L7: Efficiency of punishment and Third party punishment games Flashcards
What do Nikiforakis and Normann (2008) examine?
4 different levels of punishment effectiveness and one standard PGG without punishment
Explain the structure of Nikiforakis and Normann (2008)’s experiment? and what y, n, ci, alpha and e are?
10 period PGG
All treatments: y=20 (endowment), n=4 (no. of players) c(i)=contribution alpha=0.4 (see notes and slides!!!)
e=effectiveness of punishment (1, 2, 3 or 4)
See and learn
payoffs for Nikiforakis and Normann (2008)
3 findings of Nikiforakis and Normann (2008) ?
1) Avg. contributions increase monotonically as effectiveness of punishment rises
2) Avg. contributions: a) decrease over time for 0&1, b) remain constant for 2, c) increase over time for 3&4
3) The higher the effectiveness of punishment, the faster the existment of punishment opportunities leads to a welfare improvement compared to treatment 0
What did Nikiforakis (2008) examine?
Counter-punishment: he examined if and how punished individuals counter-punished by allowing a counter-punishment stage in the game
Wanted to find out if counter-punishment opportunities led to a decrease in co-op. (ie. does it affect social welfare)
What were the 3 treatments in Nikiforakis (2008)?
1) VCM: Standard PGG wo. punishment
2) P: PGG with one-sided punishment
3) PCP: PGG with two-sided punishment
What was the structure of Nikiforakis (2008)?
PGG
- 10 periods per treatment
- done for strangers and partners
- 2 treatments/subject (one then a diff. one)
Explain how the PGG with 2-sided punishment worked?
A third stage is added after the punishment stage. Subjects are informed of the number of points that each of the other group members assigned to them. They are then given the opportunity to reduce the income of the individuals who punished them during the second stage by assigning counter-points.
2 findings of Nikiforakis (2008)?
1) introduction of counter-punishment mitigates the disciplinary effect of one-sided punishment
2) Earnings over time: increase in P, constant in PCP, decrease in VCM
3) Total earnings: highest in VCM, then P&PCM not statistically different
What are the conclusions of Nikiforakis (2008)? (2)
1) Those who counterpunish appear to motivated by a desire to hurt those who hurt them
2) Threat of revenge -> reduced willingness to punish free-riders and therefore -> decrease in co-op.
How does Nikiforakis (2008) explain: Total earnings: highest in VCM, then P&PCM not statistically different?
The benefits from higher cooperation are insufficient to offset the punishment costs and as a result, earnings in P and PCP treatments are lower than in VCM treatment. Earnings are not significantly different between treatments PCP and P.
What did Herman et al. (2008) examine?
Whether how people punish and co-operate varies across society
Structure of Herman et al. (2008)?
Study conducts 16 different PGGs with punishment in 16 developed societies with different economic and social backgrounds
What was unusual about Herman et al’s (2008) findings?
They find widespread antisocial punishment (ie. punishing co-operators)
3 findings of Herman et al. (2008)?
1) Punishment behaviour differed strongly across societies
2) The higher the level of antisocial punishment:
- the lower the avg. level of co-operation
- the lower the rate of increase in co-operation in punishment relative to N (?)
What is rule of law?
Measures the extent to which agents have confidence in and abide by the rules of society
What are norms of civic co-operation?
Measurements of how justified people think tax evasion, benefits fraud, or dodging fares on public transport are
Main finding of Herman et al. (2008)?
Antisocial punishment is harsher in participant pools from societies with weak norms of civic co-operation and a weak rule of law
See
enforcement of social norms (midway down side 2)
What do Fehr and Fischbacher (2004) study?
Game with 2nd and 3rd party punishment - TP and SP DG and Prisoner’s Dilemma (PD)
How does the third party punishment work?
The third party observes actions of players and punishes accordingly but is costly to themself!
Why do Fehr and Fischbacher (2004) use DG and PD games?
DG has social norm of distributional fairness
PD has social norm of co-operation
Fehr and Fischbacher (2004): Explain the set up of the TP-DG? (ie. what do A, B and C each do?) (3)
1) A endowed with 100points, can choose to transfer 0-50 points to B (multiples of 10 only), B must accept
2) 2nd stage: C (TP) can punish A: endowed 50 points, 1 punish point = -3 points for A
3) B indicates (prior):
a) expected punishment for each possible transfer level
b) expected transfer level (both secret to C and A)
Fehr and Fischbacher (2004) could players incur losses?
YES if given enough punishment points, they had to pay them if so!
Explain the hypotheses in Fehr and Fischbacher (2004) with TP-DG? (3)
1) If C only cares about own payoff, will never punish
2) If B believes C is selfish, will expect A to never be punished
3) IF social equality norm is considered, would expect punishment of A to increase as transfer amount deviates further from 50 points
Main findings of Fehr and Fischbacher (2004) for TP-DG?
1) Cs punished most As who transferred less than 50%
2) Roughly 60% of Cs punished As (Bs predicted greater % and greater severity than this)
3) Increase in P and E(P) as distance fro 50points increased
What was the average A transfer and the average B expected transfer?
actual: 24.5
expectation: 21.8
Explain Fehr and Fischbacher (2004) TP-PD set up?
2 stages:
1) A and B endowed with 10 points each, can choose to transfer 10 to the other (see and learn payoff matrix in notes)
2) C is endowed 40 points, 3p/pp again and 20 max to teach player, they observe A’s and B’s decisions in 1st stage and decides whether, and if so how much to punish A and/or B
What is the strategy method?
Where decisions are elicited by asking ‘what would your response be for each different outcome’
Fehr and Fischbacher (2004) TP-PD how are decisions elicited from third party punisher?
strategy method
Fehr and Fischbacher (2004) TP-PD hypothesis?
Subjects will be willing to co-operate if others do therefore a norm of conditional co-op. will shape the TP-PD
3 findings of Fehr and Fischbacher (2004) in TP-PD?
1) 46% of Cs punished if result was (D,C) or (C,D)
2) 21% punished for (D,D)
3) Negligible punishment for (C,C)
