1. Introduction
The method of economic cost–benefit analysis (CBA) could be said to have a monopoly when it comes to informing policy decision-making. It has been extended to accommodate a variety of challenges, including policy decisions involving significant tradeoffs across people and time or else between very different types of goods. In a number of books and papers, Matthew Adler submits this popular method to careful scrutiny. He positions CBA as a mere heuristic or approximation of a proper social-welfare approach to policy decisions, and accordingly examines the advantages and limitations of CBA.
Adler’s 2025 book Risk, Death, and Well-Being: The Ethical Foundations of Risk Regulation might be placed in this tradition. The book focuses on policy decisions for which changes to quality of life must be balanced against changes to length of life, or what Adler refers to as “fatality risk regulation” decisions. This is a pressing policy challenge in the wake of the global pandemic and new climate-related concerns about public safety. The trajectory of the book involves presenting “Lifetime Welfarism”—a special form of welfarism or way of assessing outcomes in terms of the “lifetime well-being” of individuals (more on which below)—as a defensible approach to assessing policies that implicate longevity or “fatality risk” (Chapters 1–3), then showing how such an approach can be operationalized (Chapters 4, 5, and 7), and further showing how the resulting policy assessments can differ from those of the popular CBA approach, to the latter’s detriment (Chapter 6). Finally, it is shown that Lifetime Welfarism is very flexible in its ability to account for a variety of intuitions in difficult life-and-death cases (Chapter 8). The book makes a significant practical contribution by laying out a set of viable and more defensible alternatives to CBA that are tailored to the new policy challenges that we face.
What is more, the book also makes a significant theoretical contribution. Adler presents his alternative welfarist approach to CBA in a brisk, no-nonsense fashion, and yet he succeeds in opening up a whole host of provocative, philosophical questions for welfarism. The notion of “lifetime well-being” is often taken for granted as the basic unit of analysis in discussions of the social evaluation of outcomes. That is, it is assumed that a person’s life can be treated as a whole, such that promoting their interests or well-being at any particular time is valuable only insofar as it promotes the total well-being of their life as a whole. This means that egalitarian considerations would apply to the distribution of lifetime well-being and not to any smaller temporal unit of well-being. But when we look more carefully at decisions that have implications for longevity, we see, with Adler’s help, that there are many difficult questions and related ambiguity about the nature of lifetime well-being and its moral significance.
In this review essay, I begin by considering Adler’s commitments, in defending Lifetime Welfarism, regarding the nature of human lives and how we should conceive of well-being across these lives (Section 2). Then I turn to his strategy for comparing Lifetime Welfarism and CBA in practical contexts of fatality risk regulation (Section 3).
2. A Tidy Package: An Account of Lifetime Well-Being Fitted to Lifetime Welfarism
In defending “Lifetime Welfarism,” and how such an approach can be used to assess policies that have implications for longevity, Adler focuses on a highly rarefied population of persons and associated set of options. The policy options are such that the same (finite) population exists in each of the possible outcomes of the available options. Moreover, those who comprise this fixed population are what Adler (p. 2) refers to as “ordinary human persons,” in that they, “after birth, undergo the typical processes of brain development that lead to their acquiring an array of psychological properties that are jointly sufficient for personhood, properties that are typically possessed by adult humans; and who retain these properties, without breaks in intertemporal psychological continuity, until death.” The death of “ordinary human persons” is also clear-cut in that it satisfies all the plausible competing accounts of death (see pp. 76–78).
Adler, moreover, introduces some assumptions in the opening chapter that facilitate, he suggests, a simple and elegant account of Lifetime Welfarism for this restricted domain or “Focal Case.” The first assumption concerns how human persons are delineated in time and space. Adler contends that simplicity favors the so-called “Animalist” position, which takes human persons to be numerically identical to the human animal that develops the properties of personhood. The human animal begins their existence sometime between conception and birth and ends it at death. The second assumption concerns what aspects of a human person’s life contribute to, or are “integrated into,” their lifetime well-being. For the most part, Adler assumes that lifetime well-being covers all of a person’s life, including, for instance, the time in which they were an infant and had not yet developed the properties of personhood. I will refer to this as the assumption of “Full Integration.” As Adler puts it, his assumption for an “ordinary human person” is that the “age of integration” is zero (and the period of integration is zero until the age of death). He acknowledges that it is not obvious that this is justified, and the issue may depend on the nature of well-being (whether experientialist or otherwise, subjective or objective), something that he remains agnostic about in the book.
Do Adler’s domain restrictions limit his analysis and lead to complacency about substantive metaphysical assumptions? One might worry that the Focal Case excludes the really interesting and important boundary cases for reckoning with lifetime well-being and its moral significance. Adler himself says that the book is not about many things: it is not about how to balance trade-offs concerning the well-being of humans vis-à-vis non-humans (p. 2) or the permissibility of abortion at different stages of fetal development (p. 76). It is also not primarily about when death occurs and whether a life can be extended through death-defying artificial technologies (p. 78). Nor is it primarily about the moral difference between the death of a baby and that of, say, a teenager, or about how we should conceive of the well-being of someone who has developed dementia or some other psychological impairment (although these issues are addressed in Chapter 8). Adler stresses that the book is rather about policy decisions that affect fatality risk due to pollutants, transportation dangers, or other public health hazards. The sorts of issues just raised are not central to these decisions. That’s a reasonable defence of the restricted focus of the book. But I think more can be said. A great strength of the book is how it shows that, to make progress on any of these issues, we need to first address the deep questions about longevity and human well-being that arise even for a fixed population of “ordinary human persons”.
In that spirit, it is worth dwelling on Adler’s starting assumptions of Animalism and Full Integration. Some readers might feel uncomfortable with the brisk defense (in the opening chapter) of what appear to be substantive metaphysical assumptions. But I suggest that the way to think of these assumptions is as forming a package with Lifetime Welfarism. The assumptions of Animalism and Full Integration make (a strong version of) Lifetime Welfarism plausible. Even as a package, the three positions could be said to merely provide a framework for thinking about human well-being, as opposed to marking out a substantive position on human well-being.Footnote 1 Individually, they are even less substantive (since deviations in one position can be counter-balanced by corresponding deviations in the other two positions). To see this, it is worth examining them in turn.
2.1. Animalism
As noted, Adler assumes that “ordinary human persons” can be identified with a human animal, even if they only assume the characteristics of personhood some way through their animal development. All human persons are human animals, even if not all human animals are human persons (for instance, those that never develop to the point where they assume the properties of personhood).
Adler defends Animalism on grounds of (mere) simplicity. This is initially puzzling, since Animalism seems to carve up the moral world rather differently from a rival like Personalism, which holds that a human person begins only at the time of acquiring the psychological properties of personhood and ends when the continuity of these psychological properties is lost. On reflection, however, one might see that nothing normative hangs purely on the boundaries of human persons; it is only in combination with further assumptions about the full set of bearers and units of well-being that we get a welfarist moral theory.
Adler defends Animalism for its ontological simplicity with respect to the number of (kinds of) thinking beings that it entertains. This is apparently a well-known advantage of Animalism—that it provides a satisfying response to the “How Many Thinkers?” question (p. 14)—since it does not posit thinking human animals in addition to thinking human persons coinciding in space and time. Adler goes further in defending Animalism for its conceptual simplicity, in the sense that it is highly flexible and leaves open a range of substantive moral positions. Whether or not that is so, it does seem as good an anchor or fixed point as any for building a welfarist moral theory.
Animalism plays a further, more substantive role that neither Adler nor those he cites seems to emphasise. It provides a plausible account of how to identify the same person across different possible “worlds” or outcomes in which the lifespan of that person (amongst other things) may vary. The domain restrictions that Adler introduces presuppose that this can be done: the outcomes for his Focal Case all supposedly contain the same population of human persons. According to Animalism, we can identify the same human person in different worlds by the criteria that make them the same human animal, say, the fact that the same fertilised egg (or whatever is the appropriate set of biological conditions) is the starting point for development. Any alternative to Animalism would similarly need to furnish an account of how to identify the same human person across different worlds or outcomes. In discussing the ontology of human persons, Adler, in line with others, emphasises persistence conditions. He says (p. 11): “A general premise in the literature on the ontology of human persons (one I have no reason to dispute) is that each individual being (and specifically each human person) is characterized by certain persistence conditions (also referred to as ‘essential properties’).” But persistence across time is not enough. To fully characterize individual human persons and other beings, especially in the context of decision making under uncertainty, we must also attend to the conditions for their persistence across different possible worlds or outcomes.
2.2. Full Integration
Adler further assumes, more for convenience, that the entirety of the animal lives of human persons contributes to their lifetime well-being, even those parts that occur before the properties of personhood are developed. In this way, the idea that the full lives of a fixed population of (“ordinary”) human persons matter morally is made to fit with the idea that only the lifetime well-being of these persons matters morally, at least as far as welfare is concerned (as per Lifetime Welfarism).
Even when restricted to the Focal Case, however, this tidy package of views may be acceptable only with qualifications, depending on one’s further commitments regarding the determinants of well-being and the demands of equality. For instance, given that the life of an “ordinary human person” will contain infant and adult stages that have differing properties, different kinds of well-being may be applicable to these person stages. It may not be possible or desirable to then measure lifetime well-being on a single cardinal scale; a vector may be more appropriate. (When the measure of a person’s lifetime well-being is a vector, there are also more possibilities regarding egalitarian considerations in the evaluation of outcomes and policies consistent with Lifetime Welfarism. For instance, there are different ways to identify the worst-off overall, and it is open whether concern should be given to the worst-off overall or rather to the worst-off with respect to a particular component of lifetime well-being, for instance, that which corresponds to the period of continuous personhood.)
If lifetime well-being did not account for all of a human person’s life—if Full Integration was denied and the “age of integration” considered greater than zero—then Lifetime Welfarism would not be plausible, even for the restricted domain of outcomes involving a fixed population of “ordinary human persons” (as noted on page 73 and elsewhere). Adler suggests that this would require a simple weakening of Lifetime Welfarism to the effect that only when all else is equal, and not in all circumstances, the social (welfarist) evaluation of outcomes depends exclusively on lifetime well-being. For the full range of circumstances, lifetime well-being would effectively be replaced by what Adler refers to as “long-term well-being,” of which lifetime well-being is just one component. These very similar formal structures would presumably allow for the same set of possibilities for the overall evaluation of outcomes.
2.3. Lifetime Welfarism
So we see that with Animalism and Full Integration in place, a strong version of Lifetime Welfarism is an entirely plausible constraint on the social evaluation of outcomes. Otherwise, only weaker versions of Lifetime Welfarism are plausible. Either way, the package of positions that is selected might be thought to provide a structure that allows for more substantive debate about the appropriate welfarist evaluation of outcomes and policies.
Formally, Adler shows how (stronger and weaker versions of) Lifetime Welfarism can be defined in terms of Pareto Principles. For instance, the basic Pareto Indifference Principle requires that if two outcomes are equivalent with respect to the lifetime well-being of each of the human persons in the population (and all else being equal with respect to other potential sources of well-being, as well as other morally relevant factors), then the two outcomes are equally good. A further Pareto Dominance principle ensures that the social evaluation is moreover responsive to purely positive changes in lifetime well-being.
Adler defends Lifetime Welfarism (in Chapter 2) against so-called Momentary or Stage Welfarism (especially where these include a corresponding Anonymity axiom). Given that Lifetime Welfarism is presented in weaker or stronger forms, one might wonder what exactly is being defended, in the end. I suggest the following: those periods or parts of a human life that are considered to be integrated (at least, those parts integrated in personhood) are the objects of a single well-being measure. This well-being measure may be more than a simple aggregation of the well-being for the moments or stages that comprise the relevant period. For instance, the pattern of well-being across the entire period may matter. Moreover, what matters for the social evaluation of outcomes, as far as well-being for these periods or parts of a human life are concerned, is just the overall well-being for the period.
The book examines some specific versions of Lifetime Welfarism (whether the stronger kind that accounts for whole human lives or the weaker kind that accounts just for the personhood component of a life, with all else assumed equal). The versions that are defined and discussed are the familiar ones: Utilitarianism, Prioritarianism, Leximin, Sufficientism, Egalitarianism. These forms of Lifetime Welfarism are also extended to uncertain prospects (since these are the typical objects of choice in a policy decision scenario). It is made clear that here there are further questions as to whether the evaluation of uncertain prospects should be risk neutral, and also whether uncertainty should be resolved at the individual level (the so-called ex ante approach) or rather at the social level (the so-called ex post approach). Throughout the book, the default combination of positions adopted for handling uncertain prospects is the ex post approach and risk neutrality (more on which below).
3. Lifetime Welfarism versus Cost–Benefit Analysis in Practice
Let us now turn from the theoretical to the practical dimension of the book. A potential difficulty in the overall dialectic is that while theoretical flexibility allows for a stronger defense of Lifetime Welfarism, this flexibility does not seem so helpful or pertinent when it comes to positioning Lifetime Welfarism as a practical alternative to Cost–Benefit Analysis.
The contention of the previous section is that Adler’s Lifetime Welfarism is indeed extremely flexible. He lays out a framework that is extremely non-committal when it comes to how outcomes and policies should ultimately be evaluated. We see this even for the domain of “ordinary human persons”. Adler furthermore shows (especially in Chapter 8) that his framework has all the ingredients to handle a wide range of intuitions about difficult cases beyond this restricted domain. (For instance, in Section 8.2, he proposes a way to account for the death of a teenager being a greater social loss than that of a baby.) Moreover, it seems right (as Adler suggests at the outset and in the closing chapter of the book) that the framework can be extended in any way one might wish to an even wider domain, for instance, to the social evaluation of outcomes that vary in size and that include both humans and non-humans.
The potential problem, however, is that Lifetime Welfarism is thereby shown to be not one thing, but rather many: it is a broadly encompassing class of approaches to the social evaluation of outcomes and policies. How then can it be said that Lifetime Welfarism is superior to Cost–Benefit Analysis? Strictly speaking, this cannot, I think, be said. Rather, what can be readily said is that the plausible candidates for the social evaluation of outcomes and policies are versions of Lifetime Welfarism. An application of Cost–Benefit Analysis can only provide a mere approximation of any of these moral evaluations. Adler goes further than this with respect to the problem of fatality risk regulation. He says (p. 142) that he seeks “to demonstrate how welfarism gives systematic and specific guidance in making the difficult trade-offs involved in selecting policies to reduce fatality risk.” He further suggests (in Chapters 4–7) that models of Lifetime Welfarism of the kind he proposes strike a good balance between accuracy in moral evaluation and tractability (and a much better balance than does standard CBA) when it comes to decisions about fatality risk regulation.
Whether or not Adler fully succeeds in his argumentative aims, his models are extremely illuminating regarding how to make policy decisions about fatality risk tractable, in a way that surely better tracks the true social value of policies than does the standard CBA approach. In what follows, I draw attention to some of the illuminating aspects of this modelling exercise.
3.1. Modeling and evaluating lives, outcomes and policies
If we want to evaluate trade-offs between quality and length of life, it is important to depict a life as having a temporal structure, with particular goods or events occurring at different times. In Chapter 4, Adler presents a simple model of this sort (that conforms with Animalism and Full Integration). An individual’s life is described by a lifetime bundle of attributes on which the measure of lifetime well-being depends (whether intrinsically or instrumentally). The lifetime bundle involves a sequence of discernible temporal periods (say, periods of one-year duration), that are each associated with a bundle of period attributes (which might include, for those times in which the individual is alive, things like income, health status, and environmental quality). The lifetime bundle might in addition include further information, depending on the notion of lifetime well-being in play, like the individual’s “global” preferences over possible sequences of period bundles.
Adler lays out the conditions under which the ordering of these model lives may be measured by a cardinal lifetime well-being function. He, in fact, describes two such methods and associated conditions. One is the more familiar von Neumann–Morgenstern method that appeals to a goodness ranking over a rich set of lotteries, the possible outcomes of which range over the various model lives (lifetime bundles) for an individual. The other is the less familiar method that appeals to an ordering of the differences in goodness of these possible model lives. The two methods are shown to coincide (as is assumed in the book) just in case the ranking of the von Neumann–Morgenstern lotteries is risk neutral with respect to the goodness of the model lives they might yield as outcomes. Adler further illuminates special kinds of lifetime well-being measures, like those that rest on Time Separability: the strong assumption that a given period bundle makes a fixed contribution to the goodness of a life, regardless of its temporal location and the other period bundles which constitute the life.
In Adler’s model, the social evaluation of an outcome is a function of the lifetime well-being, thus measured, of the individuals populating the outcome (as per the strong version of Lifetime Welfarism). Since policies amount to probability distributions over outcomes, any given social welfare function must be coupled with an “uncertainty module,” as Adler puts it, to enable the evaluation of policies. Chapter 7 defends risk-neutral versions of, specifically, Utilitarianism and Ex Post Prioritarianism as the most promising forms of social welfare function that could be used in practice to assess policies. These social welfare functions differ in their egalitarian commitments, but otherwise share a number of properties that have theoretical and/or practical virtue. One such property is Dominance (Section 7.2), which roughly says that if all the possible outcomes of one policy are better than those of another policy, then the former is the better policy. Another such property is what Adler calls Policy Separability (Section 5.1), which roughly says that the evaluation of policies does not depend on the fate of other persons (say, those who lived in the past) who are unaffected by those policies.
3.2. The social value of risk reduction
The book facilitates a direct comparison between Lifetime Welfarist methods and Cost–Benefit Analysis by showing how the former can be defined in terms of a simple model of possible lives and associated measure of lifetime well-being. Moreover, we see that the Lifetime Welfarist methods can deliver the evaluations—what Adler calls “the social value of risk reduction” (SVRR)—that facilitate the comparison of options as they are typically conceived in a fatality risk decision problem. The SVRR is “the social value of a small change in individual fatality risk” ((p. 143); it depends on the individual’s baseline age and their past and prospective income, health, and other well-being attributes. The optimal policies, relative to the status quo, are the ones that lead to the fatality risk changes for individuals that, in view of their respective SVRRs and the size of the risk changes affecting them (in addition to any costs), maximize aggregate social value.
One might say that the key to a fair comparison of CBA with welfarist alternatives is to disentangle the model of policies from the notion of social value used for assessing these policies. For fatality risk regulation, in particular, it might seem that one needs to buy the entire CBA method: the modelling of policies, together with the account of social value. But Adler succeeds in showing that this is not so. We can embrace the modelling innovations of standard CBA that are practically useful—including modelling lives as bundles of attributes, and modelling policy options in terms of changes in fatality risk relative to the status quo—while appealing to a more defensible account of social value (and selecting the relevant attributes for the lifetime bundles accordingly).
Adler discusses how policies may be efficiently described as follows (p. 152): they confer on each individual a risk profile (chance of survival in each time period going forward) and a sequence of period bundles (describing the attributes that would be had in each time period, were the person to survive it). Together, this information suffices for determining each individual’s lottery, under the policy, for different possible full lifetime bundles. The status quo is one such policy. As noted above, the options under consideration in a fatality risk decision are typically changes in risk profile relative to the status quo. In particular, the options are typically changes in the risk of (not) surviving the next period. That is where the social value of (fatality) risk reduction comes in. It tells us the social value of each unit of reduction in an individual’s not surviving the next period.
Adler’s Lifetime Welfarist methods arguably provide an improved way of determining the social value of (fatality) risk reduction: one based on changes in what really matters, namely, social welfare. On Adler’s methods, the social value of reducing the risk (by one unit) of a person dying in the next time period depends (in a Utilitarian or a Prioritarian way) on the change in expected lifetime well-being for that person. (For simple models where period bundles are constant for the periods in which the person is alive, the change in their lifetime well-being will be a function of the fixed period bundle and the change in the person’s expected longevity.) On the CBA method, by contrast, the social value of reducing the risk (by one unit) of a person dying in the next time period depends on their willingness to pay or accept compensation in money for that risk reduction. (Strictly speaking, as Adler discusses in Chapter 6, this is the change in money for some period that “balances out,” for that person, their change in expected longevity, in the sense that the person is indifferent between the original and the money-plus-longevity-adjusted lifetime bundles.) The problem for CBA, which Adler’s Lifetime Welfarism seeks to address, is that a person’s willingness to pay in money for a change in expected longevity is a very poor measure of the change in his/her expected lifetime well-being. Accordingly, aggregate “willingness to pay” across affected persons is a very poor measure of the social value of competing fatality risk regulation policies.
4. Concluding Remarks
This book succeeds in turning the examination of a special and pertinent kind of social decision problem—fatality risk regulation, or the extent to which we should pay in quality to increase length of life—into a profound reflection on both theoretical and practical challenges for welfarism. On the theoretical front, the reader is invited to reflect on the extent of unity in a human life over time, and how this bears on the assessment of human well-being and the appropriate target for egalitarian social concerns. On the practical front, the reader confronts the question of how to translate a theoretical account of social value into a practical method for comparing outcomes and policies.
Adler presents his Lifetime Welfarism alternative(s) to standard Cost–Benefit Analysis in a way that provokes one to think about what makes for an adequate method for assessing actual policy options. Under what conditions does a simplified model of the social value of some real-world policy option adequately account for the actual social value of that policy option? Adler does not provide all the answers. But he makes significant progress on the underlying questions and shows just how rich and philosophically interesting they are.
Acknowledgments
Many thanks to my graduate seminar group at the ANU for lively discussion of this material: Rachel Cao, Thomas Graham, Ned Lis-Clarke, Darryl Mathieson, Corey McCabe, Will Moisis, and Theo Murray. I also benefited greatly from discussing this material with Kirsten Mann, and with colleagues in health economics, especially Emily Lancsar and Cam Donaldson. Finally, thanks to Matt Adler for provoking these and ongoing discussions!
Katie Steele is a Professor of Philosophy in the Research School of Social Sciences at the Australian National University. Her research concerns norms for decision making, both the ethical and epistemic dimensions.
Open access
