2 - LEC 5 - SOCIAL PREFERNCES Flashcards
How does the ultimatum game work
- One-shot game
- Two players have to divide a fixed pie p
- The Proposer (P) proposes a division (x1,x2)=(x,p-x)
- The Responder (R) observes the offer and either accepts or rejects
- If she accepts the agreed upon division is implemented
- If she rejects. Both players get nothing
- Backwards induction;
- A rational R (second player) will always prefer the alternative which yields more for her and will choose conflict only if this does not cost her anything
- Optimal decision;
- P proposes 100-ε
- R accepts ε
how does the dictator game work
- One-shot game
- P decides on division (x,p-x), R has no choice but to accept
-Any positive offer is altruistic rather than strategic (caused by the fear of rejection)
how would a experiment work using a separate dictator and ultimatum game
Instead.. Randomised into two groups
- P in either Dictator or Ultimatum
- R in either Dictator or Ultimatum
- All subjects were both P and R. But only P and R once
- Different partner in Dictator and Ultimatum
- Two randomisation procedures
-As close to two different experiments as possible
-Ultimatum game finished before Dictator offer was revealed
- don’t want to think your against anyone
-problems: Anonymity
Single-blind
Anonymity with respect to other subjects
Double-blind
Anonymity with respect to other subjects and experimenter
Incentivized
No talking amongst subjects
whats the model of Utility with Wealth and Morality
𝑈_𝑖 (𝑎, 𝑣, 𝑛, 𝑠)=𝑀_𝑖 (𝑎,𝑣,𝑛,𝑠)+𝑊_𝑖 (𝑎,𝑣) i; individual a; action (choice of action has two effects) v; monetary value of game (the stake) W; wealth (increasing in v) M; nonpecuniary costs or benefit n; social norms (against a behaviour) S; scrutiny (eg. who is watching?)
- Utility is additively separable in the moral and wealth arguments
- The greater the impact of an action on others; the more negative the moral payoff
- M is increasing in v
- A greater value of n is associated with a greater norm against a behaviour
- M is increasing in n
- A greater value of s is associated with a greater moral costs
- M is increasing in s
what are the social norms that influence behaviour in dictator and ultimatum game
Descriptive
framing effects
Eg. Ultimatum games framed as dilemmas where resources are shared tend to make players more generous
Difference between “contribute” and “allocate”
Socially shared belief regarding how one ought to behave
Some cultures have a strong norm for sharing
Reputation-building behaviour
Some play the one-shot Ultimatum Game as if they have repetition
Demographic
Eg. academic major and culture (student presentations)
Why do receivers reject positive offer?
Reciprocity (‘tit-for-tat’)
Conditional co-operators
People are prepared to cooperate as long as others in the game cooperate and will detect if others do not cooperate
Strong reciprocity
Also involves willingness to punish others
whats the behavioural aspect of it ultimatum game
Altruism
Benefit to other and involves a cost on originator
Spiteful
Cost on others and involves a cost on originator
Example; shouting at people who litter
whats the difference between pure and impure altruism
Pure altruism
An individual’s utility increases as the utility of another player increases
Impure altruism
An individual’s utility increases from the act of contributing to others
where does the nature of these social preferences come from
-Reference transactions
Normally determined by past history of transaction
In line with prospect theory’s emphasis on reference point
Fairness based on what is “normal” rather than on what is “just” (tipping)
-Outcomes to the transactors
Dual entitlement. Usually regarded as unfair if one transactor makes a gain by imposing a similar loss on the other
Prospect theory; Loss aversion
overall welfare is reduced if gains are made at the expense of equal losses
where does the nature of these social preferences come from
-Reference transactions
Normally determined by past history of transaction
In line with prospect theory’s emphasis on reference point
Fairness based on what is “normal” rather than on what is “just” (tipping)
-Outcomes to the transactors
Dual entitlement. Usually regarded as unfair if one transactor makes a gain by imposing a similar loss on the other
Prospect theory; Loss aversion
overall welfare is reduced if gains are made at the expense of equal losses
describe the Levitt, List 2007. What do Laboratory Experiments Measuring Social Preferences Reveal about the Real World. paper
Randomised into two groups :
-P in either Dictator or Ultimatum
-R in either Dictator or Ultimatum
All subjects were both P and R. But only P and R once
Different partner in Dictator and Ultimatum
was incentivised
issues:
- small sample size
- more generous because know each other
- no anonymity
describe the guth et al. 1982 ultimatum game
Subjects were in one room, but no subject knew the person with whom he/she were paired
Real monetary stakes
Two treatments:
- Naïve (inexperienced) subjects
- Experienced subjects; same experiment one week later with same subjects
naive subjects offered more than experienced
What accounts for this behaviour?
Dictator and Ultimatum:
Fairness, altruism
Ultimatum:
Strategic concerns, fear of rejection
describe the Forsythe et al. 1994. Fairness in Simple Bargaining Experiments. Comparison of Dictator and Ultimatum Game paper
Ho; Nontrivial offers in ultimatum and dictator games can be explained by proposers’ concerns with fairness
Ha; Other factors influence the offer distribution
Test; No difference in distribution of offers across the dictator and ultimatum game
If difference; we reject Ho and accepts Ha. If no difference; we cannot reject Ho
also tested incentivised and stake hypothesis
With pay;
Difference D and U
Rejects the fairness hypothesis. Accepts Ha (strategic concerns matter as well)
No pay;
Rejects the fairness when April ultimatum is used as comparison (4)
Otherwise, not rejected
Pay hypothesis:
Dictator; Rejected (pooled); i.e. incentives make a difference
Ultimatum; cannot reject (i.e. incentives don’t make a difference)
Cannot reject the stake hypothesis
Ie. the $5/$10 stake doesn’t make a difference
Do higher stakes lead to more equilibrium pay?… proof from papers
Forsythe R et al., (1994)Games and Economic Behavior) Stakes: $5 and $10
didn’t matter if stakes were high or low
in Cameron (Economic Inquiry; 1999); Ultimatum game experiments in Indonesia
they found:
-Respondents more willing to accept a given percentage in higher stakes games
-Offers approach 50/50 with increasing stakes (no spite)
describe the Cherry TL, Frykblom P, Shogern, JF, 2002. Hardnose the Dictator. paper
6 different treatments
Results; proportion of Zero offers
Double blind with earnings > With earnings > baseline
