Outline

OutlineIn-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)

Public Good Experiments

iii mmZ and

mmGSmZUij

jiiiii ))((

) and 1

ij

jiii mG(mSV

Z i

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:

Basic Insight

nSi

1If , then as n increases Vi = MPCR (marginal per

capital return) drops.

So it is unclear whether group size effect is due to changes in Vi = MPCR or changes in group size n.

Three Possible Treatments

The effect of altering MPCR but holding n constant

The effect of altering n but adjusting G(.) function so that MPCR remains constant

The combination effect in which altering n also altering MPCR.

Experimental Design

)1

ij

ji mG(mn

)

ij

ji mG(m

Experimental Procedure

Participants faced two consecutive series of ten decision trials.

In one series, MPCR was 0.3 and in another MPCR was 0.75. Order was switched to counterbalance.

12 experimental sessions.

Basic Results:% of Strong Free Riders

A person is called a strong free rider if her contribution to public good is less than 1/3 of her tokens

Data fromsecond series

Basic Results: % of Tokens Contributed

Data fromsecond series

End-Period Results

Data fromsecond series

Key Observations

Observation 1: Lowering the MPCR from 0.75 to 0.3 appears to increase significantly the incidence of free-rider behavior.

Observation 2: There are weak, if any, effects of changes in group size from four to ten (holding MPCR constant)

Intragroup Comparison I: Small Groups

Intragroup Comparison II: Small Groups

Combined Effect: Previous Experiment MPCR and Group Size

Summary