NEW EVIDENCE OF HUMAN BEINGS’ INSTINCTIVE FAIRNESS

Human instincts tell us to be fair, while our cognitive reasoning is needed to weigh up the pros and cons of bargaining for an advantage over someone else. That is the conclusion indicated by the research findings of Professor Ariel Rubinstein, who has asked thousands of people from around the world to participate in a revealing experiment that measured their response times in the Ultimatum Game.

His study, which is published in the October 2007 issue of the Economic Journal, reveals that making someone a fair offer takes significantly less time than making an unfair offer. The research also shows that women are much more sensitive to considerations of fairness than men.

The author, who is a professor of economics at Tel Aviv University and New York University, wanted to address the question, are human beings instinctively fair or instinctively selfish? Having had thousands of subjects participate in a virtual Ultimatum Game, he claims to have found indirect evidence that we are instinctively fair.

This famous game involves two players: a proposer and a responder. The proposer moves first by proposing how to divide a fixed sum of money, say $100, between the two players. Proposals are limited to natural numbers. The responder simply accepts or rejects the offer. Acceptance means that the proposal is implemented; rejection means that neither player gets anything.

The standard literature of game theory assumes that each player cares only about the dollar amount left in his pocket at the end of the game. Under this assumption, it is deduced that the rational responder accepts any offer that gives him a non-zero amount of money. Thus, it is optimal for the proposer to offer the responder not more than $1. Using game theoretical terminology, we say that in any of the game's subgame perfect equilibria, the responder gets $0 or $1.

The innovation in this study was the recording of subjects’ response times. Response time was defined as the number of seconds from the moment the server receives the request for a problem until the moment that it receives the response.

About 6,000 subjects in 25 countries participated in the experiment. Almost half of the subjects offered an equal partition of the money while 35% offered less than $50 to the other player. The other 15% offered more than half of the sum to the other player (which can be attributed either to error or the desire to give the other players more than half of the pie).

The study showed that females are much more sensitive to fairness considerations than males. The average offer made by women was about 10% higher than that made by men. 57% of the women offered an equal partition in contrast to only 46% of the men. While the responses of 15% of the men were in line with the game theoretic prediction, only 8% of the females offered almost nothing to the other player.

But the main finding of the paper is that the median response time (MRT) of those who offered less than $50 was 54 seconds, which is 25% higher than the MRT of those who offered an equal split. Given that it takes 20-40 seconds simply to read the question and respond, the difference is in fact very large. Furthermore, the cumulative distribution of MRT for those subjects who offered $50 is clearly to the left of the other distributions, indicating that the fair division is the quick answer in this case.

The paper concludes that making the fair offer takes significantly less time than making an unfair offer. This strengthens the conjecture that subjects who offer a fair division go through a different decision process than those who offer an unfair one. Subjects instinctively think first about offering the fair division.

Some stop there while others go on to consider the idea of offering less than half of the sum to the other player. This thought process involves the need to weigh the benefits of getting more against the risk of getting less. The resolution of this conflict is likely to require additional time.

Professor Rubinstein comments:

‘My research provides some evidence to be optimistic about human beings. Our instincts tell us to be fair and more cognitive resources are needed in order to use bargaining advantages in one’s favour.’

ENDS

Notes for editors: ‘‘Instinctive and Cognitive Reasoning: A Study of Response Times’ by Ariel Rubinstein is published in the October 2007 issue of the Economic Journal.

Ariel Rubinstein is professor of economics at Tel Aviv University and New York University.

For further information: contact Romesh Vaitilingam on 07768-661095 (email: romesh@compuserve.com); or Ariel Rubinstein via email: rariel@post.tau.ac.il