14 - punishment robustness

Question 1

Fehr and Gachter 2000 findings

Answer 1

if you have opportunity to punish after each stage - can prevent/reverse decay

Question 2

what were features of FG 2000

Answer 2

• students
• you can only punish based on their contributions in the current round - dont know what they did in previous rounds
• punishee doesnt know who punished them
• no scope for targeted attacks

Question 3

what did Herrman, Thoni and Gachter 2008 aim for

Answer 3

conducted the ‘same’ experiment in different countries
- do results differ depending on societites and different subject pools

Question 4

Hermman, Thoni Gachter

experiment

Answer 4

• 16 cities
• similar subjects - students age demographic
• translated everything, local curries
• consistency
• design like FG 2000 - 10 rounds, paretners matching, with and without punishment
• punishment is costly for both

Question 5

HTG
what is punishment of free-riding

Answer 5

when the punisher contributes more than the punishee

• they deserve to be punished

Question 6

HTG

what is anti-social punishment

Answer 6

punisher contribute less than punishee

Question 6

HTG findings
punishement of free riding
antisocial punishment

Answer 6

free riding
- similar results across 16 cities
- people punish those that deviate the most from the average the most (a similar proportion)
- punishment of people below the average is similar

antisocial
- punishment of people above the average group contribution varies amongst countries
- intensity of punishment varies
- some countries dont punish those above, so do

the way that punishment is used varies across countries

Question 7

HTG
what is the correlation between
antisocial punishment and contribution levels

Answer 7

the more antisocial punishment - the less contributions there are on average

Question 8

what do HTG do to determine what drives the differences between cities

Answer 8

group them by culture
how does culture affect the use of antisocial punishing
- categorise into 6 clusters

Question 9

how will we know if cultural effects are happening
evidence

Answer 9

if variation across cultural groups exceeds variation within groupings

Question 10

HTG findings
culture clusters

Answer 10

found that there is more variation across groupings than within
there are cultural effects determining anti-social punishment

• but doesnt identify specific cultural drivers

Question 11

what are 2 experiments that vary the game FG 2000

Answer 11

1. Nikiforakis (2008)
2. Nikiforakis & Engelmann (2011)

Question 12

Nikiforakis 2008

what is the difference in design

3 treatments

Answer 12

• VCM, partners and strangers, 3 treatments

1. VCM
2. P - after contributions round can punish as in FG
3. PCP - as P but each round has 3rd stage = after stage 2 subjects told how many punishment each group member assigned them
   - the punished can assign punishment points in stage 3 but only to those that punished them

allows counterpunishment

Question 13

Nikiforakis
PCP

results

Answer 13

• VCM - decay (same S and P)
• P - no decay (S and P)
• PCP - in between the 2 - closer to VCM decays
• punishment is lower in PCP than P - because of threat of retaliation

Question 14

conclusions from Nikiforakis
2008

Answer 14

• CP reduces the upward effect of P on contributions relative to VCM

Question 15

Nikiforakis and Engelmann (2011)

aim

Answer 15

see whether allowing punsihment to carry on across stages makes a difference

• Nikiforakis and FG - prevented info flowing across rounds - would only know who gave you the punishment in that round
• will creating a scope for feuds impact contribution levels - increase possibility of punishment and retaliation

Question 16

how did NE 2011
create scope for feuds

Answer 16

1. stages of punishment can continue - no restriction
2. less restriction on who can punish who (no longer case where you can only punish who punished you)
3. identity tracking across rounds

Question 17

NE 2011 experimental design

3 conditions

Answer 17

1. baseline = VCM - indentities change each round
2. short feuds = can only punish and PCP in that round - identies still change
3. short feuds but identities constant across rounds - feuding across multiple rounds

Question 18

NE 2011 resulsts

Answer 18

• feuds are rare but do occur
• SF and LF have more incidences where there are no punishing at all - people scared to punish
• punishment is most intense in SF
• LF dont maintain contributions = decay

Question 19

what is the effect of availability of punishment

Answer 19

• PCP = deters punishment - decay - less contributions