Behavioural Economics in Social-Ecological Systems with Thresholds
1. Behavioural Experiments in Social-
Ecological Systems with Thresholds
Juan C. Rocha, Caroline Schill, Anne-Sophie Crepin &
Therese Lindahl
@juanrocha
2. How does people behave when dealing with
situations pervaded by threshold (uncertainty)?
3. Framed field experiment
• 256 participants / 4 fishing
communities
• Largely depend on natural
resources as main source of income
• History of regime shifts - critical
transitions:
• Mangroves collapse (1980s -90s)
• Hypoxia events and fish
deadlines
6. • Dynamic game
• Monetary incentives: 1 fish worth 500 Pesos
• Communication allowed throughout the
game
• Participants did not know the # rounds to
be played
• Session lasted about 3 hrs
• Average earnings = ~20k COP ~€ 5.50;
range: 3,000 - 40,000 COP (+ show-up fee
of 15,000 COP), about a daily wage, paid
privately at the end of the session
Experimental design: set-up and
procedure
Earnings paid privately at the end
Communication allowed
Monetary incentives
7. STAGE 1: 6 rounds
For all groups the same,
independent of treatment
Baseline
Dynamic common pool resource game
Start
8. STAGE 1: 6 rounds
For all groups the same,
independent of treatment
Baseline
STAGE 2: 10 rounds
All groups start again
with 50 resource
stock units
Dynamic common pool resource game
Start
9. STAGE 1: 6 rounds
For all groups the same,
independent of treatment
Threshold
• climate event has happened
Baseline
STAGE 2: 10 rounds
All groups start again
with 50 resource
stock units
Dynamic common pool resource game
Start
10. STAGE 1: 6 rounds
For all groups the same,
independent of treatment
Threshold
• climate event has happened
Risk
• long-lasting climate event might
happen, known probability (0.5)
Baseline
STAGE 2: 10 rounds
All groups start again
with 50 resource
stock units
Dynamic common pool resource game
Start
11. STAGE 1: 6 rounds
For all groups the same,
independent of treatment
Threshold
• climate event has happened
Ambiguity
• long-lasting climate event might
happen, unknown probability
Risk
• long-lasting climate event might
happen, known probability (0.5)
Baseline
STAGE 2: 10 rounds
All groups start again
with 50 resource
stock units
Dynamic common pool resource game
Start
12. STAGE 1: 6 rounds
For all groups the same,
independent of treatment
Baseline
• no change
Threshold
• climate event has happened
Ambiguity
• long-lasting climate event might
happen, unknown probability
Risk
• long-lasting climate event might
happen, known probability (0.5)
Baseline
STAGE 2: 10 rounds
All groups start again
with 50 resource
stock units
Dynamic common pool resource game
Start
14. Result 1: there is a treatment effect at group level
Groups in the threshold, risk
and uncertainty treatment
achieved a higher average
stock size over time.
Pairwise Wilcox text
(Bonferroni adjusted)
Base line Risk Threshold
Risk 0.00160 - -
Threshold 0.00016 1.00000 -
Uncertainty 8.1e-11 0.01782 0.00041
15. Fishermen facing thresholds presented a more conservative
behaviour on the exploration of the parameter space of
resource exploitation.
17. Result 2: Treatment effect at individual level
• Threshold:
• Base line = 20
• Treatments = 28
• Cooperation =
(StockSize - Threshold / 4) - IndExtraction
• Are there fisher
styles?
Buenavista Las Flores Taganga
Las Flores 2.6e-11 - -
Taganga 0.00038 < 2e-16 -
Tasajera 0.02236 < 2e-16 0.10150
Pairwise Wilcox text
(Bonferroni adjusted)
18. Result 2: Treatment effect at individual level
• Threshold:
• Base line = 20
• Treatments = 28
• Cooperation =
(StockSize - Threshold / 4) - IndExtraction
• Are there fisher
styles?
Buenavista Las Flores Taganga
Las Flores 2.6e-11 - -
Taganga 0.00038 < 2e-16 -
Tasajera 0.02236 < 2e-16 0.10150
Pairwise Wilcox text
(Bonferroni adjusted)
19. Random effects panel model
shows that treatment, place and
time have significant effects on
cooperation, but effect differs if at
group or individual levels.
Taganga
20. What’s going on in these areas?
Lower education and
older population
24. Concluding remarks
• Fishermen facing (uncertain) thresholds
exhibit a more precautionary behaviour
as a group
• Speculation collapse:
• groups come often from communities
with higher livelihood diversification,
lower resource dependence
• and strongly exposed to
infrastructure development
• Higher noise on resource dynamics
masks the relationship between
fishing efforts and stock size
encouraging a more explorative
behaviour of fishing effort in real life.
26. Experimental design of dynamic CPR game:
communication of risk/ uncertainty
Risk Uncertainty
+ 8 of unknown colour
(Constant) probability of 0.5
that the event happens in
following round
(Constant) probability range between 1/10
and 9/10 that the event happens in following
round