshold to speak of a rare disease). Suppose that disease X cannot be cured, but that a newly invented treatment could lead to a prolongation of life by a couple of good quality years. The treatment is expensive, however, and reimbursing it in the collective system, together with similar treatments, would require a sharp increase in the health care budget – or crowd out the reimbursement of more common health care items. Under what conditions should we reimburse the treatment of X in the collective system?

4Easy answers to this question are not convincing. It is sometimes put forward that health has no price – or that everybody has an equal right to treatment, whatever the price to be paid. However, giving absolute priority to the sick is obviously unacceptable as a general principle. Should society pay for all additional health care expenditures as long as they generate marginal health improvements? Even if we restrict ourselves (unrealistically) to health objectives, investments in education, in the environment, in housing may be at least as effective, mainly for the poor. Another escape route is to claim that all needs could be satisfied, if only we were able to remove the glaring inefficiencies in the health care system. That there are inefficiencies is beyond doubt – but many of them are intrinsically linked to the information problems that are typical for the commodity health care (Arrow 1963). While a better organization of the system could indeed remove some waste, it is naive to believe that this would create room for reimbursing all new therapies.

5This brings us back to our case X. It cannot be the optimal policy to simply reimburse all new expensive therapies. Nor can we avoid making difficult choices by improving efficiency. We better admit that there is a problem of priority setting – or ”rationing”. How can we tackle it? Let me describe the positions defended by Daniels and Dworkin.

Daniels’s procedural approach

6Daniels’s starting point is that general philosophical theories of justice in health are not sufficiently detailed to tackle specific questions of priority setting. His own Rawlsian approach (Daniels 2008) requires managing health care and other determinants of health in order to guarantee to all, as much as possible, access to normal functioning as a precondition for an autonomous life. Although this ”opportunity-based” view yields specific policy conclusions (such as the need to fight discrimination and social inequalities), it does not give a definite answer to the question whether to reimburse treatment for X. Reasonable people may disagree about the specific trade-offs to be made in that latter setting (e.g. to what extent modest benefits for larger numbers of people outweigh significant benefits for fewer people), even if they agree that the overall goal of health policy should be to protect opportunity.

7Daniels & Sabin (2008) therefore propose a procedural approach: ”accountability for reasonableness”. Relevant decision makers should set priorities through a process satisfying a set of fairness conditions. Three of these are formal. Decisions and their rationales must be publicly accessible (publicity condition); there must be mechanisms for challenge and dispute and opportunities for revision of policies in the light of new evidence or arguments (revision and appeals condition); there must be public regulation of the process to ensure that the other conditions are met (regulative condition). The fourth is more about content: the rationales for priority-setting decisions should aim to provide a reasonable explanation, i.e. an explanation appealing to evidence, reasons, and principles accepted as relevant by fairminded people (relevance condition).

  • 2 To be fair to Daniels, I should acknowledge that he emphasizes the practical need of introducing fa (...)

8As emphasized by Daniels himself, ”accountability for reasonableness” is a practical approach. Actual decisions are taken in all countries by some sort of committee procedure and, since there is at this stage no consensus about underlying philosophical principles for priority setting, improving these decision-making procedures seems to be a definite step forward. In this respect the fairness requirements described by Daniels are eminently sensible. Yet, Daniels does hardly say anything about the kind of arguments that would be ethically justifiable in the context of priority setting (except that they should not come in conflict with the general equal opportunity-approach). The relevance condition is only vaguely defined and seems to involve some circularities in the sense that ”fair” arguments are those used by ”fair-minded” people. In fact, the openness of the approach is not by accident, but by purpose. As summarized in Daniels and Gruskin (2008: 1577): ”...we are not proposing a formula or algorithm for generating particular priorities. An algorithm would do away with the process, and it is precisely the process that is the point”. All this looks fair and democratic – this is what makes it attractive. Yet, ultimately, it also means that philosophy does not really add to the substantive discussions about the hard choices. In my view, this is deeply disappointing2.

Dworkin’s hypothetical approach

9A second possible approach has been proposed by Ronald Dworkin (2000). According to him, a free market solution applied to health insurance is unacceptable for three reasons. First, because in a society with an unequal distribution of wealth richer people can afford a better insurance. Second, because a free health insurance market would imply undesirable premium differentiation on the basis of health risks. And third, because many people have inadequate information about the value and the personal and social cost of particular medical procedures.

10Dworkin then claims that a just distribution is one that people would create for themselves by individual choices, provided these three problems were somehow corrected. He proposes the following thought experiment: what health care would we have if individuals could take insurance against adverse health events (including an unfavourable genetic endowment) on a perfect insurance market, provided that the distribution of wealth is just, that individuals are perfectly informed about medical technology, and that nobody – including insurance companies – has information on the inter-individual differences in health risks? The answer to this question describes an ideal solution of "prudent insurance", which should be mimicked in the real world, with public transfers and health care subsidies replacing the resource transfers that would be implemented under the ideal (hypothetical) system.

11How would such an insurance system look like in practice? According to Dworkin, returns to scale and the need to avoid adverse selection are strong arguments in favour of a universal compulsory system. This system would typically include regular medical care, hospitalization when necessary, periodic check-ups, etc., since all well-informed and responsible individuals would take insurance for these events. But the ”hypothetical insurance” approach also suggests that some items should not be included in the insurance coverage. Would it for a 25-year-old be rational to insure herself so as to provide for life-sustaining treatment if she falls into a persistent vegetative state? Would prudent young people buy insurance that could keep them alive by expensive medical interventions for some months, even if they had already lived into old age? Dworkin suggests that the answers to these questions would most probably be no. And, returning to our example, it is also likely that a very expensive treatment for a very rare disease would not be covered by the health insurance system that comes out of this thought exercise.

12The Dworkin-approach makes it possible to justify apparently harsh rationing decisions on the basis of a rational consensus argument. The hypothetical insurance device transforms an interpersonal redistribution problem into a kind of (intrapersonal) rational allocation decision. Since in the hypothetical situation everybody would agree on the scope of collective insurance, it seems only a matter of consistency to accept this solution also in the real world.

13Yet, this is too easy. When adverse health events decrease the marginal utility of income, rational individuals would ex ante refuse to take insurance (or would even want to counter-insure themselves and consume more when they are healthy). Of course, this refusal to insure is based on their ex ante ignorance of their final situation. Had they known that they would be hit by the adverse event, they would have insured themselves. It can be argued that social evaluation should be based on these ex post preferences since they reflect better information about the state of the world (Fleurbaey 2010). The argument gets especially strong when applied to transfers across individuals with unequal genetic endowments, as in our case X. It will be difficult to convince individuals suffering from a rare genetic disease that it is fair not to reimburse their treatment – on the ground that in the hypothetical state in which they had had a chance not to suffer from the disease, they would have preferred not to take insurance. Denying them treatment on this basis just seems grossly unfair.

The psychology of taboo trade-offs

14In my view, the limitations of these two approaches are clear. The first does not offer a substantive argument; the second disguises the decision to let a person die as a cool and rational consensus position under a veil of ignorance. How then to explain that I feel emotionally attracted by both theories? Why do they give me a feeling of relief? Psychological research can help me solve this introspective question.

15The decision about whether or not to include the treatment of disease X in the compulsory insurance cover implies a trade-off between better health (for some) on the one hand and less consumption (for the healthy) on the other hand. Psychologists have pointed out that people want to avoid such trade-offs (Fiske and Tetlock 1997; Tetlock 2003). Health and life are socalled "sacred values". Consumption is a "vulgar value". People have no problem with routine trade-offs, i.e. with choosing between different vulgar values. Tragic trade-offs, i.e. trade-offs between sacred values, are more difficult but still ”acceptable” – think about the Greek tragedies in which the hero has to choose between his nation and his life. Yet, we have serious difficulties with taboo trade-offs, with choosing between sacred and vulgar values. This is not only due to the cognitive difficulties related to the absence of a common metric for evaluating such diverse entities. More importantly, these trade-offs are morally disturbing, as our commitments to other people require us to deny that we are willing to attach a finite monetary value to their health or their life. People reject certain comparisons because they feel that seriously considering the relevant trade-offs would undercut their self-images and social identities as moral beings. Having to weigh up one life against another is hard. Weighing up material prosperity against health is scandalous.

16Empirical (including experimental) research has shown that human beings will consider escape routes when these difficult choices cannot be avoided. I mention three of them. First, we are motivated to look away and are easily distracted by rhetorical smokescreens. Second, we try to postpone these difficult decisions or, even better, to pass the buck to others. More or less secret expert commissions are installed, where the ethical choices can be packaged as technical decisions. Third, we try to disguise the taboo tradeoffs as routine trade-offs by reformulating the choice problem so that it does no longer involve our moral commitment to other people.

17Returning to case X, it is fairly obvious that denying the need of rationing fits perfectly into the first strategy. However, I suggest that the attractiveness of Daniels’s and Dworkin’s theories is also partly linked to these psychological mechanisms. ”Accountability for reasonableness” shifts responsibility from individual citizens (including the philosopher and the reader) to committees and procedures – and committee members can drown their own individual responsibility into a kind of collective group responsibility. This is close to the second strategy. Dworkin’s hypothetical approach turns an interpersonal distribution problem into an intrapersonal allocation choice, so that I do no longer have to feel responsible for or to show compassion with other human beings. This is a step in the direction of the third escape route.

18Of course, it would be grossly unfair to reduce the elaborate theories of Daniels and Dworkin to simple psychological tricks. On the contrary, ”accountability for reasonableness” can also be seen as an attempt to go beyond the escape route of secret committees by putting forward formal requirements of openness and transparency. And Dworkin’s view is an ingenious approach to structure a difficult decision problem in a coherent way. Yet, for me at least, some insights into the psychological mechanisms to which I am subjected help me understand why I feel emotionally attracted by these theories, although I am at the same time rationally convinced that they have severe limitations.

19It is obvious that I do not think that the findings about how human beings handle taboo trade-offs offer in any way a conclusive argument against the fine theories of Daniels and Dworkin. Yet, I do believe that a better insight into the psychological mechanisms and cognitive and emotional biases that influence our ways of reasoning may help to get a better perspective on the relative value of different ethical theories. In the case of priority setting in health care, they should warn us for the fact that theories may be attractive for the wrong reasons, i.e. because they make decisions more bearable rather than leading to fairer outcomes.

  • 3 According to Hall and Jones (2007), this is the welfare optimal share of health care expenditures i (...)

20For the problem of priority setting, I think we should have the courage to feel uneasy. Since we cannot escape making taboo trade-offs, we must be open about them. It is crucially important to set up fair decision procedures – but we should also have the courage to think explicitly about what are valid substantive ethical arguments in the context of priority setting. And we should be aware that a ”clean” hypothetical approach can at best be a starting point. Denying treatment to patients with a rare disease is not a hypothetical choice – nor is the decision to spend 40% of our national product to health care3 What matters at the end, is the final ex post-distribution of income and health.



ARROW, K. (1963), 'Uncertainty and the welfare economics of medical care', American Economic Review, 53: 941-973.

DANIELS, N. (2008), Just health. Meeting health needs fairly, Cambridge: Cambridge University Press.

DANIELS, N. & SABIN, J. (2008), Setting limits fairly: learning to share resources for health (2nd ed.), Oxford: Oxford University Press.

DWORKIN, R. (2000), Sovereign virtue, Cambridge: Cambridge University Press.

FISKE, A. & TETLOCK, P. (1997), 'Taboo trade-offs: reactions to transactions that transgress the spheres of justice', Political Psychology, 18: 255-297.

FLEURBAEY, M. (2010), 'Assessing risky social decisions', Journal of Political Economy, 118: 649-680.

GRUSKIN, S. & DANIELS, N. (2008), 'Justice and human rights: priority setting and fair deliberative process', American Journal of Public Health, 98: 1573-1577.

HALL, R. & JONES, C. (2007), 'The value of life and the rise in health spending', Quarterly Journal of Economics, 112: 39-72.

TETLOCK, P. (2003), 'Thinking the unthinkable: sacred values and taboo cognitions', Trends in cognitive sciences, 7: 320-324.

VAN PARIJS, P. (1996), Refonder la solidarite, Paris: Editions du Cerf.


1 I borrow from the title of the second chapter ('Reapprendre a mourir') in Van Parijs (1996).

2 To be fair to Daniels, I should acknowledge that he emphasizes the practical need of introducing fair procedures at this stage, and leaves open the possibility of formulating a coherent substantial theory of priority setting in the future.

3 According to Hall and Jones (2007), this is the welfare optimal share of health care expenditures in the US for 2050.


Erik Schokkaert is professor of welfare economics at the Department of Economics of the KULeuven, and currently Research Director of CORE (UCLouvain). He chairs the interdisciplinary think tank ”Metaforum” at the K.U. Leuven. His research focuses on (a) the modelling of different concepts of distributive justice; (b) the empirical analysis of opinions about justice in the population; (c) the application of these theories to policy problems in the fields of health, social security and taxation.

© Presses universitaires de Louvain, 2011

Conditions d’utilisation :


Volume papier