SUPPOSE I hand you a bundle of cash equivalent to a week’s salary and say you
can have it, but on the condition that you share the money with someone else.
I’m not saying who, but it’s someone you know. On the plus side, your anonymity
is guaranteed, and you can offer as much or as little as you like. If the other
person accepts your offer, each of you keeps your share of the kitty. But if
your offer is refused, neither of you gets anything. How much would you
offer?
From a strictly rational point of view, you should offer as little as
possible, and the other person should accept it—after all, free money is
free money. But people hardly ever act that way, and the chances are that you
made a generous offer. In fact, when economists try out this “ultimatum game” on
social science’s favourite lab rats—university students—they find
that the most common offer is 50 per cent of the money, and the average is
around 45 per cent. Even when the recipient doesn’t have the option of rejecting
the deal, the second most common offer (after zero) is 50 per cent. Offering
less is the rational approach, yet the subjects routinely give away more.
All around the world, from Pittsburgh, Jerusalem and Tokyo to the Slovenian
town of Ljubljana, and Yogyakarta in Java, the students behave the same way. So
could this uniform willingness to share be a part of our evolutionary heritage,
some innate sense of fair play?
Advertisement
To find out, a group of economists and anthropologists took the ultimatum
game on the road, to some of the most remote, traditional societies on Earth.
Their results, most of which are not yet published, reveal that people appear
not to share a common sense of fairness. Instead, what people from
industrialised societies consider fair is just one of a broad spectrum of
perspectives. “The way people play these games relates consistently to the way
they live their lives,” says Herbert Gintis, an economist from the University of
Massachusetts at Amherst. To put it another way, our social environment shapes
our sense of morality. As the anthropologists look deeper, they are beginning to
understand what makes one society generous and another selfish. And that, in
turn, raises the possibility that we can shape our societies to favour fair
play.
The collaboration between the anthropologists and economists began in 1996
when Joe Henrich, a young anthropology researcher then at the University of
California at Los Angeles, tried the ultimatum game on members of a traditional
society for the first time. Henrich had been studying the Machiguenga, a tribe
which lives in a remote part of the Amazon rainforest in Peru and gets by on a
combination of subsistence farming, hunting and gathering. Machi-guenga society
is very different from our own, so he thought it would be interesting to play
the ultimatum game with these people. “They are one of the least social human
groups known,” says Henrich, who is now at the University of Michigan.
He was in for a surprise. The Machiguenga’s idea of what was a fair offer
proved very different to those of university students. They routinely offered as
little as 15 per cent of the kitty and their average offer was little more than
a quarter. Even more surprisingly, their counterparts did not punish this
apparently mean behaviour: there was only one rejection in the 21 games played.
“There is a delicious irony about the Machiguenga results,” says Colin Camerer,
an economist from the California Institute of Technology, who coached Henrich in
game theory. “The idea that very skilled college students around the world
violate the theory [of rational behaviour] but the illiterate Machiguenga ‘play
correctly’ is just too funny.”
Determined to get to the bottom of the mystery, Henrich and Rob Boyd, his
professor at UCLA, assembled a group of 12 field anthropologists and a handful
of economists including Camerer, Gintis and his Massachusetts colleague Samuel
Bowles, and Ernst Fehr from Zurich University. The group got together in 1997 to
plan a wider study. Then the anthropologists set off on their research.
Last December, they met to discuss their findings. It turned out that in the
15 traditional societies where they tested the ultimatum game, the commonest
offer ranged from 15 per cent to 58 per cent. The meanest—most rational in
economic terms—included the Machiguenga and some Tsimané
subsistence farmers from the Bolivian Amazon, and Hadza foragers from Tanzania.
The most generous were the Lamalera whalers from Indonesia whose hyper-fair
offers frequently exceeded 50 per cent.
People from some societies consistently offered the same share of the
kitty—just like the university students—but in other societies
offers varied widely. And there was also a huge spread in people’s tendency to
reject offers. Au and Gnau speakers from Papua New Guinea often rejected
hyper-fair offers, while members of some societies never rejected even very low
offers. People in industrialised societies were among the groups most inclined
to punish.
What could account for these differences? Age, sex, level of education and
participation in paid labour turned out to have no bearing on how people played
the game. Nor were poorer people more reluctant to share the money because they
valued it more highly. Most of the anthropologists used a stake equivalent to
one day’s wage, but even those that used stakes up to ten times higher found it
had very little effect on behaviour. And for each additional day’s pay at stake,
the average offer increased by only 0.7 per cent. What’s more, when Henrich
offered the equivalent of a week’s teaching pay to a group of graduate
anthropology students at UCLA, he found that this high stake had little effect
on the fraction they offered to give away. Even offering ten weeks’ salary, as
Fehr had done in previous research, had no effect.
The results seem to show that society, not the individual, is what counts.
“The most interesting result we have is that the group matters and individual
preferences matter less,” says Boyd. You can see this most clearly where some
groups of people living side by side play the ultimatum game very differently.
In Conambo, in the Ecuadorian Amazon, for example, the Achuar and Quechua have
distinct histories but these days live together and even intermarry. Yet John
Patton from Washington State University found that the Quechua offer a mean 26
per cent on average, compared with their neighbours’ 43 per cent.
In Bolivia, Mike Gurven from the University of New Mexico played the
ultimatum game with five separate communities of Tsimané from villages
scattered along the same river bank. Although village membership is fluid and
people spend lots of time visiting relatives in other villages, the average
offer varied widely between the communities, ranging from 27 per cent to 46 per
cent. The best predictor of how much players would offer was their home village,
rather than their status, wealth, or other personal circumstances.
What makes some societies generous and others mean? One obvious answer might
be biology. The late William Hamilton was the first to recognise that apparently
altruistic behaviour can often be explained as kin selection—it makes
sense to be generous toward your relations because they carry many of the same
genes as you. If our sense of fairness were a reflection of biological ties, you
would expect individuals from small, interrelated societies to be more generous
than those from large, genetically diverse ones. But family-based societies such
as the Machiguenga and Tsimané are the least generous. What’s more, in
all but one study, group size did not influence behaviour.
In the one society where generosity did depend on group size, the results
were the opposite to what you would expect. Among the nomadic Hadza of Tanzania,
one of the world’s few remaining hunter-gatherer societies, members of four
small camps offered a meagre 27 per cent on average, while members of a large
camp offered 40 per cent. “People who are forced to share all the time like to
escape it when they can,” explains Frank Marlowe of Harvard University, who did
the study. People from small camps have little opportunity to behave selfishly,
and in larger camps generosity is motivated more by fear of reprisal. “There is
both less reason to advertise generosity and fewer repercussions from being
labelled stingy in small camps than large, so Hadza in smaller camps made lower
offers,” says Marlowe.
If group size isn’t the key, then perhaps our attitudes to generosity are
somehow linked with the complexity of the society to which we belong. At first
glance this approach is promising. People from sophisticated industrial
societies make generous offers, while the meanest offers come from the
Machiguenga, who are among the least complex societies known. But in his study
of five Tsimané villages, Gurven found that these people, who have a
simple, family-based society like the Machiguenga, show a range of responses to
the ultimatum game. In fact, close scrutiny reveals no link between the social
complexity of a society and the way its members play the game. The only trend
that emerges is that groups of people from the most basic societies vary most in
their attitude to fairness. Conversely, students from industrial societies show
remarkable homogeneity even when separated by thousands of miles.
Nonetheless, we can still begin to explain a group’s behaviour in the
ultimatum game in terms of more specific aspects of their lifestyle. For
example, the hyper-fair offers made by some Au and Gnau speakers in Papua New
Guinea make sense in the light of their culture of gift giving. But in these
societies, even unsolicited gifts come with strings attached. By accepting an
offer, recipients commit themselves to reciprocating whenever the giver calls in
the favour. In real life, rejections of large gifts are common, just as they are
in the ultimatum game.
The hyper-fair offers made by Lamalera whalers have a different root. Michael
Alvard from Texas A&M University, who performed this study, believes their
generosity stems from a life based on mutualism. Harpooning whales from open
boats is a dangerous business, and the hunters must cooperate to succeed.
Catches are shared out along traditional lines to everyone with a stake in the
catch, from hunters, boat owners and builders to other members of the clan that
owns the boat. The whole group benefits from working together. “In societies
where these sorts of coordination problems are common, people will be inculcated
to both trust more and punish more,” says Avard. He argues that this, together
with the benefits of establishing a good reputation, explains why almost one
third of players from Lamalera offered more than half the stake.
At the other end of the spectrum, the Machiguenga’s idea of what’s fair may
also be a product of their way of life. In this society there is very little
cooperation, exchange or sharing beyond the family unit. This insularity is
reflected in the fact that until recently they didn’t even have personal names,
referring to each other simply as “father”, “patrilineal same-sex cousin” or
whatever.
So the most cooperative society is also the most generous, while the least
cooperative is the meanest. Could there be a pattern emerging? After all, a
society’s level of cooperation is central to the way it runs its economy. The
researchers suspected that this and other economic variables might be the
crucial factors that determined people’s behaviour in the ultimatum game. To
test their ideas they ranked the 15 traditional societies, plus Henrich’s UCLA
students, according to each of four different qualities: pay-offs to
cooperation, market integration, anonymity and privacy.
“Pay-offs to cooperation” is a measure of the importance of cooperation in
the way the society makes a living. “Market integration” describes the extent to
which it relies on trading in daily life. High-tech Western society, for
example, would collapse if trading disappeared, whereas life for the Hadza would
change very little. The “anonymity” category considers who you normally interact
with. For example, the Zimbabwean farmers studied by Abigail Barr from Oxford
University often trade with people they don’t know and may never see again,
whereas such interactions are extremely rare among the Achuar of Ecuador.
“Privacy” means people’s ability to hoard goods and keep secrets. The Hadza, Au
and Gnau lack even pockets in which to hide the smallest item.
The researchers found no correlation between generosity and either anonymity
or privacy. But there were links with cooperation and market integration. These
two factors turned out to predict two-thirds—68 per cent—of the
variability in average offers made by the 16 societies.
So we can start to predict how different societies might play the ultimatum
game. We already have some idea of how cultural factors can influence our
attitudes to what’s fair, but how does the degree of cooperation and market
integration in a society influence the way we act?
The most obvious explanation is that different societies have different core
values, which their members absorb through their interactions with one another.
These values influence every moral decision we make. But surely this can’t be
the whole story. The findings from the Hadza and other societies suggest that
people’s tendency to cooperate depends on context—size of camp, for
example. This supports an alternative idea: that our social interactions allow
us to build up mental models of the appropriate behaviour in certain contexts.
In this case, rather than a person being guided by a predisposition to certain
values, each new situation simply cues a set of behavioural rules, or
“heuristics”. These heuristics serve as a short cut to behavioural
decision-making, a gut feeling that points to a response without need for
detailed analysis of the particular situation (New Scientist, 4
September 1999, p 32).
This could be the key to why different societies play the ultimatum game in
such a variety of ways. Industrialised societies, the researchers hypothesise,
have developed a standard set of heuristics for dealing with anonymous people,
hence the similarities in how students around the world played the game.
Traditional societies, where such dealings are rare, have no equivalent
rules—though some, such as the Lamalera, developed a different set of
fairness rules that lead to even greater sharing. But in other societies, such
as the Tsimané, the game may weakly activate two or more different sets
of rules, or fail to activate any set at all, so players’ responses vary
widely.
It might all sound rather academic, but the work has important practical
implications. Bowles points out that if our moral values are shaped by our
lifestyle rather than by human nature—as traditional economic theory
holds—then we should be able to promote good social behaviour in the
correct social context.
Barr agrees. In particular, she believes that understanding what motivates
fair play could help in the fight against corruption in developing economies.
She cites a study done in 1998 which showed that 70 per cent of Ugandan
government money allocated to education never reached the schools for which it
was intended. The simple act of posting budgetary details in schools cut this
loss to less than 20 per cent. In a society where reputation is important, the
prospect of being caught out shamed local administrators into doing the right
thing. “To my knowledge there was never any use or threats of court action,”
says Barr.
In fact, mainstream economists are increasingly interested in adding human
behaviour to their calculations. Where once the economic models assumed that we
all act rationally to maximise our returns from any transaction, they are now
beginning to reflect the complexity of the real world. “The fact that a
substantial fraction of people is motivated by concerns for fairness and
reciprocity is very important for many areas of economics,” says Fehr. Gintis
adds: “It’s not wrong to say that people are rational, but rather to think that
it’s rational to be a selfish, sociopathic brute!”
