Friday, December 19, 2014

Markets and Morals (2): The Roots of Human Sociality Project

There is nothing inherent in market exchange that make markets morally corrupting.  Actually, one might argue, markets depend on morals.  Voluntary exchange in markets requires trust and a sense of fairness.  Before you deal with strangers, you have to trust that they won't cheat you.  You have to trust that your property is secure.  You have to trust that social norms of fairness and the rule of law will enforce contracts, protect your property from confiscation, and keep banks sound.  You have to trust that the legal system will punish violence, fraud, and corruption.

The modern commercial society did not arise in the modern world until the development of the moral infrastructure of the bourgeois virtues.  When those bourgeois virtues are absent or weak, markets fail to work.

Our recent experience with the global financial crisis illustrates the failure of markets without trust and fairness.  Financial markets are built on trust in promises.  In primary financial markets, borrowers sell their promises to repay their debts to lenders.  In secondary financial markets, investors buy and sell these promises.  The problem is that once these promises are made tradable, the networks of trust are weakened, and it becomes hard to judge the trustworthiness of the promises.  Imprudent risk-taking, unscrupulous greed, and fraudulent deception can then lead to financial collapse.

If markets do depend on morals, then we should expect that societies with extensive market experience will show the moral norms of trust and fairness on which markets depend.  Experimenting with economic games is one way to test this.

Consider the research program that has come to be called "The Roots of Human Sociality Project."  Since 1997, about two dozen anthropologists and economists have been studying the evolution of prosocial norms by combining experimental economics and anthropological field ethnography in gathering evidence from 24 small scale societies around the world.  This research has been presented in a series of articles and two books (Henrich et al. 2004; Ensminger and Henrich 2014).

Much of neo-classical economics and classical game theory has been dominated by the Homo economicus model of human beings as rationally selfish maximizers of their utility.  But then in the 1980s and early 1990s, experimental economists discovered that the predictions of the Homo economicus model were not being fulfilled in the way people played the Ultimatum Game.  In this game, the experimenter provides some amount of money (say $10) for two players.  One player designated the proposer will propose a split of this money between the two players.  The other player designated the responder will respond by either accepting this proposed split or rejecting it.  If he accepts it, the money is split as proposed.  If he rejects it, neither player receives any of the money.  The prediction of the Homo economicus model is that the proposer will take most of the money for himself ($9) and offer the responder a small amount ($1), and the responder will accept this, because a small amount of the money is better than none at all.  In most cases, this is not what happens.  In most cases, the proposer offers to split the money in a proportion close to 50-50 ($5 for each); and the responder accepts.  When the proposer offers a smaller amount to the responder, the responder usually rejects the offer.  Apparently, responders are expressing their moral indignation against unfair offers.  This expression of moral sentiments in the Ultimatum Game has been shown to be correlated with activation of the brain's reward systems, which suggests that social norms of fairness have been internalized in the brain (Fehr and Camerer 2007; Sanfey 2007; Sanfey et al. 2003).

In 1995, Joseph Henrich was a graduate student in anthropology studying under Robert Boyd.  After hearing about the results in the play of the Ultimatum Game--mostly conducted with American undergraduate students--Henrich wondered how the game would be played by the people he was studing--the Machiguenga, who live in the Peruvian Amazon in small family-level groups that subsist on a combination of hunting, gathering, fishing, and slash-and-burn agriculture.  When he had them play the game that summer, he discovered that most proposers offered no more than 15 percent of the pot to responders, and that almost all of these offers were accepted.  So, in contrast to the American students, the Machiguenga were acting as rationally selfish maximizers, apparently confirming the Homo economicus model!

Boyd and Henrich decided that they should organize a large group of anthropologists and economists who would administer the Ultimatum Game to some small-scale societies around the world, representing a wide range of culturally diverse social organizations.  Twelve field researchers recruited subjects from fifteen societies.  The Ultimatum Game was played at each site.  At a few of the sites, the Dictator Game and the Public Goods Game were played.  All of the games are played anonymously.  The Dictator Game is played like the Ultimatum Game, except that the responder has no chance to accept or reject the offer.  The proposer dictates the split of the money, and so if he is generous, this must express his sense of fairness and not any fear of rejection.  In the Public Goods Game, the players are individually allocated some amount of money.  They can then contribute some of their allocation to a common pool, which is then increased and divided equally among all the players regardless of their contribution.  The selfish free-rider will not contribute anything to the common pool.

The fifteen societies were from twelve countries on four continents and New Guinea.  One was a purely foraging/hunting-gathering society (the Hadza of Tanzania).  Others included slash-and-burn semi-nomadic horticulturalists, pastoralists, and sedentary farmers.  The money allocated for each game was calculated to be the equivalent of an average day's wage for each society.  The studies of these societies were completed in 2000.

In 2002, a second phase of this project was started.  Four of the sites from the first phase were included in this second phase, and twelve new sites were added, including a group of Africans in a large city (Accra, Ghana) and a group of Americans in a small rural town in Missouri.

In this second phase, three games were played at every site: the Dictator Game, the Strategy Method Ultimatum Game, and the Third Party Punishment Game.  In the Strategy Method Ultimatum Game, the responder must say what his response would be to a range of possible offers from the proposer; and that determines the response once the proposer has made the offer, without the proposer knowing ahead of time what the response is going to be.  In the Third Party Punishment Game, two players are allotted a sum of money (the stake), and a third player gets half of this amount.  The first player must decide how much of the stake to give to the second player, with the second player making no decisions.  The third player must decide whether to pay 20 percent of his allocation to punish the first player across each of all possible offers.  This will measure the willingness of an individual to engage in costly third-party punishment to enforce social norms of fair behavior.  As in all of the games, the stake was set at one day's wage in the local economy, which meant that more money was involved in these games that is typically the case with the games using university students.

This experimental research supported four major findings (Henrich et al. 2014, pp. 131-33).  First, "fairness and punishment show both substantial variability and reliable patterns across diverse populations."  There is great cultural variability in human societies.  In all three experiments, there is substantial variability in the average offers and the willingness to punish low offers.  People in Western industrialized societies are very different from people in other societies, and so when undergraduate students in Western countries play these games, there is no reason to assume that their conduct shows a universal human nature. 

And yet this cultural variability does not mean that human beings are infinitely malleable blank slates on which culture can write just anything.  There is a clear pattern in this variation that shows how human culture is constrained by human nature.  Offers of 50 percent were always the most acceptable offer.  There were very few offers above 50 percent.  No society showed an average offer above 60 percent.  So there are no societies where most people give more than half, or where most people give zero.  The Hadza foragers were the most selfish people, but even they are not completely selfish.  In the Dictator Game, 71 percent of the Hadza offered more than zero.  On the other end of the scale, neither do we see completely other-regarding behavior.  In the play of the Dictator Game, only three individuals out of 427 offered 100 percent.  Hume and Smith were right in observing that human beings naturally show limited benevolence.

The second major finding is that "fairness increases with market integration."  Market integration was measured as the percentage of the average diet purchased in a market.  In all three games, the strength of fairness (making more equal offers) is correlated with increasing market integration.  The lowest average offers were made by the Hadza, a foraging band society in Tanzania with almost no market integration.  So it seems that markets promote morals by fostering social norms of fairness and cooperation.  A market society does not make people selfish, greedy, and amoral.

The third major finding is that "fairness increases with an individual's participation in a world religion."  In the societies studied, "world religion" means either Islam or some form of Christianity.  As opposed to the indigenous religions in some of these societies, Christianity and Islam teach that God is a powerful and moral divinity who punishes the bad and rewards the good.  This belief seems to reinforce social norms of fairness as reflected in the play of the Dictator Game and the Ultimatum Game.

The fourth major finding is that "willingness to engage in costly punishment increases with community size."  People from larger societies tend to punish more.  This is manifest in both second-party and third-party punishment.  This could explain the cultural evolution of social norms that made it possible--beginning about 10,000 to 5,000 years ago--for human societies to expand in size far beyond the small foraging bands that characterized most of human evolutionary history.  Those groups with social norms enforced by costly punishment could expand and outcompete those groups that lacked this cultural enforcement of group morality.

Henrich and his colleagues have argued that these findings of their experimental game project confirm the claims of Montesquieu, David Hume, and Adam Smith that more market-integrated societies foster social norms of fairness that facilitate expanded cooperation in the extended order of commercial exchange.


Jean Ensminger and Joseph Henrich, eds., Experimenting with Social Norms: Fairness and Punishment in Cross-Cultural Perspective (New York: Russell Sage Foundation, 2014).

Ernst Fehr and Colin F. Camerer, "Social Neuroeconomics: The Neural Circuitry of Social Preferences," Trends in Cognitive Science 11 (2007): 419-27.

Joseph Henrich, Jean Ensminger, Abigail Barr, and Richard McElreath, "Major Empirical Results: Markets, Religion, Community Size, and the Evolution of Fairness and Punishment," in Ensminer and Henrich, Experimenting with Social Norms, 89-148.

Joseph Henrich, Robert Boyd, Samuel Bowles, Colin Camerer, Ernst Fehr, and Herbert Gintis, eds., Foundations of Human Sociality: Economic Experiments and Ethnographic Evidence from Fifteen Small-Scale Societies (Oxford: Oxford University Press, 2004).

Alan G. Sanfey, "Social Decision-Making: Insights from Game Theory and Neuroscience," Science 318 (2007): 598-602.

Alan G. Sanfey, James K. Rilling, Jessica a. Aronson, Leigh E. Nystrom, and Jonathan D. Cohen, "The Neural Basis of Economic Decision-Making in the Ultimatum Game," Science 300 (2003): 1755-58.

Some posts on related topics can be found here, here,  here., here., here, and here.


CJColucci said...

I'm curious about how the authors arrived at the third finding, about participation in a world religion. Did they administer the game to people who adhered to "small" religions lacking a big-time deity overseeing our morals, and, if so, where did they find them? Or did they use non-religious Westerners, which would be problematic because of the inescapable influence of Big-God Religion even on people who no longer believe in it?

Larry Arnhart said...

Some of the people in the small-scale societies they studied were believers in indigenous religions rather than Christianity or Islam. Believers in indigenous religions were less fair in their Ultimatum Game offers.