Utility functions predict variance and skewness risk preferences in monkeys

Area of Science:

• Behavioral economics
• Neuroscience
• Decision-making under risk

Background:

• Utility functions are central to understanding economic choices and risk preferences.
• Formal definitions of risk, including variance-risk and skewness-risk, are crucial for validating utility theory.
• Previous research primarily focused on variance-risk, leaving other risk dimensions less explored.

Purpose of the Study:

• To investigate whether empirically derived utility functions in macaques can predict preferences for both variance-risk and skewness-risk.
• To formally define and test preferences for different types of risk in non-human primates.
• To assess the conformity of macaque choices to internal reward valuations.

Main Methods:

• Calculated expected utilities for symmetrical and skewed gambles to define stochastic dominance.
• Conducted direct choice experiments with macaques involving gambles with varying variance and skewness.
• Analyzed preferences based on second-order (variance) and third-order (skewness) stochastic dominance.

Main Results:

• Macaque preferences aligned with both variance-risk and skewness-risk predictions.
• Monkeys exhibited preferences for high-variance gambles at low expected values (EVs) and low-variance gambles at high EVs.
• Animals consistently preferred positively skewed gambles over symmetrical and negatively skewed ones, demonstrating strong transitivity.

Conclusions:

• Empirically derived utility functions accurately predict macaque preferences for variance-risk and skewness-risk.
• Macaque decision-making aligns with internal reward valuations that incorporate both variance and skewness.
• This study supports a representation of utility in monkeys that accounts for multiple dimensions of risk.