L7: Efficiency of punishment and Third party punishment games

Question 1

What do Nikiforakis and Normann (2008) examine?

Answer 1

4 different levels of punishment effectiveness and one standard PGG without punishment

Question 2

Explain the structure of Nikiforakis and Normann (2008)’s experiment? and what y, n, ci, alpha and e are?

Answer 2

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)

Question 3

See and learn

Answer 3

payoffs for Nikiforakis and Normann (2008)

Question 4

3 findings of Nikiforakis and Normann (2008) ?

Answer 4

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

Question 5

What did Nikiforakis (2008) examine?

Answer 5

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)

Question 6

What were the 3 treatments in Nikiforakis (2008)?

Answer 6

1) VCM: Standard PGG wo. punishment
2) P: PGG with one-sided punishment
3) PCP: PGG with two-sided punishment

Question 7

What was the structure of Nikiforakis (2008)?

Answer 7

PGG

• 10 periods per treatment
• done for strangers and partners
• 2 treatments/subject (one then a diff. one)

Question 8

Explain how the PGG with 2-sided punishment worked?

Answer 8

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.

Question 9

2 findings of Nikiforakis (2008)?

Answer 9

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

Question 10

What are the conclusions of Nikiforakis (2008)? (2)

Answer 10

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.

Question 11

How does Nikiforakis (2008) explain: Total earnings: highest in VCM, then P&PCM not statistically different?

Answer 11

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.

Question 12

What did Herman et al. (2008) examine?

Answer 12

Whether how people punish and co-operate varies across society

Question 13

Structure of Herman et al. (2008)?

Answer 13

Study conducts 16 different PGGs with punishment in 16 developed societies with different economic and social backgrounds

Question 14

What was unusual about Herman et al’s (2008) findings?

Answer 14

They find widespread antisocial punishment (ie. punishing co-operators)

Question 15

3 findings of Herman et al. (2008)?

Answer 15

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 (?)

Question 16

What is rule of law?

Answer 16

Measures the extent to which agents have confidence in and abide by the rules of society

Question 17

What are norms of civic co-operation?

Answer 17

Measurements of how justified people think tax evasion, benefits fraud, or dodging fares on public transport are

Question 18

Main finding of Herman et al. (2008)?

Answer 18

Antisocial punishment is harsher in participant pools from societies with weak norms of civic co-operation and a weak rule of law

Question 19

See

Answer 19

enforcement of social norms (midway down side 2)

Question 20

What do Fehr and Fischbacher (2004) study?

Answer 20

Game with 2nd and 3rd party punishment - TP and SP DG and Prisoner’s Dilemma (PD)

Question 21

How does the third party punishment work?

Answer 21

The third party observes actions of players and punishes accordingly but is costly to themself!

Question 22

Why do Fehr and Fischbacher (2004) use DG and PD games?

Answer 22

DG has social norm of distributional fairness

PD has social norm of co-operation

Question 23

Fehr and Fischbacher (2004): Explain the set up of the TP-DG? (ie. what do A, B and C each do?) (3)

Answer 23

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)

Question 24

Fehr and Fischbacher (2004) could players incur losses?

Answer 24

YES if given enough punishment points, they had to pay them if so!

Question 25

Explain the hypotheses in Fehr and Fischbacher (2004) with TP-DG? (3)

Answer 25

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

Question 26

Main findings of Fehr and Fischbacher (2004) for TP-DG?

Answer 26

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

Question 27

What was the average A transfer and the average B expected transfer?

Answer 27

actual: 24.5
expectation: 21.8

Question 28

Explain Fehr and Fischbacher (2004) TP-PD set up?

Answer 28

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

Question 29

What is the strategy method?

Answer 29

Where decisions are elicited by asking ‘what would your response be for each different outcome’

Question 30

Fehr and Fischbacher (2004) TP-PD how are decisions elicited from third party punisher?

Answer 30

strategy method

Question 31

Fehr and Fischbacher (2004) TP-PD hypothesis?

Answer 31

Subjects will be willing to co-operate if others do therefore a norm of conditional co-op. will shape the TP-PD

Question 32

3 findings of Fehr and Fischbacher (2004) in TP-PD?

Answer 32

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)