The Psychology of Rare Events: Challenges to Managing Tail Risks

Transcription

1 Workshop on Climate Change and Extreme Events: The Psychology of Rare Events: Challenges to Managing Tail Risks Elke U. Weber Center for Research on Environmental Decisions (CRED) Columbia University Resources for the Future, February 3, 2010

2 Need for a psychological analysis Best solutions (e.g., to reform National Flood Insurance Program) only as good as they are implementable Depends on perceptions of and reactions to proposed choice options by the public and politicians Knowledge of human perception of and reaction to rare and adverse events should guide design as well as implementation of policy

3 Value of psychological analysis and tools Complements economic and political analysis Suggests solutions that might have greater individual or social utility

4 Human Cognition and Motivation Limited attention and processing capacity Limited emotional capacity Multiple goals

5 Limited attention and processing capacity Need to attend selectively Guided by expectations (past experience) and goals (fears or wishful thinking) Need to encode and evaluate locally Thurber story Buckets-of-water thought experiment Hardwired automatic emotion- and association-based processes often dominate effortful analytic processes

6 Reactions to rare events Depends on how we find out about them From personal experience Like all other animals, by repeated sampling Only outcomes are experienced, time after time From description Based on vicarious past experience or model predictions Possible outcomes and their likelihood are described numerically or visually (pie chart, frequency distribution)

7 Rare events are Overweighted (or ignored) in decisions from description Captured by decision weight function of Prospect Theory Underweighted in decisions from experience Unless they recently occurred, in which case they are strongly overweighted Captured by reinforcement learning models with strong recency weight Many decisions based on a combination of outcome distribution description and personal experience Personal experience typically the stronger determinant of choice More emotionally engaging, vivid

8 Expert vs. Public Disagreements about Importance of Risk Often explained by differences in how public vs. experts learn about the risk Flood risks, airplane crash risk (flight insurance) Experts by description (actuarial rates), public from experience Vaccination side effect risks Public by description (website, pamphlet), pediatricians from experience

9 Reactions to rare events Also depends on how we feel about them No worry, no action Risk is a feeling (Loewenstein, Weber, Hsee & Welch 2001) Analytic concern neither necessary nor sufficient Psychological risk dimensions Multidimensional scaling analysis of health and safety risks (Slovic et al, in 1970s and 80s) dread risk and unknown risk Risks scoring high on two factors are attended to and receive social amplification (e.g., by media) Flood risk is in lower left psychological risk space: no dread, well known

10 Limited emotional capacity Single action bias Tendency to engage in a single risk reduction or risk management behavior when action is triggered by concern (rather than analysis) Farmers who are concerned about climate change engage in either production, pricing, or policy path to protection, but not all three (Weber, 1999) Finite pool of worry Increases in concern about one risk are accompanied by decreases in another

11 Variation in concern about global warming over time finite pool of worry (Weber, 2006)

12 Human Motivation People s motivation not as simple or predictable as that of rationaleconomic man Outcome goals for self Growth, procreation, and promotion Preservation and prevention Social goals Other-regarding preferences, fairness, future generations Process goals Feel in control Uncertainty and ambiguity are aversive Predict events and understand causes Mental models, science Learning to predict and thus survive is one of main function of our brain Large neural responses to unexpected events, especially when they are perceived as losses Neuroeconomics, Reward prediction error Loss aversion, Prospect Theory

13 Psychological Contributions to Solutions Capitalize on flexible nature of motivation and information processing Multiple goals Social (and other) goals can be primed by prior tasks, choice context (group vs. individual setting) or background Wallpaper on first page of website Voting in church or school Multiple ways to represent information (framing) Prospect theory (Kahneman & Tversky) Query theory (Johnson & Weber)

14 Preferences are constructed, not revealed Construction process can be influenced or nudged (Thaler & Sunstein, 2008) (Re)labeling options Carbon offsets more palatable than carbon taxes to Republicans (Hardisty et al., 2010) Preferences change when defaults change Merits of default option considered first, leading to strong status quo bias (Johnson & Goldstein, 2003)

15 What are the new tools in your bag? If all encoding is relative, then you can change people s reference point ( compared to what? ) Prospect theory Different risk attitudes for gains (averse) and losses (seeking) Loss aversion Explains Equity Premium Puzzle (Thaler, 1990), insufficient diversification of portfolios, etc. Perceptions of change or variation also relative Coefficient of variation (std/ev) a better predictor of risk taking than variance (Weber et al., 2004)

16 Know about effects of providing information about fat tail events in different ways By description or by direct experience (e.g., in simulation games) depending on goal Provides readiness to capitalize on effects of rare events when they occur Hurricane Katrina and other potential game changers in public attitudes and willingness to act

17 Ability to predict people s changes in preference better than they themselves Hedonic forecasting generally poor People underestimate their adaptation to changes in status quo (winning lottery or becoming paralyzed) Observed changes in preference well described by Query Theory (Johnson and Weber) Should increase politicians willingness to help people reach their own longer-term or social goals by fiat or regulation If you change it, they will like it Example of Mayor Bloomberg s smoking ban in bars in 2005

18