Room mates cleaning the house each most prefers that

• Room-mates cleaning the
house
– Each most prefers that
everyone else cleans but she
does not
• Best for each
– Next, each prefers she cleans
and everyone else cleans
• Best for all
– Next, each prefers that she
does not clean and that
everyone else does not
• The equilibrium outcome - tragic
– Each least prefers that she
cleans and everyone else does
not
• Total loser
• Conservation of commons
• Overfishing
• International arms race
• Obeying just laws and having a
government
53
Social Dilemma Games
• Can be resolved by
– Moral,
– Social,
– Legal Regulation, or
– Combinations thereof
• Regulation changes it to a different
game
54
Humans Don’t Do
What Homo Economicus Predicts
• Experimental testing of anonymous one-round
social dilemma games
– Prediction: zero cooperation
– Reality: about half of subjects cooperate
• Logical inference: Thus, humans in these
experiments must be motivated by something
additional to dollar payoffs
• Do lots of experiments to discover those other
motivations
55
“Altruistic Punishment”
• In a repeated, anonymous, no-communication
social dilemma cooperation declines in each
round
– Economist Binmore: because they’ve learned how to
play the game
– No. Reshuffle players for another 10 rounds, and
cooperation again starts out high. Why?
– Why does cooperation decline with repetition?
• Repeated social dilemma with second-party
punishment option yields high cooperation
56

11/4/19
15
Strong Reciprocity
• In repeated, anonymous, two-person games,
cooperation is initially observed to be
reciprocated with cooperation,
– Independently of the dollar payoffs, i.e., contrary to
the rational choice of always defect based on the
dollar payoffs alone
• Further, third parties observing the games of
others will, at cost to themselves, reward
cooperators and punish defectors
– Doing so helps unknown strangers and the subject
takes home less money than she would otherwise
57
Communication and punishment in voluntary contribution experiments
B = std. social dilemma; R = punishment; FF = face to face
communication; FFwr = face to face & punishment
Olivier Bochet, Talbot Page, Louis Putterman
58
Communication vs. Punishment
• Communication is more effective than
punishment
– B&G are mistaken,
• communication is not symbolic punishment,
• rather punishment is a crude and clumsy form of
communication
– Praise and rebuke motivate; reward is a form of
praise and punishment a form of rebuke.
59
Behavioral Evidence
vs. Homo Economicus
• Weak moral motivations contrary to homo
economicus
– (dictator game)
• Strong reciprocity motivations contrary to
homo economicus
– Help those who help; harm those who harm
– Conditional cooperation: cooperate if and only if
others do
– (ultimatum, social dilemma games)
60

11/4/19
16
Evaluation of Strong
Reciprocity Evidence
• Third party punishment and reward by individuals
strongly established in repeated laboratory
experiments (Fehr)
• Objection: third-party punishment by an individual is
rarely observed in ethnographic studies or in life
(Guala)
• Bowles and Gintis reply: