Risk Attitudes & Arrow-Pratt Measures

Risk attitudes describe how people feel about uncertainty—some avoid it, some seek it out, and some don't care either way. Think of it like this: if someone offers you a guaranteed $50 or a coin flip for $100, your choice reveals your risk attitude. The Arrow-Pratt measure is a mathematical tool economists use to quantify this intuition by measuring how sharply the value of money changes for a person (the curvature of their utility function). The more sharply that value curves downward, the more risk-averse someone is. This matters because it explains why some people buy insurance while others gamble their savings on startups. Try this: When you face an uncertain choice, ask yourself what guaranteed amount would make you just as happy—that's your personal measure of risk attitude.

How Risk Attitudes Work

Risk attitudes fall into three categories. Risk-averse people prefer certainty over uncertainty with the same average payoff. Risk-neutral people only care about the average outcome, so they're indifferent between the two. Risk-seeking people prefer the gamble even when it has the same or lower average value.

The Arrow-Pratt Measure Explained

The Arrow-Pratt measure quantifies risk attitude by measuring how sharply a person's satisfaction (utility) curves as they gain more money. Imagine a graph where the x-axis is money and the y-axis is how much that money means to you. Risk-averse people have a curve that flattens quickly—each extra dollar matters less than the previous one. The Arrow-Pratt measure captures this curvature mathematically. A higher value means more risk aversion, zero means risk-neutral, and negative values mean risk-seeking.

Real-World Examples

Insurance: A risk-averse person buys home insurance even though the premium costs more than the expected loss. They're paying to eliminate uncertainty.

Investing: The Arrow-Pratt framework helps explain why individuals with different wealth levels or time horizons may tolerate different levels of risk.

Career choices: Some people take stable government jobs with lower pay (risk-averse), while others join startups with equity that could be worth millions or nothing (risk-seeking).

Why This Matters

Understanding your own risk attitude helps you make consistent decisions. If you're naturally risk-averse, you might overpay for safety. If you're risk-seeking, you might take dangerous gambles without realizing it. The Arrow-Pratt framework shows that there's no 'right' risk attitude—different levels make sense for different situations and wealth levels.

The Arrow-Pratt measure of risk aversion, developed independently by Kenneth Arrow (1965) and John Pratt (1964), provides a local measure of risk attitude based on the curvature of the utility function. The absolute risk aversion coefficient (ARA) is defined as the ratio of the second derivative to the first derivative of utility, capturing how marginal utility changes with wealth. This measure is invariant under positive affine transformations of utility, making it a robust tool for comparing risk attitudes across individuals [1].

Key research findings:

• Arrow (1965) introduced the measure in his Yrjö Jahnsson Lectures, showing that the curvature of utility functions directly corresponds to risk attitudes and that decreasing absolute risk aversion is a plausible empirical assumption [1].
• Pratt (1964) independently derived equivalent measures, demonstrating that the local risk aversion coefficient determines the risk premium an individual would pay to avoid a small risk [2].
• The measure also connects to observable behavior through the concept of decreasing absolute risk aversion (DARA) and constant relative risk aversion (CRRA), which are widely used in macroeconomic and financial models. [3]

The Arrow-Pratt framework has several important limitations. First, it assumes people have stable, consistent risk attitudes across all situations, but behavioral research shows risk preferences can depend on framing, context, and even mood. Second, it treats risk as a single dimension, ignoring how people react differently to known probabilities versus deep uncertainty. Third, prospect theory (Kahneman & Tversky, 1979) shows people evaluate gains and losses asymmetrically—they're more sensitive to losses than equivalent gains—which the standard utility model doesn't capture. Finally, measuring real-world risk attitudes requires inferring utility functions from observed choices, which is difficult and often yields inconsistent results.