Contributing for myself, but free-riding for my group

1
Contributing for myself, but free-riding for my
group

• Christiane Schwieren
• Universitat Pompeu Fabra, Spain
• Yoshio Iida,
• Kyoto Sangyo University, Japan

2
Motivation

• Decisions in many real-world economic settings
  are taken by representatives of groups
• Little consideration in the experimental economic
  literature
• Intuition psychological literature suggest
  effect on behaviour

3
Basic design

• Public good game
• Representatives of small (3-person) groups
• 210 periods
• Previous to play Group discussion

4
Potential moderating variables

• Identification with own group
• Social identity theory
• Cultural aspects
• Individualism vs. collectivism
• Accountability

5
Design of our first experiment

• 2 2 design
• Country Spain vs. Japan
• Treatment Standard PG vs. PG of representatives
• Discussion in groups of 3 subjects (5 min., any
  topic)
• PG-game with 3 participants
• 5 tokens to distribute
• Token in personal account - 1 token for oneself
  (own group)
• Token in group account - 0,75 token for everybody
• Control-group design 2 (anonymous)
  representatives of each group play at the same
  time
• 2 10 periods, change in representatives

6
Results of first phase by country

• Spain
• No significant difference between representatives
  1 and 2
• Contribution to PG in standard PG significantly
  higher than in game with representatives over all
  periods in period 1
• Japan
• After exclusion of one outlier, no significant
  difference between representatives 1 and 2
• Over all periods, contribution to PG in standard
  PG significantly higher than in game with
  representatives
• No significant difference in period 1

7
Descriptive statistics over all periods
Man-Whitney U-test between treatments over all
periods
8
Differences between treatments for first period
9
(No Transcript)
10
(No Transcript)
11
Comparison between countries

• No significant differences in contributions
  between countries
• Neither in baseline nor in representative
  treatment
• Potential reason
• Hofstede Spain is high in collectivism
• No good comparison group if interested in
  influence of western individualistic values
• Maybe later data collection in Germany

12
Discussion

• Basic hypothesis about effect of being a group
  representative weakly confirmed
• Country-differences not confirmed
• Weak effects
• group identification not very strong
• No consequences, accountability low
• Further data collection necessary

13
Further research

• Collection of more data necessary
• Prior discussion with knowledge of game
• Explicitly to discuss strategy of game
• Second discussion period
• Unknown before playing
• Known before playing
• Selection of representatives by group