Selfish Cultures? Cross-Cultural Valuations of Fairness

Neoclassical economic game theory assumes that players solely consider their own payoffs when making decisions; however, experimental research continually suggests that people do not behave this way in reality. In one of the simplest decision settings, the ultimatum game, players routinely deviate from the subgame perfect equilibrium prediction; in fact, players frequently reject positive offers even if the alternative is zero payoff (Camerer 2011). While deviations from maximizing a player’s own payoffs may appear to suggest that said player behaves irrationally, these deviations are routinely replicated. Consequently, research suggests that players factor in economic reasoning, but that self-interest is not the only factor at play. Further research in this field demonstrates that players are averse to unequal payoffs: in other words, players care about fairness. Furthermore, players have been shown to factor in the intentions that led to these payoffs as well (known as the concept of reciprocity). However, this research has been almost entirely conducted on university students who might only reflect behavior from culturally-evolved beliefs characteristic of urbanized, industrial societies. Cross-cultural studies indicate substantial differences between the results of experiments conducted in these societies and those conducted in others. In order to properly explore these differences in fairness across different cultures, I will first discuss the concept of fairness (both inequality aversion and reciprocity) by assessing academic studies on simple games. I will then Shen 2 analyze two cross-cultural studies to demonstrate that the desire for an equal outcome is universal in all cultures, but variations in the rejections and punishments between different groups of people depend on cultural and societal norms.

Variations of the ultimatum game demonstrate that both the first mover and the second mover value fairness. In a study conducted in Indonesia to investigate high stakes rejections, Cameron concludes that while increasing the stakes decreases second mover rejection rates, some second movers still rejected large sums of money if the split was not even (Cameron 1999). But confined to the ultimatum game, in cases where the first mover offers half the share, we cannot rule out the possibility that the first mover acts solely out of self-interest: the first mover may have offered to split not because they value fairness but because they expect a lower offer will be rejected. To test this logic, researchers introduced the dictator game. The only difference between the dictator game and the ultimatum game is that second movers cannot reject offers. We expect that a player who offered half of the share out of self interest in the ultimatum game would offer nothing in the dictator game, but the first mover often offers positive amounts even in a dictator game where the first mover has complete anonymity (Hoffman et al. 1994).

Inequality aversion alone cannot explain behavior in some variations of the ultimatum game; reciprocity -- where intentions of the first mover matter -- contributes to experimental outcomes as well. Falk et al. compares ultimatum games where the first mover is limited to offering either one payoff outcome or the other. In a game where the first mover could choose between offering a payoff (8, 2) or (5, 5), the (8, 2) offer was rejected around 45% of the time; meanwhile, in games where the first mover chose between a payoff of (8, 2) or (10, 0), the (8, 2) offer was rejected less than 10% of the time (Falk et al. 2003). Inequality aversion would predict Shen 3 that the (8, 2) offer would be equally rejected in both game variations because the share obtained by the second mover is the same in both scenarios. Thus this study demonstrates that not only do the outcomes of the first mover matter, but so do the intentions. In games where the first mover offers the better alternative to the second mover, even if the alternative is equally unequal, the second mover is almost five times as likely to accept the offer (Falk et al. 2003).

While the experiments discussed thus far demonstrate payoff outcomes are reliant on both fairness and reciprocity, most of the experiments discussed overwhelmingly use university students as their subjects and apply results gathered from Western, industrialized countries to broader human psychology. The results of these experiments could reflect universal human cognitive behavior, but they could also be the result of specific social environments. Some economic studies examine the variation between different social environments. One two-round economic and anthropological study found that, contrary to many previously noted experiments (where the mode offer was frequently 0.5 across the board), mode offers ranged from 26% to 58% of the ultimatum share (Henrich et. al 2005). Rejection and acceptance rates also varied depending on the group of people. The figure on the right is a bubble plot of the ultimatum game rejection rates of different groups: some groups rejected offers both below and above 50% at nearly equal rates (Henrich et. al 2005, Shen 4 2006). This deviation from the experimental norm in Western studies, where rejection rates do not increase as offers exceeded 50% of the share, suggests that deviations from expected equilibriums are due to cultural differences rather than a universal psychological tendency. We can apply the concept of reciprocity here: perhaps groups where people rejected offers above 50% have cultural norms where the intent behind these offers are read differently. However, it’s worth noting that in all societies studied, offering an equal share minimizes rejection probability, suggesting that a desire for fairness is either present in all cultures or is an innate human trait.

To further assess these similarities and differences among different societies of people, we can look at games involving punishment. One study conducted research using an adapted version of the public goods game -- a game modeling dilemmas where cooperation generates socially optimal outcomes but individual players have the incentive to slack off. Players in this adapted public goods game could contribute “tokens” to a group project where each token invested in the group project generated a return of 0.4 for each group member (Herrmann et al. 2008). Afterward, players in this game could choose to punish other members, at the expense of their own payoffs, based on others’ contributions to the group project. Experimental results showed a strong correlation between the strength of the group’s “rule of law” in their respective society and the severity of punishment (Herrmann et al. 2008). Players from places with a rule of law deemed relatively strong (where law enforcement institutions were generally seen as just) punished slackers much more harshly than those with weaker rules of law (Herrmann et al. 2008). These results reflect how social norms may seep into decision-making because socially unjust behavior is punished accordingly in societies with stronger rule. Although punishments Shen 5 were more severe for freeloaders among societies with strong rule of law, the punishments increased among all cultures the more contributions deviated from the mean.

In both cross-cultural studies, further deviations from the societally expected norm (an even split in the first study and the mean in the second) generated less payoff for different parties. These results suggest that there is some innate human psychology at play in rational decision-making that factors fairness into account; however, the results also suggest that fairness and reciprocity materialize differently based on societal norms. These societal norms are dictated by a number of factors, but both experiments note that the cultural differences are less dramatic between societies with similar economic and judicial systems. Furthermore, results are predictable in both experiments when factors such as levels of market exposure, strength of law enforcement institutions, and size of society; these findings signal that cross-cultural differences are not arbitrary (Henrich et. al 2006; Herrmann et al. 2008). While it may be a leap to say that the decision behavior of a group of people is due to their cultural experience, the idea that average individual experiences contribute to the evolution of cultures over time could help explain cultural differences in decision-making. All in all, while simple psychological behavior certainly plays a part in deviations from neoclassical game theory predictions, it is likely not the only contributor to these deviations.