experiments with indians and economic theory
A while back, Jeremy turned me on to the work of John Henrich, a psychologist at UBC. Henrich is one of those people who travels to far off places and has people play games, like the Prisoner’s Dilemma or the Ultimatum Game. One of the things you learn is that there is substantial variation in how people play these games outside of Western cultures. There’s a nice article in the National Post about his research and his hypothesis that we are WEIRD (Western, Educated, Industrialized, Rich, Democratic). In other words, comfortable Westerners really view things differently than non-industrialized people.
Take the Ultimatum game, where person #1 can split $100 and offer it to person #2. If person #2 rejects, no one gets anything. If person #1 accepts, then person #2 keeps what is offered and #1 gets the rest. Economic theory suggests that something is better than nothing, so people will accept what is offered, even if it is just $1. What happens in Western societies is that player #1 offers around $48, while player #2 rejects if the offer is under $40. When Henrich’s team discovered is that there are cultures were people will accept the low ball offers because, as economic theory suggests, something is better than nothing – it’s free money! Westerners reject low offers, maybe because they feel insulted, or they want to punish player #1 for making a low offer.
If you follow rational choice sociology, this should not be surprising. The work of Werner Raub and Vincent Buskens is all about “embeddedness” of games – the social context of the game affects how people play it. Even in Western culture, researchers have found real differences in game play. I learned from Werner, when he visited Chicago a while back, that students and business people play the trust game differently. His argument was that business people show more trust because they work in a world of cooperation, where students live in a world of short term relationships.
The lessons I take from this research:
- For game theory: Rather than going on insane mathematical excursions, maybe game theorists should work on simple games and variation in play. How do people interpret the game? How does heterogeneity of players affect the outcomes? I think this is more interesting than the umpteenth behavioral paper about risk that’s based on college students.
- For economic theory: Let me phrase this in econo-talk so that you guys can understand… People have different utility functions for money and they have complex utility functions. If you believe the research, there are cultures where people have nice linear utility functions for money. In the West, we may decide that we don’t value money and that we value fairness, or that we value punishing other players. Showing real differences in utility functions across cultures is pretty friggin’ important.
- For economic sociology: What sorts of cultures produce “homo economicus?” What needs to happen to create a society of cooperators? Are the behaviors in these games observed in real economic transactions? Henrich’s writings cite anthropological research on these cultures, but more can be done, especially in Western cultures and focusing on transactions. For example, maybe someone like Sudhir Venkatesh could have people in inner city neighborhoods or low income schools play these games. Then we could see later if game play is correlated with certain job search behaviors or life course outcomes.
Good stuff.
Fabio,
I’m tremendously interested in economic anthropology – clearly we need to “learn” how to operate in a market system, and more insights along these lines would help, for instance, communist transition countries. That said, I’m curious what you think of the obvious rejoinder to these types of psychological studies of game theory: they don’t seem to match real world experience. For example, there is a version of Ultimatum where the guy offering gets his share regardless, and the guy getting the offer can accept or reject, anonymously. In the lab, you still see people offering something other than $0. However, we simply don’t see people on the street going up to strangers with $100 bills and offering to split that money. This says to me (and, based on seminar reactions I’ve seen, the same reaction applies to many, many theoretical game theorists) that a lot of experimental methodology is a massively flawed way of learning how people will play games. Further, from the point of view of theory, what cultural differences exist concerning altruism and fairness are better dealt with by suitably changing the *payoffs*, rather than by changing the equilibrium concept or the idea of rational play, where “rational” is just interpreted to mean “acting to maximize some goal.” Indeed, the most standard forms of game theory can deal with seemingly symmetric games where players in fact have different payoffs (i.e., some players are altruistic and some are not).
afinetheorem
August 26, 2010 at 12:17 am
There was a very good paper in Behavioral and Brain Sciences recently on WEIRD populations, along with open peer commentary.
Also, the work of Christina Bicchieri on social norms seems to follow some of your lessons (“The Grammar of Society”).
Xavier
August 26, 2010 at 12:52 am
Links to the BBS paper: http://orgtheory.wordpress.com/?s=henrich+et+al+bbs
And, I think this link will get you directly there: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.163.1068&rep=rep1&type=pdf
teppofelin
August 26, 2010 at 1:48 am
This sounds like a very promising avenue of research.
Mikaila
August 26, 2010 at 3:05 am
[...] Experimental anthroposociologiconomics August 26, 2010 by Michal Commenting on this: [...]
Experimental anthroposociologiconomics « Brokering the Closure
August 26, 2010 at 12:08 pm
I wanted to comment, but it got too long. So I wrote it here http://brokeringclosure.wordpress.com/2010/08/26/experimental-anthroposociologiconomics/
Michal
August 26, 2010 at 12:11 pm
Delia Baldassari wrote a post about “lab in the field experiments” during her guest blogger stint. She’s been running game experiments in Uganda. Her conclusion: This finding shows that behavioral games can be used not only to capture underlying behavioral tendencies common to all human beings (or profound cultural differences across societies) but also to measure differences between individuals (or groups) that derive from their individual and group experiences.
brayden king
August 26, 2010 at 1:33 pm
most of these experiments are with small amounts of money, which induces variation in cooperative responses based on the COL of the people comprising the sample. Indians, college students & businesspeople are not different because they have inherently different-shaped utility functions but because their utility functions have a different ‘origin’ (n graphical terms). To really make the claims the authors are making, teh experimenters need to be able to play with money that matters to their sample ($5 for college students, but $50 for business-people, etc). For paltry amounts, business people are nice not because they are more cooperative, but because they don’t really care.
sd
August 26, 2010 at 6:22 pm
A fine theorem: I’m usually an experimental skeptic, but I feel better about simple games that try to mimic some real world situation. I’ll write a longer post about this later.
Sd: I think you have a good point. But I wonder, are we setting the bar too high? Sure, I bet business people would act like students if the ultimatum game was about $1M rather than $100. But what would that prove? There is no maximal state – you can always find some better state. We can always find something that will induce different behavior if we try hard enough.
Also, small amounts are ok w/me. Transactions with small amounts of resources are much more common than those with large amount.
fabiorojas
August 26, 2010 at 6:35 pm
Hey, what’s going on here? Rampant relativism? Cultural determinism? Collapse of positivist methodology? Any moment someone will come up with something like ‘industry recipes’!
JC
August 26, 2010 at 8:55 pm
Take the Ultimatum game, where person #1 can split $100 and offer it to person #2. If person #2 rejects, no one gets anything. If person #1 accepts, then person #2 keeps what is offered and #1 gets the rest.
If you had a multiplayer version of this, it would be a metaphor for employment. If nobody accepts the employer’s offer, the business fails and nobody gets anything, but if enough workers accept, the acceptors get their wages and the employer gets the rest of the value of the business’s output.
Inter-worker competition would drive down the offers pretty strongly, unless the employer happened to have a strong sense of fairness.
chris
August 27, 2010 at 1:19 pm
Sd, many of the behavioral games reproduced in small-scale societies are played for amounts of money that exceed a day’s wage there, and might even be for as much as a month’s wage. (I forget, but this is in the Henrich piece, as well as in the more focused “In Search of Homo Economicus”). This doesn’t answer your point about business people, but at least the differences amongst small-scale societies cannot be attributed to floor effects (and these were every bit as large as differences between small-scale societies and WEIRD societies).
By the way, getting back to an earlier thread about poli sci’s usage of methods inspired by econ: it seems to me like political scientists have been especially good at using behavioral games for these “anthropological” means. This, for example, seems to be along the lines of research Fabio wants to see: http://www.russellsage.org/publications/books/090112.677920 .
p.s., I also found Henrich’s WEIRD piece to be very inspiring. I hope it leads to as much excellent research as it should.
Andrei Boutyline
August 28, 2010 at 6:28 am