The risk elicitation puzzle revisited: Across-methods (in)consistency?

Area of Science:

• Social Sciences
• Behavioral Economics
• Experimental Psychology

Background:

• Laboratory methods for assessing risk attitudes are common in social sciences.
• Existing evidence indicates that risk preference measurements can differ across elicitation tasks.
• This variability raises questions about the consistency of revealed risk preferences.

Purpose of the Study:

• To investigate how different risk preference elicitation tasks affect individual and aggregate risk attitude estimates.
• To compare the heterogeneity of risk preference estimates across methods with simulated random variation.
• To examine subjects' awareness of the riskiness of their choices within experimental settings.

Main Methods:

• A within-subject experimental design was employed.
• Four widespread risk preference elicitation tasks were utilized.
• Simulation exercises were conducted to establish behavioral benchmarks.

Main Results:

• Different methods produced significantly varying estimates of risk preferences at both individual and aggregate levels.
• The observed heterogeneity across methods was comparable to random variation from choice distributions.
• Subjects demonstrated a notable awareness of the varying riskiness of their decisions.

Conclusions:

• The substantial variation in risk preference estimates across methods is a key finding.
• Subjects' awareness of choice riskiness suggests their behavior is not necessarily inconsistent.
• This challenges the conventional interpretation of fluctuating revealed risk preferences as indicative of instability.