Transcript Collective Action Behavior and Social Institutions

Institutions and the Evolution
of Collective Action
Mark Lubell
UC Davis
Defining Collective Action

Collective-action problem: Individual
decision-making leads to socially
undesirable (Pareto-inefficient) outcomes

Cooperation: Adjusting behavior to
minimize socially undesirable outcomes
Tragedy of the Commons

Garrett Hardin (1968):
“Therein is the tragedy. Each man is locked into a system that compels
him to increase his herd without limit—in a world that is limited. Ruin is
the destination towards which all men rush, each his own best interest
in a society that believes in freedom of the commons.”
“Mutual coercion, mutually agreed upon”

Flip side of resource use: Maintenance of
ecosystems/public goods
 Collective action problems are ubiquitous!
From Global….
To Local…
To Local…
Studying Collective Action
Major Research Questions
1.
Factors explaining cooperative behavior
2.
Role of institutions (e.g., punish defection, reward cooperation)
Theoretical

Philosophy

Game theory

Evolutionary game theory

Evolutionary simulations (This talk)
Empirical

Field research (qualitative and quantitative)

Experimental research
Prisoner’s Dilemma
Player 2
Cooperate
Defect
Cooperate
R1= 6
R2= 6
S1 = 3
T2= 8
Defect
T1= 8
S2= 3
P1 = 4
P2 = 4
Player 1
Conditions: T>R>P>S; 2R>T+S
Nash equilibrium: Both players defect
Collective Action Agents
Five “gene” strategies; 32 possible
 Each gene determines behavior in current
round on basis of outcome in last round

<Nice (1st round), Reciprocal (CC), Sucker(CD), Forgive (DC),
Protect (DD)>

Important Examples:
All Cooperate <1,1,1,1,1>
GRIM Trigger <1,1,0,0,0>
PAVLOV(Win-stay, lose shift) <1,1,0,0,1>
Tit-for-Tat <1,1,0,1,0>
Structure of Simulation
Generation
1:
Randomly
Select 40
Strategies
Round Robin
Tournament: Each
strategy vs. itself and all
others
Next Generation:
Survival of Fittest
1% Mutation Rate
on Each Gene
Generation 1
Proportional Fitness
Reproduction:
P(reproduction)=
Fitnessi/Fitnessall
Generation
5000
A “Punishing” Experiment
Design
 Baseline 2-player repeated PD, with discount
rate= .9
 Examine the effect of $2 punishment for defection,
with increasing probability ranging from [0,1] in .10
increments
 10 runs of each experiment; 40 strategies, 5000
generations
Hypotheses
 Increasing levels of cooperation
 Increased population stability
 Shift in the population dynamics of cooperation
4914
4625
4336
4047
3758
3469
3180
2891
2602
2313
2024
1735
1446
1157
868
579
290
1
7
6
5
4
3
Mean
Fitness
2
1
0
45
40
35
30
25
20
grim
pavlov
15
10
5
0
4999
4705
4411
4117
3823
3529
3235
2941
2647
2353
2059
1765
1471
1177
883
589
295
1
Hobbes: Punishment p=1.0
7
6
5
4
3
Mean
Fitness
2
1
0
45
40
35
30
25
20
grim
pavlov
15
10
5
0
Mean Fitness Increases With Punishment
Probability
5.8
5.6
Mean Fitness
5.4
5.2
Generation Mean
Fitness
5
4.8
4.6
4.4
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Punishment Probability
1
Gene Frequency: All Regimes
Nice
CC
CD
DC
DD
Average Gene Frequency Per Generation
40
35
30
25
20
15
10
5
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Punishment Probability
0.8
0.9
1
Strategy Frequency: All Regimes
14
Average Strategy Frequency
12
10
8
GRIM
PAVLOV
SPAVLOV
6
4
2
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Punishment Probability
0.8
0.9
1
Gene Frequency: Cooperative Regimes
(Avg. Fitness>5.9)
Nice
CC
CD
DC
DD
45
40
Gene Frequency
35
30
25
20
15
10
5
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Punishment Probability
0.8
0.9
1
Strategy Frequency: Cooperative Regimes
34
Average Strategy Frequency
29
24
GRIM
PAVLOV
SPAVLOV
TFT
19
14
9
4
-1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Punishment Probability
0.8
0.9
1
Some Correlations
Overall Fitness
.21
Genes
Nice
.11
CC
.22
CD
.10
DC
.06
DD
.24
Strategies
All Defect
-.18
GRIM
-.03
PAVLOV
.10
Suspicious PAVLOV
.10
TFT
.04
Conclusions




Punishment institutions increase cooperation and
stability, even in noisy environment
As punishment increase, basis of cooperation shifts
towards PAVLOV
Institutions change population dynamics of
cooperation, even if same behaviors observed
Must square with observed human behavior; e.g.;
resistance to coercion, reduced effectiveness of
reciprocity in coercive environments