For instructors

Teaching notes

A supplier ships goods before the payment for them arrives. An investor wires money to a founder who will decide later how much to return. An employer pays more than the market requires and hopes for effort that no contract can specify. Each transaction creates value only if the first mover trusts and the second mover reciprocates.

What the game shows

The game measures trust, as the amount an investor sends, and reciprocity, as the share a trustee returns, in a setting where backward induction predicts neither. It shows how much potential surplus depends on behavior that self-interest does not explain. The gift-exchange variant applies the same structure to employment and tests whether higher wages are answered with higher effort.

The theory

In the investment game of Berg, Dickhaut and McCabe (1995), whatever the investor sends is multiplied on the way to the trustee, who then decides how much to send back. A trustee who cares only about money returns nothing, an investor who foresees this sends nothing, and the gains from the multiplication are lost. The prediction is the same in every round of a game of known length.

The gift-exchange game takes its motivation from Akerlof (1982), who described the employment relation as a partial exchange of gifts in which a wage above the minimum is met by effort above the minimum. In the experiment of Fehr, Kirchsteiger and Riedl (1993), an employer offers a wage and the worker then chooses a costly effort. Since effort cannot be enforced, the prediction is the lowest effort and, in anticipation, the lowest wage. Sending money can also express generosity without any expectation of return, and Cox (2004) shows how control treatments with dictator games separate the two motives.

What to expect

In Berg, Dickhaut and McCabe (1995), investors had ten dollars and the amount sent was tripled. Thirty of 32 investors sent something, the average amount sent was $5.16, and the average amount returned was $4.66, so that trust came close to breaking even without quite paying. The meta-analysis of Johnson and Mislin (2011), covering 162 replications, puts the average amount sent at about half of the endowment and the average amount returned at about 37 percent of what the trustee had available.

In the gift-exchange experiment of Fehr, Kirchsteiger and Riedl (1993), the average wage was 72 where the market-clearing wage was 30, and average effort was 0.4 where the minimum was 0.1. Effort rose with the wage, which is the relation that makes a high wage profitable for the employer.

Tips and things to look out for

References