The ultimatum game was run in fifteen small-scale societies, from Amazonian foragers to whale hunters. Offers ranged from stingy to hyper-generous, and some groups even rejected offers that were too high. There is no single human bargaining instinct.
Culture, context, and design
The fairness this unit measured is not a universal constant. Cross-cultural experiments show offers and rejections vary widely with how a society is organized, and undergraduate subjects turn out to be outliers. A mechanism that assumes one fairness norm can fail badly on players who hold another.
In the ultimatum game, a fairness-minded responder rejects offers that are...
A mechanism is strategy-proof if honesty is a...
Predict, then play
Switch the responder’s cultural norm and see how the acceptable offer changes. The same game, different people, different thresholds.
Why did that happen?
Take a guess, then reveal the reason
Ultimatum behaviour varies across populations, and observed offers and rejections are associated with market integration and local forms of cooperation. These comparisons do not by themselves prove that markets caused particular fairness norms. The safe conclusion is that a designer should not assume one laboratory population represents everyone.
The math
Most experimental results come from WEIRD populations (Western, Educated, Industrialized, Rich, Democratic). Henrich and colleagues ran the ultimatum game in fifteen small-scale societies: mean offers ranged from about 26 percent (the Machiguenga) to 58 percent (the Lamalera whale hunters), and rejection behaviour was associated with market integration and cooperation in daily life. The study establishes substantial cross-population variation, not a single causal channel.
The interactive’s presets are illustrations of how different inequity-aversion parameters change a threshold. They are not estimates for the societies named by the buttons. And the historical problems in medical matching and Boston school choice arose from concrete incentives and market rules, not simply from assuming the wrong fairness parameter.
Work it out
One game, three thresholds
Take the same $10 ultimatum and vary the responder’s illustrative envy weight . At , the threshold is about $2.90. At , it is $2.50. At , it is about $1.15. These calculations show what the parameter does; they do not estimate any society’s preferences.
A designer optimizing offers, or building a market that relies on fair play, gets very different answers depending on which holds. Assume the WEIRD value everywhere and you will over- or under-shoot elsewhere. The lesson: measure the population’s actual norms, and design the mechanism for them, not for an idealized universal player.
Key idea
Fairness is learned from institutions and varies across cultures, and the usual lab subjects are outliers. Effective mechanism design measures the norms of the actual players rather than assuming a universal one.
Check yourself
Sources
- Henrich, J., et al. (2001). In search of Homo economicus: behavioral experiments in 15 small-scale societies. American Economic Review, 91(2), 73-78.