Dictator Game

This entry introduces and defines the Dictator Game (DG), a commonly used game in economic experiments. The DG was first used by Robert Forsythe, Joel L. Horowitz, N. E. Savin, and Martin Sefton in 1994, in which two anonymous players split money provided to only one of them. The person who receives money, the “dictator,” is given an opportunity to share this with the person who received nothing, the “recipient.” The DG was designed to reduce complex human behavior to simple steps to test game-theoretical predictions that assume rational utility maximization (i.e., maximization of the value of the money kept) based on mathematical calculations. It was assumed that dictators would not share money with anonymous strangers without some enforcement or punishment. It was a great surprise to learn that people shared, often splitting the money equally among themselves.

When two people who know each other or are related play the DG, it is expected that they will share. Standard economic theories, such as the neoclassical theory, for example, do not suggest that when playing with kin, one should maximize wealth at the expense of that kin. Even if the players don't know each other, if the game is played repeatedly and if the player's identity may be revealed later, economic theory suggests that there is reason to share. [Page 248]Building good reputation is a valuable asset, whose utility may be maximized. However, when complete strangers play a single game (one shot) with forever hidden identity, standard economic theories assume rational choices and expect the player to maximize wealth to self and to not share. However, in experiments played in various countries and cultures around the world, researchers found that the majority of people share with anonymous strangers in one-shot games. The game theoretic prediction that the “dictator keeps everything” happens only about 20% of the time. There is some cultural variation in how much money dictators keep. In some societies, zero offers simply do not exist. Dictators typically share between 20 and 50% of the money with anonymous recipients. The amount of money given to dictators as an endowment varies from a few dollars to substantial sums, sometimes several hundreds of dollars. Still, the size of the endowment seems to make little difference; 80% of the people share regardless how big an endowment they start with, and the percentage of money shared is stable.

Figure 1 Dictator game

Because in the DG sharing of the money is voluntary and the anonymous recipient benefits at the expense of the dictator, this game represent a classic example of altruistic behavior. There is gender bias in assessing who might behave more altruistically in the DG. When two boxes, one labeled from men and the other from women containing dictator donations, are placed in front of recipients, 80% of women and nearly 50% of men recipients prefer to choose from the from women box. Replacing human recipients with charities reduces donations. This indicates that there is something special about sharing specifically with people even if they are anonymous and even if the opportunity to share surfaces only once.

...