The amplifying role of need in giving decisions

Abstract

Hamilton's rule predicts that altruism should depend on costs incurred and benefits provided, but these depend on the relative needs of the donor and recipient. Rewriting Hamilton's rule to account for relative need suggests an amplifying effect of need on relatedness, but not necessarily other relationship qualities. In a reanalysis of three studies of social discounting and a new replication, we find that relative need amplifies the effects of relatedness on giving in two samples of U.S. adults recruited online, but not U.S. undergraduates or Indian adults recruited online. Among U.S. online participants, the effect of genetic kinship was greater when the partner was perceived to be in higher need than when in lower need. In the other samples, relatedness and greater partner need were associated with greater giving, but the effect of relatedness on giving was not significantly amplified by need. Need never amplified the effect of social closeness on giving, although it did diminish the effect of closeness in U.S. undergraduates, likely reflecting a ceiling effect. These results confirm predictions from a modification of Hamilton's rule in a sample of U.S. adults, but raise important questions about why they hold in some samples but not others. They also illustrate the importance of understanding how contextual factors, such as relative need, can moderate the importance of common variables used in evolutionary cost-benefit analyses.

Introduction

Evolutionary theories of individual sacrifice often predict that decisions should be sensitive to the ratio of fitness gains to costs. As an example, Hamilton's theory of inclusive fitness predicts that individuals should sacrifice for a genetically related partner as long as the fitness cost of sacrificing (c_d) relative to the partner's benefit (b_p) is less than the donor's genetic relatedness to the partner (r) (Hamilton, 1964). Similarly, theories of reciprocal altruism predict that an individual should sacrifice for a partner as long as the partner's benefit b_p discounted by the probability that the partner returns the benefit is greater than the cost, c_d (Trivers, 1971).

Complicating these simple calculations is the fact that observable costs and benefits are usually only proximate currencies (e.g. food, money) that may have different fitness consequences for the donor and recipient at any one point in time. For example, the fitness benefit of a blood meal for a starving vampire bat is much greater than the fitness cost incurred by a well-fed partner giving up that same blood meal (Wilkinson, 1984, Wilkinson, 1988). For this reason, we should expect individuals to be attuned not only to proximate benefits and costs of a decision (i.e. the amount of food or money), but also to the relative fitness gains from the same quantity of a currency for the donor and recipient.

 One way that humans conceive of these relative gains is in terms of a recipient's need and how it compares to the donor's need (Aktipis, Cronk, & de Aguiar, 2011; Gurven, 2004; Hruschka, 2010). Experimental studies have shown that a range of proxies for a partner's need at varying times scales (e.g. being sick, being poor, a recent loss) increase both self-reported likelihood of helping the partner (Burnstein, Crandall, & Kitayama, 1994; Hackman, Danvers, & Hruschka, 2015; Korchmaros & Kenny, 2006) and transfers of actual money (Howe, Murphy, Gerkey, & West, 2016). Moreover, observational studies of Tsimane forager-horticulturists in Bolivia show that families with objectively greater need—measured as daily consumption relative to production of calories over a year—are more likely to receive food from other households and are less likely to share food outside the household (Hooper, Gurven, Winking, & Kaplan, 2015).

While most studies have examined main effects of need, formal evolutionary models of individual sacrifice suggest that relative need should also modify the effect of other factors, such as relatedness, probability of a return, and the ratio of proximate benefits to costs. Here, we define need in terms of the fitness change from a unit change in a proximate currency (e.g. food, money). For example, suppose we decompose the cost incurred by a donor as c_d = cf_d where c is the proximate cost of sacrifice and f_d is the donor's fitness loss for a loss of one unit of the relevant currency. Similarly, the benefit to recipients can be rewritten as b_p = bf_p where b is the proximate benefit and f_p is the fitness gain from the receipt of one unit of the relevant currency. With these substitutions, Hamilton's rule becomes the following:

$$\frac{b}{c}\frac{f_p}{f_d}r>1$$

Thus, the decision to sacrifice should depend on the product of three quantities: the ratio of benefits to costs in a proximate currency (e.g. food, money) $\left(\frac{b}{c}\right)$, the need of the partner relative to one's own need $\left(\frac{f_p}{f_d}\right)$, and one's relatedness to the partner (r). Notably, when a partner is in greater relative need, the same increase in r should lead to a larger increase in the amount one is willing to sacrifice (c) to give the same proximal benefit (b) for a partner. In short, a partner's greater relative need should amplify the effect of genetic relatedness on willingness to give. In this generalized form, b and c can represent different types or amounts of resources, but in specific cases such as the direct transfer of resources (e.g. a donation of money from one person to another), b and c could also be equivalent. In those cases, what may matter most is the relative need of each partner.

Models of repeated beneficial interactions among unrelated individuals provide different predictions about the moderating effect of need. In this case, genetic relatedness in the equation, r, is replaced with the probability of the partner returning the benefit in the future, p. The value p depends on a number of factors, including the expected duration of the relationship, the degree of discounting of the future benefit, and one's trust that the partner will return the benefit in the future when needed (Fehr & Fischbacher, 2003). The current fitness gains to the partner (f_p) are replaced with the expected future fitness gains for the donor (f_d′) from a one unit change in the currency.

$$\frac{b}{c}\frac{f_{d\prime }}{f_d}p>1$$

In this case, as one's future need relative to one's present need increases $\left(\frac{f_{d\text{'}}}{f_d}\right)$, one should be willing to sacrifice more (c) all else being equal. Crucially, in this case, there should be no amplifying effect of the current need of the recipient relative to that of the donor.

The decomposition of costs and benefits illustrates how assessments of relative need should play a role in decisions to help others, and how relevant indicators of need (e.g., self vs. partner's need, self's present vs. future need) matter differently depending on key qualities of the relationship (e.g., genetic relatedness, probability of a future return). Although numerous studies have examined the additive effects of relatedness, need, and proxies for mutually beneficial reciprocal relationships (Delton & Robertson, 2016; Hackman et al., 2015; Howe et al., 2016; Jones & Rachlin, 2006; Korchmaros & Kenny, 2006; Lieberman & Lobel, 2012), far fewer have examined how relative need might amplify the effect of genetic relatedness on generosity. One recent observational study suggests that need—operationalized in terms of a household's caloric consumption minus production—amplifies the effect of genetic relatedness on food sharing in a small-scale society (Hooper et al., 2015). In that paper, the researchers used objective proxies of need, measured in caloric surpluses or deficits. However, perceptions of important social categories such as genetic relatedness or need can influence behavior even when they do not objectively track features of the world. For example, perceived kinship can influence altruism and sexual attraction even when individuals are not genetically related (Lieberman & Lobel, 2012; Lieberman, Tooby, & Cosmides, 2007). It is still unclear how perceptions of a recipient's versus one's own need also amplify the effect of genetic relatedness on giving.

In this study, we assess whether perceptions of relative need amplify the effects of genetic relatedness and a proxy for expectations of future return (e.g. social closeness to a partner) according to the expectations described above. Specifically, a donor's need relative to the recipient's should amplify the effects of genetic relatedness. However, relative need should not amplify the effects of social closeness, as reciprocal giving involves a trade-off between the donor's current and expected future need, not the recipient's need. To assess these expectations, we re-analyzed three existing data sets from three distinct samples (U.S. undergraduates, U.S. MTurk users, and Indian MTurk users) using multi-level modeling with interactions among the key variables. We also collected and analyzed new data for a direct replication of the study in a U.S. Mturk sample.