The Ethics of Rationing

The initial Romanell Fellows blog, The Ethics of Rationing, was open to discussion in a one-week forum, at the end of April, 2020. Although this paper is prior to Covid-19, it is an article that will give the “lay of the land”  on the ethics of rationing. The paper touches on plenty of topics, albeit briefly, it will provide Romanell Fellows with their varied interests and specializations with many issues to pursue in their comments.

Paper 1. Principles for allocation of scarce medical interventions, by Govind Persad, Alan Wertheimer, Ezekiel J Emanuel


Jack Freer April 23, 2020 at 8:40 AM

Persad, Wertheimer, & Emanuel (PWE) have described a resource allocation system they call the "complete lives system." They compare it to 3 better known system (UNOS, QALY, DALY).

Even accounting for the difference in familiarity in comparing a new concept to older ones, there still seems to be a qualitative difference between complete lives and the three older systems. As I thought about it, I realized that UNOS, QALY & DALY are real systems. Complete lives is something else. UNOS is the most detailed system (in large part because it has been the benchmark for real life allocation decisions for decades). As such, it has the authenticity that comes with real life and death decisions. QALY and DALY have a fundamental connection with a time scale--a relatively precise, if crude tool.

PWE say, "it is not an algorithm" and indeed, it can not be used in a systematic fashion. Although they say the principles should be ordered lexically, there is no example of this and no obvious way it can be applied. "Less important principles should come into play only when more important ones are fullfilled" sounds prescriptive but the guidance is limited to modifications in the "youngest-first" principle (resulting in the age-based priority graph shown in the article).

For example, let's compare a couple of hypothetical patients. Both are hospitalized with COVID-19 pneumonia. One is a 40 year old manager of a fast food restaurant, applying for unemployment benefits. The other is a 60 year old hospital billing clerk with hypertension and mild emphysema. Both patients are deteriorating and need intubation/ventilation if they are to survive. There is only one available ventilator. I don't see a path to applying complete lives to the dilemma and coming up with anything useful and practical (much less just).

David H April 23, 2020 at 10:25 AM


Good point. The Complete lives System (CLS) is rather unsystematic. They say the complete life is the focus of distributive justice but don't really say much directly about a guiding principle or directive but we must patch it together from the constituent considerations lottery, prioritization of the young, saving the most lives and prognosis, saving the most lives,and instrumental value, and then balance them with an eye towards providing complete lives.

By the way, I think they are unclear in the section on objections. The first time I read it, I thought they were asserting claims about lexical order but on the second read I think that each paragraph of that section begins with an objection that they are then rebutting or mitigating. So the first paragraph is the objection that the Complete Lives System discriminates again older people; the second paragraph starts with the objection that the CLS system is insensitive to international differences; the third paragraph begins with the objection that principles must be ordered lexically and and less important principles come into play only when more important ones are fulfilled. I don't think they are endorsing this claim of lexical ordering and fulfilling more important principles first as they then propose, "as an alternative" balancing

David H April 23, 2020 at 10:31 AM

I want to focus this post on the second paragraph of the section on “Complete Lives System” on the right side of page 428 where Emanuel et al. are defending prioritizing young adults over infants. One claim is that parents have invested greater efforts in adolescents than infants and these will be wasted without a complete life. “Adolescents have received substantial education and parental care, investment that will be wasted without a complete life. Infants, by contrast, have not received these investments.” This is then qualified as being based upon the investments that “adolescents and young adults are morally entitled to receive at a particular age” rather than those they did receive, “consequently, poor adolescents should be treated the same as wealthy ones even though they may have received less investments…” The first claim is dubious, the qualification seems wholly misguided.

Prioritization should depend first and foremost on the harm that death brings the deceased, not the survivors. On the one hand we can imagine the parents of the adolescent having outsourced the childrearing, perhaps even starting with a surrogate and then nurses, nannies and boarding schools. Or we can imagine the infant being the end result of years of expensive infertility treatments then suffers all sorts of illnesses requiring round the clock care that emotionally, physically, and financially taxes the parents. So, in terms of efforts being in vain, more were exerted without return by the infant’s parents than outsourcing parents of the adolescent, and the former may even add up to more than the division of labor in the rearing of the adolescent. My point isn’t to argue that for those reasons we should prioritize the infant, but we should not prioritize on the basis of investment

Furthermore, the qualification of the author to handle unjust underinvestment seems to completely defang the idea of parental efforts being in vain. What sense can we make of wasted investment when the investments were not made but should have been? The authors write “…the prioritization of adolescents and young adults considers the social and personal investment that people are morally entitled to have received at a particular age…Consequently poor adolescents should be treated the same as wealthy ones, even though they may have received less investment owing to social injustice.”


Bob Kelly April 24, 2020 at 6:51 PM

David, I think you raise good objections here. I think part of the problem you identify has to do with a point you made in your previous comment responding to Jack, namely, that the authors don't define 'complete life', or even gesture really at what that means. Hence, we can't really understand what it means for an investment to be "wasted without a complete life." They can't just mean life-years saved since, all else equal, this would favor the infants.

It seems like they might have been vague intentionally, wanting only to focus on shifting attention from "events and episodes" to "entire lives," whatever the latter means (except counting years). Indeed, in response to an objection that they are not sensitive to cultural differences in valuing lives, they say on p. 429 that, "By contrast, the complete lives system requires only that citizens see a complete life, however defined, as an important good, and accept that fairness gives those short of a complete life stronger claims to scarce life-saving resources."

I suspect they were attempting to do something McMahon-esque here by adding a feature (like McMahon's psychological connectedness) that is supposed to capture the intuition that it's worse for an adolescent to die because they've got more to lose (e.g. the foundation for building a complete life, whether from actual investment or the capacity to take advantage of investment). If something like this is right, then they might have two ways to respond to your first point. First, they could agree with you that harm is the metric for prioritization. It's just that, due to (i) a complete life being a (the?) high(est) good, and (ii) adolescents having the most to lose regarding (i), an adolescent is typically going to be harmed more than an infant. Second, perhaps they can even agree with this principle AND agree with your cases, since you may have found cases where the infant is the one who has more to lose in terms of a complete life. I'm not exactly sure how they want investment to work. This comes to your second objection, which I also think is good.

To be clear, I am not ultimately trying to say they have an escape from your objections. I am agreeing that there is a problem here, grounded in their lack of definition for what a complete life is supposed to be. But I wonder if trying to see them as actually agreeing with your idea that harm is the right metric will help us (i) begin to get at SOME understanding of what they want 'complete life' to mean, and (ii) see why they think, in general, adolescents ought to be prioritized over infants.

David H April 27, 2020 at 1:44 PM

My first reading of the Emanuel et al, perhaps primed by my familiarity with McMahan’s Parfit-inspired Time-Relative- Interests-Account of Harm, made me think they were relying on harm being a function of not just what valuable future is lost out on but the degree of psychological ties to the future. But a second reading suggests that is not the case. The authors offer a claim that they consider somewhat related to the argument they just made against prioritizing infants in my first post about the adolescent having capabilities and interests and plans that an infant doesn’t, which I think is meant to be treated as quite different from McMahan’s approach. They write “Similarly, adolescence brings with it a developed personality capable of forming and valuing long-term plans whose fulfillment requires a complete life.” They then appeal to the authority of Ronal Dworkin quoting him from his book defending abortion “It is terrible when an infant dies, but worse, most people think, when a three-year dies, and worse still when an adolescent does.” Dworkin does not account for this tragedy in terms of less psychological ties to the future meaning death is less of a deprivation.

In the book, Life Dominion, from which the Dworkin quote comes, Dworkin defends a very odd labor theory of value that the more efforts made, either social or natural, the greater the value of an entity and the greater the tragedy if the destruction of that entity renders those efforts made in vain. Pro-lifers, Dworkin believes, are moved by the “efforts” that nature has put into the fetus, while pro-choicers are more concerned with the efforts of the unwilling mother that will be frustrated if the law requires that she bear the child, and then perhaps succumbs to non-legal pressures to raise the child, which puts her earlier (heavily invested-in) life plans on hold or wasted. Dworkin appeals to an impersonal value of the fetus do to “nature’s investment”, and not the interests of the fetus who, Dworkin insists, lacks the conceptual apparatus to have interests. Leaving aside that this doesn’t capture pro-life attitudes whose focus in on the harm to the fetus, not some impersonal value lost, it is also an unpersuasive theory of value. The latter point can be seen by comparing the destruction of a lousy work of art that involved great efforts of an untalented artist and the destruction of a great work of art that was very quickly and easily made by an incredibly talented artist. There might be a sense in which it is more tragic when great effort is in vain, but that makes the efforts of the untalented tragic even if the lousy art work is not destroyed. The artist just wasted her time. But there is no increase of impersonal value because of effort. So, investment in a fetus or infant or adolescent can’t change their value. Imagine a species that developed minds like ours automatically without efforts of others. Their value would be the same as ours. So insomuch as investments wasted depend upon Dworkin’s (insufficient, at best) account of value, Emanuel et al have not provided a good reason to favor adolescents over infants. Nor, I suspect, would there be much public support for allowing infants to die to save adolescents.

Now, I would agree with you that it is philosophically better if we don’t treat the endorsement of Dworkin idea as “similarly” to the wasted investment claim, but read it more in the McMahan/Parfit manner that the mentally more advanced adolescent is more closely psychologically tied to the future than the infant and so deprived of more by death. The adolescent has interests that can be frustrated that the newborn lacks; moreover, the adolescent is more intimately tied to the future lost than the infant. The infant, McMahan says, loses out on a future life that is almost like the life of someone else.

David H April 27, 2020 at 1:45 PM

If my last reply is the best interpretation of why the earliest deaths are not worse despite losing out on more of a complete life, it is worth drawing out some consequences, some might be disturbingly inegalitarian. One is that is if the if a child, adolescent, or adult is going to suffer considerable psychological defects from perhaps prolonged time on ventilator, that makes them infant-like in mind, even if they can develop like a normal infant, their prioritization for medical care would be dropped to that of a newborn. That puts pressure against to maintaining the second class treatment of infants as I don’t think the past can justify why adolescents and adults with infant like minds warrant better treatment than infants, though I can appreciate the psychological factors that mind lead us to treat them differently.

David H April 23, 2020 at 3:51 PM

The authors claim they are not engaged in invidious discrimination against the young or old because we all age, while we are not all sexual or racial minorities (p. 429). I think there is more to be said for the claim that we are not discriminating against the elderly than the young. We in the rationing debate are all too old to be the sacrificed young. It doesn’t matter that we were once young. To make the point that we can mistreat the young despite having been young, just imagine someone who successfully changes sex (just grant the possibility for the sake of argument) and discriminates against members of their past sex.

David H April 23, 2020 at 4:08 PM

If a reason the authors deny the elderly a scarce treatment is that they don’t think it is as bad for the elderly to die than the adolescent because the former have enjoyed a complete life, what if the elderly can recall little of that earlier life? Is it just objectively good for them that they had such a long life of good experiences, achievements, and meaningful attachments? If there is supposed to be a subjective justification in that they have a sense of completion or satisfaction, memories of a good life with loved ones and friends, rewarding experiences, desires satisfied, and projects fulfilled, this will be absent in the impaired geriatric suffering memory loss. The authors may have some sympathy for the death of infants not being as bad as the deaths of adolescents who have a developed psychology with projects and plans that psychologically tie them to the future. But the demented geriatric lacks the psychological ties to the past. So symmetry considerations perhaps suggest that the objectively good past doesn’t do the demented much good without psychological connections to the past just as the loss of an objectively good future doesn’t harm the infant who is psychologically unconnected to the future. Perhaps the bad prognosis is sufficient to deny them prioritization, but that has little to do with their having lived a complete life.


Bob Kelly April 24, 2020 at 7:19 PM

Again, good objection. I think your points here again belie the fact that they don't say enough about what a complete life is. Your objection rightly assumes it cannot be mere number of lives, and then seems to assume that it just means 'objectively good life'. This, then, brings up your symmetry worry: they capture the dementia case, but get the infant cases wrong (when infants have good futures).

What do you think about the suggestion that the directions are asymmetrical because what matters is knowledge of both (i) amount of complete life already lived, and (ii) amount of complete life yet to be lived. In the geriatric case, we already know they lived basically a complete life with not much left to make a difference. The 40 yo has lived a decent amount of a complete life, but also has a decent amount left. This might also explain the adolescent vs infant case. While both have a lot of their would-be complete lives to live, it is the adolescent that has already secured a portion of their complete life, which itself adds to the prospect of completing the rest of it (i.e. connectedness, investment, etc.). Since what we know matters, it counts that we're certain about the adolescent's past and only predictive about the infant's future. This would also explain why the 3 yo is favored over the infant as well, but not as much as the adolescent.

Weighing what we're certain of (the complete life portion already lived in the past) and what we predict (the potential for completing the complete life in the future) might have us putting more emphasis on the years lived when evaluating complete lives. In this way, ticking years off as you age starts to add weight to the future years. Having already lived some of a complete life is like insurance for continuing to do so into the future. The more insurance you have, the greater we weight the rest of your potential future life. But only so far. Once the scales start to tip so you have less potential and more lived, the former incur less weight from the latter. This is why the graph flattens in the middle and goes back down towards the end.

I don't know if that made sense. I may have blacked out in the middle there. Thoughts?

Jack Freer April 25, 2020 at 7:33 AM

Bob and David,

The more we try to pin down and characterize a "complete life," the more slippery it becomes. It seems to be more of a poetic device than an analytic tool. It may resonate with me (as a 70 year old recently retired physician, who is now dabbling in art, foreign language and philosophy). It would be hard to argue that I've not lived a complete life--but how does one distill that narrative into something that could be plugged into a resource allocation scheme?

David H April 27, 2020 at 1:47 PM


I may not be following your point. Is it just that if a youngster has lived X+ N amount of time we have more reason to believe he will continue to survive than someone even younger who has lived only for X time? Is it analogous to our knowledge that early embryos are far more likely to miscarry than later embryos, so triage considerations favor saving the latter? I would of course accept that but I think you are saying more. Is it that greater investment makes it more likely that they will survive? If that is, so then prognosis is the guiding concern and I would agree with that being a consideration in a triage situation. But you also mention psychological connections. If these don’t have to do with increasing susceptibility to harm, but just the psychologically connected are better able to take care of themselves, follow directions, need less resources and assistance to survive, then again, I could perhaps agree with you that is a factor in a triage scenario. So, there is a knowledge-based asymmetry. But if it because the actual harm rather than the expected disutility is less, then things get tricky; though I am sympathetic to a modified time-relative-interest-account that expands the sense of interests in a way that includes the interests of the mindless and minimally minded. But this isn’t the place for my trying to put that idea forth. I also have egalitarian worries that moral status is somewhat independent of degree of harm that one is susceptible. Older people have just as much of a right to life as younger, despite death being a greater harm for the former. Now an equal right not to be killed may not mean an equal right to life-sustaining aid, but I am wary of strength of rights varying with degree of harms, despite my intuitions in the triage cases. I am also wary of our ability to make fine grained evaluations of well-being and harm between lives. Best to stop this my tangent now

Phil Reed April 23, 2020 at 4:22 PM

Regarding Jack's point, my initial thought was that their system would obviously give the ventilator to the 40 year old, since, from what we know about the case, he has more life years than the 60 year old. On the graph the 60 year old has a lower probability of receiving an intervention than the 40 year old. The "defining feature" of the complete lives system is to "consider entire lives rather than events or episodes" (428).

However, reading the exchange between David and Jack, I'm not sure my interpretation is correct. And even if it is, I still agree that Jack is correct to say that their proposal is much less systematic than the alternatives they discuss.

Regarding David's point, I also find dubious the significance of parental investment. I'm not sure how worried they would be about the counterexample, though, because I suspect the contrast between infants and adolescents is supposed to be a general claim that could fix a policy of favoring adolescents (since, in most cases, they receive more investment).

Phil Reed April 23, 2020 at 4:35 PM

Saving the most lives in a situation where we cannot save everyone has to be the first value for any allocation decision. This is in fact what Emanuel, et al. argue in their recent paper in the wake of the COVID-19 global pandemic. Saving the most lives is consistent with treating people equally, since it does not discriminate against individuals except according to their ability to receive the thing we are trying to maximize, i.e. a life-saving resource.

Let us consider two kinds of attempts to de-prioritize saving the most lives in allocating scarce medical resources in a crisis situation. The first is Emanuel’s 2009 article co-authored with Persad and Wertheimer under discussion here. Their view explicitly de-prioritizes saving the most lives in favor of prioritizing the young on the basis of upholding the value of living a complete life. They deny that this policy is ageist, insisting that ageism occurs only if we treat a class of people differently “because of stereotypes or falsehoods” (429). While stereotypes and falsehoods add insult to injury, it’s hard to see why they would be strictly necessary for discrimination. They criticize using QALYs because it fails to treat people equally, but they neglect to mention that the only part of their complete lives system that values treating people equally is their use of lottery as a tie-breaker.

On my reading of Emanuel et al. 2020, he changes the view he develops in the 2009 article (to say nothing of his personal goal to die by age 75). In 2009, the objective was to maximize life-years. In 2020, the objective was to save the most lives, with the maximization of life-years being used only as a tie-breaker or subordinate aim. This difference between the two papers is not explicitly announced in the later paper, which confused some bioethicists ( and caused them mistakenly to read maximizing life-years into the recent paper.

Why does Emanuel change his view? He does mention in the Covid paper that there is “limited time and information” in an emergency to be assessing life-years. But I suspect also that when push comes to shove, i.e. when your policy is going to be looked at carefully and probably implemented, the controversial nature of discrimination against the old in allocating scarce medical resources becomes a lot less attractive.

The second attempt to de-prioritize saving the most lives comes from those who are (rightly) concerned about structural and societal injustices, especially as these play out in health care disparities. Serife Tekin says ( that protocols that maximize lives saved “are limited because they don’t pay attention to existing health disparities in society.” L. Syd M. Johnson argues ( that saving the most lives should be balanced by the goal “to promote justice by minimizing the added burdens and disadvantages for those…already disadvantaged by health disparities.”

These writers don't provide specifics on their allocation alternatives. Therefore, they seem to ignore the reality of the situation. If their proposals are taken seriously, it would result in the following situation: a person of color (who experiences systematic discrimination) who is unlikely to survive with a ventilator should be prioritized over a privileged person who is likely to survive with a ventilator because our allocation policy should correct for past discrimination. Most would find this just as unjust as what we are trying to correct.

Policies for allocation of scarce resources aren’t equipped to address past or systemic wrongs. They are meant only to do the most good in a tragic and limited situation. While it is important to acknowledge and bemoan that saving the most lives can entrench health care disparities, these are side-effects of the allocation policy and not integral to it. While these writers want to avoid a whole-sale utilitarianism, sometimes maximizing benefits is necessary. We should all want to end up with fewer dead rather than more.


David H April 23, 2020 at 5:40 PM


Although you are right that there isn't a talk of complete lives which would clearly prioritize the young in the 2020 paper, Wertheimer and Persad leave themselves some wiggle room when they claim to maximize lives "with a reasonable life expectancy." They leave "reasonable" unspecified. But since they are still contrasting their maximizing the number of patients saved against maximizing improvements in length of life of those saved so your point about a switch may still hold. But they may be still getting that earlier 2009 outcome with their appeals to prognosis. So I don't know if your skepticism is correct and the reason is as you suggest that it is worries about discriminating against the old and not just as they say due to time and information constraints in an epidemic. They do write about the prognosis considerations of recommendation #4 in the more recent piece " long the patient is likely to live if treated - which may mean giving priority to younger patients...this is consistent with the Italian guidelines that potentially assign a higher priority for intensive care access to young patients with severe illnesses than to elderly patients" (2020:5)

Jack Freer April 23, 2020 at 8:11 PM

David, yes, you are correct that I misinterpreted the paragraph on lexical order--it is, in fact, the third hypothetical objection. That said, there is not a lot of guidance in balancing CLS principles (Youngest-first; prognosis; save the most lives; lottery;

instrumental value, but only in public health

emergency). Ironically, the attractiveness of CLS is the number of morally relevant (non-flawed) principles it incorporates (five). While this bolsters its theoretical attractiveness, it complicates any practical attempt to balance competing principles. Instrumental value, in particular is problematic (and valid for CLS in public health emergencies, such as a pandemic). Who gets priority for a ventilator in COVID-19 pneumonia? The ICU physician? The ICU nurse? The personal care aide who cleans up the patient's bodily secretions?

David H April 23, 2020 at 9:36 PM


You are right. The unsystematic Whole Life System has just specified some intuitively attractive principles, stated they are insufficient rather than inherently flawed (and I am not even quite sure how to cash out that distinction), need to be combined, weighed, balanced, made specific and explicit, and said no more, leaving us within a maze (or haze) of intuitions. It seems they are to qualify the pursuit of complete lives. I think your examples of instrumental value show how hard it is to weigh factors within a principle (comparing different kinds of apples), and that probably pales between comparisons between the principles (comparing apples and oranges). Maybe Emanuel et al. believe they have set some parameters that future research will refine. Perhaps those philosophers who advocate some sort of value pluralism would have more to say about how to proceed. Perhaps the authors think we can't do much better than have some agreed upon well publicized principles in place, some consistent precedents or paradigm cases of trade offs, and a committee to employ these and expand on them rather than leaving the decisions to doctors to improvise in the heat of the moment at the bedside.

Jack Freer April 24, 2020 at 9:34 AM


As you point out, PWE try to sidestep the charge of ageism by stipulating that ageism only occurs if we treat that class of people differently “because of stereotypes or falsehoods.” I think this confuses prejudice with discrimination. There are two senses of "discrimination" (with the more benign being the simple act of differentiating a group and treating them accordingly). The other sense, of course, is basing such action on stereotypes etc. Prejudice leads to this second type of discrimination and is always destructive.

Since treating young and old differently is such an integral part of PWE's case for CLS (and EE's personal goals re: age 75 etc), the distinction between prejudice and discrimination demands more discussion and clarification. I'm not sure where that discussion would lead but it might well come to some uncomfortable conclusions about society's treatment of the elderly and disabled (not to mention the socio-economically disadvantaged).

Stephen Kershnar April 24, 2020 at 1:59 PM

I wonder whether there is moral space for these principles to do the work that Ezekiel Emanuel et al. want it to do.

Govind Persad, Alan Wertheimer, and Ezekiel Emanuel put together the Complete Lives System (CLS) to allocate scare medical interventions such as organs and vaccines.

Except when there is a consequentialist override, coercive rationing is wrong. To see this, consider organ donation. If a father wants to donate his kidney to his son, the government has no right to make him give the kidney to someone else’s child instead on the basis of social justice. A similar thing is true for a father who designs a vaccine for his son and generates a single dose via his 3-D printer. A similar thing is true for a physician’s labor.

To block this result, Persad et al. might argue that coercive rationing is permissible because people’s rights to organs, vaccines, or labor are overridden, undermined, or lost (via waiver or forfeiture). Given the dim view most of us have on involuntary organ lotteries and wartime drafts, even in times of catastrophe, I doubt such a position is plausible. We should take liberty seriously.

Alternatively, Persad et al. might argue they are only concerned with government rationing, although I doubt it given Emanuel’s track record. In this context, they misidentify the right-action maker. In the case of an individual, valid consent is required for permissible medical treatment (leaving aside incompetents). Along similar lines, when a collection owns (or has rented) the organs, vaccines, and labor, its valid consent is required for permissible production and distribution of medical goods and services. For example, there is nothing wrong with a group (for example, Japanese citizens) using their ventilators to save their own elderly rather than other people’s elderly. A collection’s valid consent need not track a principle or set of them any more than does an individual’s valid consent need do so. Thus, just as an individual’s consent need not track CLS, nor must a collection’s (for example, country or state) do so.

Here, then, is my argument. For simplicity, I set aside consequentialist overrides.

(1) If a government distribution of goods and services is permissible, then valid consent justifies it.

(2) If valid consent justifies a government distribution of goods and services, then CLS does not justify a distribution of goods and services

(3) Hence, CLS does not justify a distribution of goods and services. [(1), (2)]

Even if CLS were true, there is an issue as to whether it would be better satisfied through rationing or the market. Given the vast advantages of the market over socialism in providing more goods and services (consider the size of the pizza versus how to cut it up), CLS-considerations are probably more likely met if the government does not pursue them.

Stephen Kershnar April 24, 2020 at 2:08 PM

Jack, David, and Phil:

I agree with you about the implausibility of the principles, inadequate arguments for them, and problems with weighting the principles whether in general or in particular cases. Here is a way of seeing this objection.

Let us assume that the state has a right to distribute the medical goods and services and needs a non-market principle to guide it. How should it do so? Let us put forth a competitor to CLS. Consequentialism asserts that medical resources should be distributed to maximize the good. There are two types of consequentialist theories: welfarism and non-welfarism. Welfarism says that goodness is a function of, and only of, well-being. Non-welfarism says that welfarism is false. One version of non-welfarism incorporates justice. It asserts that intrinsic value is a function of, and only of, desert-adjusted well-being. Let us call this theory desert-adjusted consequentialism (DAC).

DAC has advantages over CLS. First, DAC explains and justifies the relevant factors. Consider, for example, age and prognosis. DAC justifies the use of these factors because they make people’s lives go better and make them get what they deserve. DAC explains and justifies the systematic consideration of desert. It does this by noting how justice applies to different distributions none of which infringe on rights.

In contrast, CLS makes an ungodly mess of justification. It does not adequately explain or justify why these factors apply and not others. Nor does it explain how to prioritize these factors. What can CLS say other than to cite isolated intuitions? It is not even clear if the intuitions occur at a more fundamental level (How to weight principles?) or at a less fundamental level (How to weight principles in deciding whether Al or Bob should get a ventilator?). CLS makes no attempt to systematically consider justice. For example, the theory does not make clear whether its concern for justice occurs via rights or desert (or equality or fairness).

Second, DAC allows for principled, and to my mind correct, trade-offs. Consequentialism trades off rationing-resources by whatever distribution maximizes desert-adjusted well-being. In particular, it tells us how to trade off factors such as age, instrumental value, prognosis, and QALYs. Consequentialism also trades off rationing-related resources against other the need for resources in other areas of medicine and other parts of the economy.

On a practical matter, economics guides DAC in determining correct trade-offs because money reliably tracks well-being. The tracking is not exact and does not take desert into account. Still, it is a valuable tool. CLS cannot use economics because it is not trying to maximize well-being, whether justice-adjusted or not. As a practical matter, the lack of economics will produce exactly the sort of corruption, discrimination, and waste with which Persad et al. are concerned. Again, if CLS were true, it might best be satisfied by focusing on DAC considerations.

Here, then, is a comparison of DAC and CLS.


(1) DAC justifies rationing via a theory of how to weight lives.

(2) CLS justifies rationing via isolated intuitions.


(1) DAC justifies trade-offs within rationing, between rationing and other parts of the medical economy, and between rationing and other parts of the non-medical economy.

(2) CLS justifies trade-offs via isolated intuitions.

Practical Considerations

(1) According to DAC, cconomics guide rationing. If adopted, there would likely be less corruption, discrimination, and waste. DAC likely best achieved by focusing on DAC-related factors.

(2) If CLS were adopted, economics would not guide rationing. There would be more corruption, discrimination, and waste. CLS likely best achieved by focusing on DAC-related factors.

Stephen Kershnar April 24, 2020 at 2:08 PM

One objection to this comparison is that we need to first determine whether DAC or CLS is true. This should be done in general and not in the specific context of rationing. I think this is right. I merely illustrate some of the things that make DAC more attractive.

Stephen Kershnar April 24, 2020 at 2:16 PM

Phil writes the following.

Saving the most lives in a situation where we cannot save everyone has to be the first value for any allocation decision. This is in fact what Emanuel, et al. argue in their recent paper in the wake of the COVID-19 global pandemic. Saving the most lives is consistent with treating people equally, since it does not discriminate against individuals except according to their ability to receive the thing we are trying to maximize, i.e. a life-saving resource.

I don't see why saving the most lives as opposed to maximizing QALYs is obligatory or why it is always good.

It is not obligatory, because this depends on rights. For example, one might want to save two 15-year-olds rather than three 95-year-olds with his personal stock of life-saving pills.

It does not maximize the good because it does not maximize total well-being, average well-being (well-being per life), or a diminishing marginal well-being function. I'm not following your reasoning. I'm not running as well as you either. Perhaps there is a similar cause.

Stephen Kershnar April 24, 2020 at 2:18 PM

Jack writes the following.

"Instrumental value, in particular is problematic (and valid for CLS in public health emergencies, such as a pandemic). Who gets priority for a ventilator in COVID-19 pneumonia? The ICU physician? The ICU nurse? The personal care aide who cleans up the patient's bodily secretions?"

I am not seeing your argument. If saving one physician will cause the saving of two people down the road, then his life has a total value of three lives (his own + the two people he will save). So, priority is given to the total value of a healthcare worker, specifically, the value of his own life plus the value of the lives he will save.

How we know someone's instrumental value is another issue.

Stephen Kershnar April 24, 2020 at 2:22 PM

For David, I have the following question.

An illegal alien teenager sneaks into the country in the dead of night. She was very sick and in need of a ventilator before she arrived. Does she get to cut to the front of the ventilator line ahead of the 45-year-olds that are citizens and paid for the ventilator?

In general, does the country have to distribute ventilators different from how a country club would distribute them?

Thanks to the three of you for the interesting comments.


David H April 27, 2020 at 1:57 PM


I don’t know why this question is put to me since I didn’t post anything on the topic, have never published on distribution medical resources, nor, alas, even read much about it as my focus has been on other issues in bioethics. Maybe you just assumed that I have some wishy-washy liberal beliefs about a universal right to health care or an internal morality of medicine in which doctors should prioritize saving the medically neediest and you are ready to pounce if I express such views. (Actually, I have wishy-washy hopes that a universal right to health care can be sustained as I haven’t worked on the issue to the point where I have beliefs that it can be defended in a rigorous manner.) I am just hopeful that something like a right to health can be given a principled defense against your claims that others don’t have a right that we save them or pay for their health care). Well, having made some excuses for why I shouldn’t be criticized too harshly for the following post being worthless, I will take the bait. Let’s see how far I can go with a duty to rescue even without something more systematic in terms of universal right to health care.

I think a non-citizen who shows up at United States hospital should be treated. Here’s an analogy. Some guy who is not a citizen of New York but is from Jersey, say Exit 109, commits a crime in New York. It is not a federal crime or a crime deserving death, so he is sent to a state jail in upstate New York near Buffalo where he gets sick and is sent to the criminal wing of a state hospital. He is sicker than another patient who is a resident of New York and so gets the scarce resource which would have helped the New Yorker who has a better chance of surviving without it. The New Yorker’s state taxes pay for the hospital. Most of us don’t think out of state criminals forfeit the right to life or any rights to medical care when a ward of the state in a state in which they don’t reside or pay taxes. Then why should a someone whose crime was entering the country get less medical care if Jersey guys don’t get less medical treatment in NY? Is it that the states would have some reciprocity that countries don’t? Or is it that the illegal alien is not in jail and a ward of the state but he could become one. Should it matter that he is not yet arrested and a ward? Surely not. It would be odd that if border patrol agents arrested him in the hospital that was not treating him then he would get to stay there and receive medical services. Your view may be to put him on a stretcher and dump him over the border. But it turns out he is not Mexican but South American and can’t be quickly repatriated and is a ward of the state for a much longer time.

Your example was that the illegal alien came here for medical treatment and wasn’t here illegally for economic reasons. That raises worries and feeling about being taken advantage of that perhaps other illegal alien scenarios doesn’t. Resources are limited and so we can’t supply the world and we don’t want to incentivize runs on American hospitals. I suspect that there are reasons to favor citizens (some efficiency based, some fairness based, some just politically expedient) that are probably consistent with also providing less but some health care to non-citizens. Take an analogy of wartime. Captured soldiers, even those whose countries aren’t signers of war accords or humanely reciprocating, should receive medical treatment, but lose out to American soldiers when there are shortages. These reasons need not be that we want to get the American soldiers back to the front based on other considerations probably having to do with group membership and fairness and other factors. So, while I don’t have anything worked out, I am partial to prioritizing Americans, without denying medical resources to non-citizens. What are the principles for such waiting lists and prioritization? I don’t have anything worked out. I am just hoping the analogies of the states and wartime have some promise.

Stephen Kershnar April 28, 2020 at 2:45 PM

David and Bob:

Many good points. I directed at you guys because I assumed you had some sympathy with Robert Nozick's argument against end-state and non-historical patterned principles of distributive justice.

You seem to put forth the following theory.

(P) There is a just rationing principle and it allows for lower prioritization for illegal aliens.

This theory allows for lower prioritization for illegal aliens, but, perhaps, not denying them medical resources altogether.


I am guessing that this principle, (A), allows the state or a private medical group to deny illegal aliens organs, vaccines, or ventilators when there are not enough of them and every one will be used to save a life. Who are you, the Grinch? kidding.


This lack of prioritization suggests that rationing principle can be adjusted by contract (promise or consent) of the citizens or group membership (perhaps, an associative political obligation).

This moves us away from the sort of contract- and group-membership independent principles put forth by Emanuel et al.'s principle, QALY, lottery, etc.

That is, the price for this change is to make a rationing principle frequently overridden or undermined or, alternatively, to make such a principle unjust.

Here is a Robert Nozick-like argument.

(1) If there is a just rationing principle, then it is distinct from a standard theory of justice in acquisition (JIA) and justice in transfer (JIT).

[Note: JIT might involve a democratic principle.]

(2) If a rationing principle is independent of a standard theory of JIA and JIT, then it is a just end-state or non-historical patterned principle.

[Note: The end-state or patterned principle might be a complex one such as a mathematical version of the priority-for-the-worst-off principle.]

(3) There is no just end-state or non-historical patterned principle.

[See, for example, Hershenov's intuition - I'm guessing Kelly has the same - that illegal aliens get lower prioritization for lifesaving organs, vaccines, etc. Note this does not appear to be an override of a just rationing principle because, intuitively, there does not seem to be someone who is owed an apology or compensation - that is, a residue duty.]

(4) Hence, there is no just rationing principle. [(1)-(3)]

If (4) is true, and I think it is, then Emanuel et al.'s whole project had to fail. This is independent of whether, as I argue above, it is clearly inferior to a desert-adjusted-consequentialist principle.

Jack Freer April 25, 2020 at 6:53 AM

Steve, certainly the physician is the one who gets the credit for the "save" but (as nurses will tell you), it was not a solitary act by the doc. The analogous argument can be made for the personal care aide who changes the patient's diarrhea drenched gown, or the housekeeper who cleans the ICU unit. I don't know many docs who will grab a gown and mop to pitch in and clean the patient and environs. Therefore, those jobs are instrumental in allowing the docs to do their jobs. As you point out, there is no universally accepted means to establish or compare value for those activities (no matter who performs them).

Stephen Kershnar April 25, 2020 at 9:47 AM


I like your point, but have to disagree. The same problem occurs with regard to picking out the MVP in sports league. For example, we try to measure how much two running backs contribute to wins above their likely replacements.

Here I would separate two issues.

(1) Metaphysics. How much did an individual contribute? Here the baseline is against the likely replacement. Because this is sometimes hard to determine, sports leagues use an average replacement value to get Wins Above Replacement Player.

(2) Epistemology. How do we measure contribution? Here we can use an analogue to WAR in the context of medicine. Perhaps we don't do this statistics, but we could in theory and likely in practice.

We need to do something like (1) and (2), albeit in economic terms, to determine the salaries of ICU docs compared to personal aides. They judgment is similar, although more focused on economics.


Steve K


Jack Freer April 26, 2020 at 7:52 AM


I agree with both of your critiques (metaphysical and epistemological). There are probably other aspects of evaluating contribution (including theatrics, since "top docs" are often brilliant showmen). Rather than help focus on the correct individual to credit with instrumental value, your arguments discredit the whole notion of including this factor (in an already crowded field of possibilities).

Allocation of scarce resources has been a focus of clinical ethics for over half a century. Indeed, the inception of healthcare ethics committees was a response to decisions about too few dialysis machines to treat kidney failure. It has remained unresolved despite an army of bioethicists addressing the problems. Identifying a just, workable allocation scheme that is consistent with the medical realities (especially poor prognosis) is as remote now for ventilators as it ever was for dialysis. I think PWE's analysis is still useful in promoting reflection on the meaning of a complete full life, but this is probably best done on an individual basis in private.

Stephen Kershnar April 26, 2020 at 10:13 AM


As always good points.

(1) I think the individual contribution can be measured my marginal contribution. This requires a baseline. This has some theoretical problems, but is a standard, and to my mind, largely correct way to look at an individual's contribution. I think the baseline is the nearest world in which an individual does not do a job.

(2) I think we have a workable and correct principle to allocate organs, machines, medicine, etc. Here is the principle: The owner may give the good to whatever individual he wants. This sounds trivial, but it is not. It means that the choice of recipient is a moral free zone (except in the case of consequentialist override). This is similar to the price of a good in the free market.

I don't think there is a correct guidance principle, but if there were, it would be the following: You ought to allocate a good to maximize utility.

I hope your weekend is going well.

Stephen Kershnar April 25, 2020 at 9:56 AM

David, Jack, and Phil:

My concern about illegal aliens takes the form of a dilemma: Consider illegal aliens teens (assume there are five thousand of them) snuck into the country in the dead of night, need organs (ones that are in very short supply), and would go to the front of the line given any plausible rationing principle (e.g., QALYs, prognosis, young-first, or prioritism). Should they go to the front of the line?

Horn #1: Yes. This is absurd. You don't get to live instead of ten thousand 45-year-old Americans who donated (or their families donated) the organs, who pay for the surgery, whose country it is, and so on.

Horn #2: No. If so, then there is no plausible principle or set of principles for rationing. Hence, a just rationing principle is whatever the owners of something choose.

By analogy, Robert Nozick argues convincingly that liberty destroys end-state and non-historical patterned principles of distributive justice. A similar thing is true for ownership.


Steve K


Bob Kelly April 25, 2020 at 5:25 PM


Why doesn't your set up of the dilemma take us directly to the first horn, and also undermine its even being a horn. In other words, it seems pretty straightforward from your set up ("and would go to the front of the line given any plausible rationing principle") that, yes, they should go to the front of the line. Moreover, I suggest that this is exactly what you told us should happen in the description, and hence, why would it be absurd? You specifically said that on any plausible rationing system, then should go to the front of the line. This either accounted for ownership (if it is part of any plausible system) or it didn't (if it isn't). If the former, then horn #1 is where we obviously go, and this isn't absurd but simply directly follows from the set up. If the latter, we still go to horn #1, and it isn't absurd because ownership is an implausible consideration when thinking about rationing (per the set up). So, you have your own dilemma to deal with, which leads us, non-absurdly, to horn (or more like nice comfy cushion) #1.

Stephen Kershnar April 26, 2020 at 10:06 AM


Good point. My argument is that the following is true.

(1) Illegal aliens who just snuck in should not go to the front of the line.

(2) If these sort of rationing principles (rather than whatever people consent to) should govern distribution of private medical goods and services, then the illegal aliens should go to the front of the line.

(3) Hence, these sort of rationing principles should not govern distribution.

Perhaps the dilemma was not a helpful way to consider the problem.

The basic idea is two-fold. First, ownership gives complete moral freedom with regard to distribution (outside of infringing someone's right). Second, even if ownership did not give such freedom, it should take into account those with whom we have a special relationship (for example, family, friends, and fellow citizens).

Thanks for the helpful points.