Care as altruism: interdependent utility functions

Since Arrow's (Reference Arrow1963) article on the economic features of medical care, medical ethics has been acknowledged in standard health economics as means of tempering ‘consumer exploitation’ arising from information asymmetries in the supply of health care. The most prominent interpretation treats medical ethics as both an assurance and insurance against moral hazard associated with behaviour by the agent-physician that is self-interested and misaligned to and damaging to the interests of the principal-patient. It also moderates the potential for adverse selection, which reflects the information asymmetry between buyers (patients) and sellers (physicians) that implies caveat emptor (see Mooney and Ryan, Reference Mooney and Ryan1993; Williams, Reference Williams1988). McGuire (Reference McGuire, Culyer and Newhouse2000, 521) argues:

McGuire is explicit that the physician utility function consists of net income accruing to the clinician and the benefits/utility received by the patient. Standard microeconomic accounts of altruism view the benefits accruing to another as an argument in the individual's utility function as a special type of preference having the form: agent X's preference for satisfying agent Y's preferences (Folbre and Goodin, Reference Folbre and Goodin2004). As Khalil (Reference Khalil2003: 116) expresses it, the altruist (qua charity) lowers ‘his interest in order to buttress the recipient's interest’. Yet there is a tension here. Mainstream economics assumes that scarcity is ubiquitous. Famously Arrow (Reference Arrow1972) in his critique of Richard Titmuss’ (Reference Titmuss1971) study of altruistic blood donation used the conventional scarcity assumption to argue that a heavy reliance on altruism – or other similar sentiments – in the supply of a good or service, would deplete the ‘resource’. Under Arrow's formulation the price mechanism remains the dominant force as it rations scarce commodities and resources efficiently without diminishing altruism. Therefore, altruism, according to Arrow, should be confined to those practices and instances where the price mechanism ‘breaks down’.

The standard framework, in fact, can be seen as distinguishing three approaches to altruism: ‘egoistic’, where altruism revolves around the expectation of future benefits accruing to the benefactor; ‘egocentric’ (which resonates with Becker's approach) where the donor's utility reflects the utility of beneficiaries; and ‘altercentric’ where altruistic actions are associated with a personality trait (Khalil, Reference Khalil2003).

Health economic accounts of ‘humanitarian spillovers’ between the wider population and those who are ill (Culyer, Reference Culyer1976), or care interventions as ‘spillovers’ on ‘significant others’ (Bobanic et al., Reference Bobinac, Job, van, Rutten and Werner2010) correspond to the Becker-inspired ‘egocentric’ orientation. A ‘caring externality’ in the representative agent's utility function generates a benefit flow to this agent (Mooney and Ryan, Reference Mooney and Ryan1993) from the knowledge that other members of the population have the ability to access health care regardless of their ability to pay.

A variation on the foregoing is presented by Wiseman (Reference Wiseman1997), who associates process utility with care in arguing that utility may be derived from the act of caring or giving, as well as the consequences of care-giving. Drawing from Margolis’ (Reference Margolis1982) fair shares model, Wiseman notes that individuals may be conceived as possessing two utility functions: one deriving from ‘selfish utility’ and the other from group or social activities. In recommending this approach, Mooney (Reference Mooney2009) invokes Sen's (Reference Sen1977) reference to ‘commitment’ in articulating the idea of a social welfare function ‘independent’ of an individual's utility functions, so that the social welfare function is no longer simply the aggregation of individual utility functions. This dual utility function approach may be viewed as a heuristic device to account for how seemingly irrational acts, seen from a strictly selfish perspective – such as altruistic care giving at personal cost – can be explained as rational acts consistent with utility derived socially. This contrasts with other analyses where the process of care-giving is modelled as burdensome (Bobanic et al., Reference Bobinac, Job, van, Rutten and Werner2010).

In effect, two interpretations are offered: care as an argument in the physician utility function and care as a component in a separate ‘collective’ or ‘social welfare function’. Mooney and Wiseman associate the latter with commitment, and say it thereby goes beyond the standard case, which they believe can only accommodate ‘sympathy’ in the form of interdependent functions (Mooney, Reference Mooney2009: 22). Mooney also argues that health economists have ‘ignored’ the contextual element of caring: caring that takes place in a community, which bears some resonance with a process perspective, and thereby identifies a potential distinction between processes and outcomes of care. We agree with Mooney's assessment and commend his emphasis. At the same time, however, we feel the dual utility function approach does not capture all the relevant dimensions of care, as we argue below. Moreover, in the dual utility case it remains theoretically plausible for a meta-function to serve as the appropriate maximand with all the resulting trade-offs (see also Brennan’s, Reference Brennan1993 critique). Further, the dual function approach does not depart from the standard frame in its exclusive reliance on consequentialist reasoning. Care is a form of other-regarding behaviour, but limitedly so. As the quotation from McGuire above makes explicit, arguments (or utility functions) in the individual's (meta) utility function may be traded off against one another.

Care as social capital

The health economics literature consistently refers to health care as a ‘special commodity’ (see for example, Becker, Reference Becker2007; Culyer and Newhouse, Reference Culyer, Newhouse, Culyer and Newhouse2000; Jones et al., Reference Jones, Rice, Contoyannis and Jones2006, and McGuire et al., Reference McGuire, Henderson and Mooney1982, see also, Mooney’s, Reference Mooney2009, advocacy of a ‘community’ health care commodity). Indeed, in echoes of Culyer's (Reference Culyer1976) ‘humanitarian spillovers’ idea, Tuohy and Glied (Reference Tuohy, Glied, Glied and Smith2011) discuss the distribution of health care in terms of its status as a merit good. Governments then intervene to ensure certain levels of minimum health care for those on lower incomes by constraining the health care choices of those on higher incomes through the redistribution of resources. Tuohy and Glied's analysis and others like it do not explicitly attempt to analyse conceptions of care; instead they are focused on the distribution of (medical) care resources (see also Williams’, Reference Williams2012, study of geographical variations in medical care utilization).

Treating care as an externality makes it as an aspect of social capital, since most contributions to this approach define social capital in terms of non-rivalrous public goods (Folland, Reference Folland2006) or overtly as an externality (Portes, Reference Portes1998). The social capital literature supposes that caring social relationships have powerful beneficial effects on health. A considerable literature also recognizes the potential for the indirect effects of social capital on health through its associations with politico-economic and environmental variables that influence health, such as access to resources (for example, Dahl and Malmberg-Heimonen, Reference Dahl and Malmberg-Heimonen2010).

Although they do not refer explicitly to social capital, Bobinac et al.’s (Reference Bobinac, Job, van, Rutten and Werner2010) model of informal care is informative in that it seeks to expand health economics beyond its principal focus on the physician-as-carer by using the standard externality account. Those close to an ill individual benefit from improvements to this individual's health status by virtue of their relationship to that individual – ‘the significant other(s)’. In contrast, Bobinac et al.’s ‘caregiver effect’ refers to the welfare effects of providing care as a set of activities. Their hypothesis reflects the standard assumption of the disutility of work in that the greater the caregiving activities involved the greater the adverse impact of the care provider's welfare, through for instance, emotional stress and opportunity costs in time commitment. The authors claim the two effects are difficult to disentangle as informal care tends to be provided by a common unit – the family.

Bobinac et al. also argue that the family effect of informal care offers an important insight: care provision has welfare effects beyond the care provider–recipient relationship, and therefore that their research ultimately shows: ‘patients should not be seen as isolated individuals’ (Bobinac et al., Reference Bobinac, Job, van, Rutten and Werner2010: 555). Again, such claims have some correspondence to Culyer's ‘humanitarian spillovers’, and they also represent the social dimension of care in interdependent utility function terms. We agree with Bobinac et al.’s advocacy: the patient should not be viewed as an isolated individual. Indeed, their allusion to the social hints at relations nested in further relations, and thus departs from the standard view of social capital as a ‘plausible’ economic concept when it is reformulated at the individual level (Folland, Reference Folland2006). Nonetheless, Bobanic, et al. and Folland share their analytical grounding in Becker's individualist model of altruism within the family.

Care as altruism

The orthodox economic analysis of altruism has been heavily criticised, and by extension much of this criticism may be applied to health economics’ view of care as altruism. In Khalil's (Reference Khalil2003) egocentric account of altruism, altruists favour health care on account of its expected benefits to them. However, Khalil argues, altruism thus understood has a paradoxical quality in that it can be equivalent to masochism as the altruist gains more the more wretched the state of the parties receiving care. This is logically consistent with Culyer's idea of ‘humanitarian spillover’ and the interdependence of clinician-patient utility functions. Nonetheless, it seems hardly humanitarian or caring in a positive sense that others’ misery sustains a vicarious sort of utility for the altruistic/caring agent. More specifically, in extremis, under this conception medical cures are the last thing physicians want: therapeutic treatments are ruled out, and medical procedures are reduced to some form of intermittent palliative episodes of ‘care’ followed by periods of tolerating the patient's worsening medical condition (a source of disutility), to be followed by the patient's improvement as a consequence of further episodes of palliative ‘care’ (a source of utility), and so on. Thus, a process of infinite regress is entered as Khalil's masochist mutates into a sadomasochist with a stethoscope. Obviously Arrow's ‘medical ethics’ constrains a physician's toleration of a patient's deteriorating condition, but it is only a constraint: the logic of the mainstream conception of vicarious utility implies the physician is motivated to behave in the manner outlined above.

Moreover, the idea of a ‘caring externality’ is problematic in itself in that the spillover is not on a third-party, as in the standard case, but actually on one of the parties to the exchange, namely, the patient. If we say then, that care providers gain utility when their patients gain utility, does their care for those individuals constitute something over and above the health care services being delivered? Caring externalities might then be thought to counterbalance the negative effects of moral hazard in that the incentives that asymmetric information creates for physicians to pursue individual gain, such as supplier-induced demand, or the sadomasochist case, above, could be offset by the special regard for the well-being of the patient that comes from this caring preference.

However, this will not work, since on the standard view caring defined as an externality still acts as a secondary, non-essential type of factor since health care is delivered whether or not physicians exhibit caring preferences. In contrast, physicians are clearly incentivized to induce patient demand. Thus there should generally be a tendency for caring attitudes, such as compassion, and hence related activities, such as therapeutic care, to be crowded out, or at least become peripheral. Ethical motives, as it were, are always at risk when economic ones dominate. This conclusion is reinforced by the standard view in microeconomics that externalities are only addressed through the intervention of government in the market system. In the case of clinician caring preferences, this implies that unless government or some other institutional authority (medical associations, codes of ethics, etc.) weighs in on the side of patients, one should expect caring preferences to be continually under pressure to be crowded out of medical care markets.

The view of care as altruism is subject to a further problem: how does the altruist know what another agent's preferences are, especially assuming the standard view that preferences are understood to be revealed preferences? Drawing from Bourdieu (Reference Bourdieu1977), Dewey (Reference Dewey1922) and Veblen (Reference Veblen1994 [1899]), altruism, we suggest, is more appropriately seen as a disposition, in effect a habit, in that it is more suitably conceptualised as a cause of an individual's actions and practices than a consequence. Habit, moreover, can be either dynamically reinforced or weakened through changes in institutional patterns of social interaction.

As Folbre and Goodin (Reference Folbre and Goodin2004, 19) argue, the standard account gives:

This contrasts markedly to Arrow's (Reference Arrow1972) scarcity of altruism analysis. If the supply of altruism is limited, as Arrow argues, then there is a case for economizing on it. However, as Singer (Reference Singer1973) argued in his critique of Arrow, such a position conflates altruism and other similar virtues and dispositions to those properties associated with conventional economic goods. This, as Sandel (Reference Sandel2012, 128) maintains,

What we take from this is that treating ‘care’ as a preference in an agent's utility function makes the level of care something determined through an algorithm of utility maximisation. But this framing makes care a highly mechanical affair, and imposes an instrumentalist interpretation on the concept of care. This in turn suggests the body might be understood as a machine, and that ethical values, particularly non-consequentialist deontological ones, have no place in the analysis of care. Following the philosopher Bernard Williams (Reference Williams1985), we say that instrumentalism removes care from the realm of ethical consideration. For Williams, the outcomes of actions certainly warrant classification as ethical considerations, but so to do obligations and duties, as do character dispositions (virtues), given that they affect how individuals deliberate in undertaking or avoiding actions of certain types. Williams’ arguments stress deontological value and the social embeddedness of the individual. In contrast, a health economics based on socially dis-embedded individuals imparts, at best, a thin notion of care. Yet it is well recognised in various literatures that therapeutic care, for example, can provide beneficial health outcomes (Adams and Nelson, Reference Adams and Nelson2009; Armstrong et al., Reference Armstrong, Armstrong and Scott-Dixon2008; Kontos and Naglie, Reference Kontos and Nalie2009; Watson and Smith, Reference Watson and Smith2002).

Care as social capital

As noted, health economists consider health care a ‘special’ type of commodity – one exhibiting external benefits. We believe there are sufficient grounds for saying that health economists do not appreciate the implications of this narrative. Commodities are things that are tradable, therefore monetizable, produced for sale in a market, and therefore where property rights may be identified and transferred (Fine, Reference Fine2001; McMaster, Reference McMaster2013). Culyer (Reference Culyer, Baldwin, Godfer and Propper1990) is one of the few health economists to seriously reflect on this conceptualisation of health care. He describes commodities as:

This is appealing, but Culyer goes on to compare the ‘characteristics’ of health care; timeliness, clinical efficacy, and readmission rates, with those of a steak; juiciness, tenderness, and flavour. The implication is that care is solely instrumental in producing its special characteristics, and that the social relations and institutional arrangements involved in health care provision are essentially the same as those governing commodity exchanges, including those involving principal-agent relations. In the extreme, this would seem to suggest that the nature of dependency, as typified by the principal-agent relationship, is reducible to one of the distribution of information between two parties. Regrettably, Culyer's allusion betrays a reductionism that cannot acknowledge let alone analyse the relational qualities of care.

By contrast, much of the care literature emphasises that care and caring are interactive, and hence possess profoundly relational characteristics (Engster, Reference Engster2005; Mol, Reference Mol2006; Tronto, Reference Tronto2013). In an economics context Himmelweit (Reference Himmelweit2007: 585) expresses an important implication of the relational dimension of care in a compelling manner:

In articulating this, Himmelweit draws on Baumol and Bowen's (Reference Baumol and Bowen1965) analysis of the economics of the arts. They argue that productivity rises more quickly in areas of the economy which benefit directly from innovation, investment and/or technological enhancements. However, the arts, according to Baumol and Bowen, cannot benefit from these effects. They use the example of a string quartet – reducing the number of players or demanding that the musicians play more quickly, inter alia, may raise productivity in terms of the number of pieces performed per musician over a given period of time, but would substantially alter the nature of the piece. The problem arises from the attempted objectification of the arts. The same can be said about the nature of care.

In the context of nursing and gender, Adams and Nelson (Reference Adams and Nelson2009) similarly note that many desirable activities focused on therapeutic care are necessarily time costly and attempts to impose codified, standardised procedures are likely to diminish important aspects of caring, such as in the development of relations that potentially elicit more effective diagnoses (see also, Armstrong et al., Reference Armstrong, Armstrong and Scott-Dixon2008; Groopman, Reference Groopman2007; van Staveren, Reference van Staveren2001). This is consistent with Kontis and Naglie's (Reference Kontos and Nalie2009) argument that caring does not develop through theoretical learning, and involves experience (tacit knowledge) and ‘embodied selfhood’, i.e. a synthesis of primordial and social being. In short, medical, and other forms of care cannot be captured adequately by appealing to the characteristics of a steak.