Brainscape's Knowledge GenomeTM


Top Flashcards (Certified)
All Flashcards

ECON 3028 Adv Exp +Behavioural Economics > 5. Public Goods Game > Flashcards

5. Public Goods Game Flashcards

1
Q

Issues with experimental games

A

-Games may be complex
-Should we expect NE to hold unconditionally or only on completion of a convergence process
-Is utility entirely determined by payoffs

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Describe the generic public good game

A

Subjects play in groups. Each player chooses to split tokens between private and public account. Tokens in public account give everyone money

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Typical findings of public goods game

A

If game is played once only 20% if people free ride. On average people contribute 40-60% to the public account. If the game is repeated with feedback the contribution rate decays

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What are the explanations for why people contribute?

A

-Error: players are confused about the rules
-Strategic: players may think contributing in early rounds of repeated game will raise future contributions of others
-Preference explanations: altruism, warm glow, conditional cooperation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What is the one sided error problem?

A

It is the fact that since the dominant strategy is 0 in the standard VCM design, all errors are in the same direction, making it hard to distinguish from intended contributions

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What was the aim of the Keser 1996 experiment?

A

To redesign the game so errors are possible on both side of the NE contribution level. That way we can distinguish between random errors and intended behaviour

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Results of Keser 1996 experiment

A

Average contributions barely drop over time and are always above optimum. This shows over contribution isn’t purely due to errors

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What was the aim of Andreoni 1988 experiment?

A

Aims to separate the learning hypothesis and the strategic hypothesis

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Design of Andreoni 1988 experiment

A

-partners: play game repeatedly in same group
-strangers: groups are randomly redrawn after each round
-surprise restart: 3 extra rounds added on

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Results of Andreoni 1988

A

-people contribute more in strangers- weird result, not been consistently replicated
-decline in contributions over time until new restart then there is a large jump in partner group

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Design of Yamakawa 2016 experiment

A

-Hunan treatment (H): standard VCM with n=2, partners protocol
-Computer treatment (C): same as H, except each group comprises of one human player and a computer with pre determined choice
-HC treatment: same as C except computer plays on behalf of another human who receives its payoff

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Results of Yamakawa 2016

A

Least contribution in C, slightly more in HC. A lot more contribution in H. Relatively constant over time, drops for H in final round

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What does the contribution table capture?

A

The preferences of a subject- how much they are willing to contribute at each level of contribution from others

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What is the P experiment?

A

It is the FGF form of the strategy method used to classify subjects into types of contributors (ie conditional contributors, free riders, and triangle contributors)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What is the C experiment?

A

Standard 10 round repeated VCM game plus elicitation of beliefs of average contributions of others round by round

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What are the incentives in the P experiment?

A

Any element of the contribution table could affect the payoff of the subject

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

Breakdown of subjects in FGF 2012 by preference types

A

55% are conditional cooperators, 23% are free riders, 12% are triangle contributors

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

Aggregate findings of FGF 2012 experiment

A

Decay of contributions and decay of beliefs about others’ contribution

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

Results of individual findings in FGF 2012

A

-some free riders contribute in early rounds- evidence of strategic thinking
-in one shot game, free riders become perfect free riders
-imperfect conditional cooperators want to undercut contributions they expect from others
-imperfect conditional cooperators and free riders mean that sustained contribution is unlikely

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

Reciprocity

A

The desire to reciprocate whatever others give you

21
Q

When does individual i get positive utility from an increase in the other persons income?

A

When i has a greater income than the other person

22
Q

What is the punishment condition in Fehr and Gachter 2000?

A

Players are told contributions of others and can assign punishment costs to them anonymously. This is costly to both parties

23
Q

What does the theory predict will happen in punishment?

A

There will be a solution where no one contributes and no one punished in every round of a finitely repeated game

24
Q

What are reasons for punishing?

A

-Strategic
-Reciprocal reasons
-Reduce unfavourable inequality

25
Results of strangers with punishment in Fehr and Gachter 2000?
-with no punishment contributions decay over time -with punishment contributions are higher and stable -immediate jump when punishment is introduced/ removed -others who contribute a similar small amount to them are punished
26
Results of partners in punishment game Fehr and Gachter 2000?
-with no punishment contributions decay over time -with punishment contributions are higher and even rising -jump when punishment is introduced/ removed -80% contribute entire endowment in final round -average contribution across all rounds with punishment is 17/20
27
Describe the Herman, Thoni and Gachter 2008 experiment
The same experiment as FG 2000 but across 16 universities in different cities around the world.
28
Results of Herman, Thoni and Gachter
-Punishment of free riding is similar in all cities -Anti-social punishment varies drastically -Without punishment, all cities show decay of contributions at similar rate but at different absolute levels -With punishment, all cities have stable contributions but larger variance between countries -Negative correlation between anti-social punishment and contributions
29
What were the findings of Gachter, Herman and Thoni 2010?
They found evidence of cultural effects by grouping countries into cultural groups and see greater variation across groups than within
30
What are the three punishment treatments in the Nikiforakis?
-VCM treatment- no punishment -P treatment- each round after stage 1 contributions, punishment opportunity at stage 2 -PCP treatment- those who have been punished in stage 2 have the opportunity to retaliate in stage 3 against those who punished them
31
Results from Nikiforakis 2008?
-VCM and P show usual results. PCP decays over time but not as quickly as VCM -Among strangers, PCP is very close to VCM -Punishment is lower in PCP than P -Approx 25% of punishments provoke counter punishment
32
What is the aim of Nikiforakis and Engelmann 2011?
To allow for more realistic retaliation in the form of feuds
33
How do Nikiforakis and Engelmann allow for feuds?
-make number of stages of punishment in each round endogenous -less restrictions on who may punish who -identify tracking across rounds
34
What are the three treatments in Nikiforakis and Engelmann 2011?
-Baseline: standard VCM game with one stage of punishment opportunity each round. Players have identities which change each round -Short feud: same as baseline except punishment stages can continue as long as subjects want. Players told punishment points assigned by each player to others -Long feud: same as shirt feud but now identities are held across rounds
35
Results of Nikiforakis and Engelmann 2011
-feuds happen but are rare -short feuds and long feuds have more instances of no punishment at all. People may be deterred from punishing by threat of feud -similar levels of contribution across all 3, slightly less in long feud because people are less inclined to punish
36
Describe the ultimatum game
Proposer offers a division of the total money between the two players. Responder can accept or reject the offer
37
What are the typical experimental findings of the ultimatum game?
-average offers are around 40% of T -50% offer is most common -offers over 50% are very rare -offers of less than 20% are rejected half of the time
38
Describe the dictator game
The proposer splits the money pot between himself and the other player. The other player can’t reject what Is offered
39
Describe the set up of the restricting the proposer version of the ultimatum game by Falk, Fehr & Fischbachter 2003
-four versions of ultimatum games except the proposer is only allowed a choice between two offers: 20% or (50%,80%,20%,0%) depending on the treatment. -in each treatment the responder knows the choices available to the proposer -all 4 rounds are played and the responders behaviour for both offers is elicited using the strategy method
40
What is found from the “restricting the proposer” paper by Falk, Fehr & Fischbachter 2003?
The rejection rate of the 20% offer varies with the alternative offer. Rejection of highest in order of 50%,80%,20%,0% as the alternative offer
41
What is the aim of Blanco, Engelmann & Norman 2011?
It considers if the Fehr- Schmidt model can explain subject’s behaviour across different games in a unified way
42
What determines the upwards slope of i’s ICs?
Alpha
43
What determines the downwards slope of i’s ICs?
Beta
44
How can we interpret alpha and beta?
Alpha is an envy parameter, beta is a guilt parameter
45
Describe the set up of BEN 2011
All subjects play all of -Ultimatum game, as proposer and responder -Modified dictator game -Sequential prisoners dilemma, as 1st and 2nd mover -2 player public good game Use play in two roles/games to work out alpha and beta and then use these values to predict behaviour in other games
46
In BEN 2011 how do we measure alpha?
We use decisions as responder in ultimatum game. Ask responder for their MAO. Set this equal to a situation where the responder receives nothing
47
In BEN 2011 how do we work out beta?
Use a modified dictator game where the subject chooses between (£20,£0) and (£x, £x) for different values of x. Let x* be the value where indifference occurs. Use these points to work out beta
48
How effective are alpha and beta in predicting ultimatum proposals?
Very accurate on aggregate but not for individuals
49
Does it matter that individual predictions of alpha and beta aren’t accurate if the aggregate predictions are accurate?
-One view: if we are only interested in aggregate findings then it does the job. -Alternative view: even if we are often interested in aggregate findings, a theory that doesn’t predict we’ll at the individual level hasn’t gotten to the bottom of things

ECON 3028 Adv Exp +Behavioural Economics (10 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2025 Bold Learning Solutions. Terms and Conditions