Outline In-Class Experiment on the Provision of Public Good Test of Free-Rider Hypothesis I: Marwell and Ames (1979) Test of Free-Rider Hypothesis II: Marwell and Ames (1980) Test of Free-Rider Hypothesis III: Issac and Walker (1988) Why: Kindness versus Confusion? Andreoni (1995)
Public Good Experiments Individual endowment is and n individuals in a group Invest in Private and Public exchanges: i’s returns from Private and Public exchanges are: Individual’s utility function:
Dependent Variables Investment in Public exchange: % of individuals who invest 0 in public exchange, Pr
Instructions – Private returns
Instructions – Public returns
This Experiment Individual endowment is $0.6 and 5 individuals in a group Invest in Private and Public exchanges: $0.6 – m i, m i i’s returns from Private and Public exchanges are respectively: Introduce a new treatment: Actual payoff based on ranking of experimental payoffs rather than the experimental payoff itself remove incentives for kindness such as altruism
Three Treatments Three treatment conditions Regular – actual payoff computed directly from experimental returns Rank – rank of experimental payoff is known, and actual payoff depends on the rank of the experimental returns RegRank – rank of experimental payoff is known, and actual payoff computed directly from experimental returns % of Free Rider (Rank) = an estimate of cooperation due to confusion % of Free Rider (Rank) - % of Free Rider (RegRank) = an estimate of cooperation due to kindness % of Free Rider (RegRank) - % of Free Rider (Regular) = An estimate of either cooperation or kindness
Rank Condition Actual Payoff = If there are ties in ranking, payoff is averaged off, e.g., if three subjects rank 1, actual payoff to each of the 3 subjects is ( )/3
Comparison
Pr
Findings As predicted, Cooperation in Rank < RegRank < Regular On average, 75% of subjects are cooperative Half of those who cooperate are really confuse, while half are really kind Information about rank decreases the amount given much more than it decreases the number of givers For some subjects, kindness depend on reciprocity
Summary