Social Preferences Flashcards
What is game theory equilibrium?
set of best strategies players pick to maximise payoff.
What is Nash Equilibrium?
point where no player benefits from changing their action if others keep their decision constant.
What can be said about prisoners dilemma regarding Nash?
1 - Nash equ. = cheat, cheat
2- Nash does not mean pareto
3- In a repeated game backwards induction shows sublime perfect equilibrium is cheat, cheat
Who is Robert Axelrod?
economist that showed tit-for-tat tactic was better than cheat, cheat
- Got game theorists to come and submit a strategy and then competed the tactics
What is a tit-for-tat strategy?
- Repeats the opponents action in the last round, starting with coop.
- promotes cooperation and punishes cheat
What did Axelrods investigation discover regarding the tit-for-tat strategy?
- Tit-for-tat strategy collected the most points overall
- In a head to head, cheat, cheat wins but two t4t players will coop every round
What’d Adreoni and Millar’s repeated PD experiment show?
-people cooperated initially
-coop collapsed towards the end, reasons:
1-Rational cooperation (building good rep for private gains)
2- Altruism (natural tendency to coop)
What are the altruism models?
1- Pure altruism: care directly about others (Ui = pi + apj)
2- Duty: cooperating brings you enjoyment (Ui = pi + a, a>0 if i coops)
3- Reciprocal Altruism - both need to coop to improve utility (Ui = pi + a, a>0 if i&j coop)
How did Adreoni and Millar test for cooperation reasoning?
3 conditions:
1- Strangers, play one round with a random person
2- Partners, play ten rounds with a random partner (opp. for rep building)
3- Computer 50s, play with either stranger or computer for 10 rounds, computer plays tit for tat
What did Adreoni and Millars test for reasoning find?
- Findings are consistent with reputation building
- Coop still present with strangers so stable fraction are altruists.
What was the ultimatum game?
- one person is given £10 and they have to propose how to split the £10, receiver can accept or reject the offer
- If receiver rejects the offer, both get nothing
What is the nash equilibrium for the ultimatum game?
offer £0.01
What were the results of the ultimatum game?
- mainly offer £4-5
- almost no offers below £2
- Lower offers higher chance of rejection
- Reciprocal Altruism to foster cooperation
- Reciprocal punishment against unkind behaviour.
- Don’t know how much altruism is due to ‘rational irratinality’ - not wanting to get rejected so offering more
What is the dictator game?
- Proposer given £10
- They choose the offer and it cannot be rejected
What are the findings for the dictator game?
- Less offered than ultimatum game, shows some of altruism must have been reciprocal
- Forsythe ‘94: 20% give half, 79% give >0
- Hoffman ‘94: <6% give half, 33% positive
- Hoffman’s study had double anonymity, this suggests people are intrinsically selfish but don’t want to appear so.
What is the public good game?
- 80 subjects, each given 225 tokens
- Can keep token or add it to the public pot
- Payoff from public pot depends on contribution, more contributed, more paid out
- if all pay public, payoff = £4.95 each
- if one free rides, payoff = £4.40 + £2.25 = £6.65
- Repeated 10 times
What is the Nash equilibrium for the public good game?
- All private pot
What are the findings of the public good game?
- first round ~1/2 public pot
- Public pot falls after each round, virtually no public pot by the end
What are the Causes of decay in the Public good game?
1- Learning, Players learn free riding is optimal
2- Rational Coop, understand free riding is optimal but tries to coop
3- frustrated kindness - gives up when others are selfish
What was the Isaac and Walker experiment and what did it show?
- played two games consecutively
- end of first game 80% free ride
- start of second game public pot high again
- rules out learning!
What was Adreoni’s experiment in 1995?
- investigation into whether Isaac and walker results were due to kindness or confusion
- 2 game modes:
1) Regular
2) Rank - total payout fixed, private payout depends on if you beat others. No incentive to be kind
What were the findings of Adreoni’s experiment in ‘95?
- Lower Public contributions in rank, more free riding
- 1/2 of cooperation moves are due to kindness
- Decay is due to frustrated kindness
How did Fehr/ Gachter suggest we can prevent the decay in public good games and how well did it work?
-punishmwnt mechanism
-standard game theory suggests costly to punish
Findings:
-More public contributions
-Cooperatino was sustained
-Heavy punishment for poor contributors
What is the nash equilibrium for the dictator game?
£0
