MathIsimple
← Utility Theory Hub
Uncertainty & Expected Utility

Utility Under Uncertainty & Expected Utility Theory

Move beyond certainty by modeling lotteries, risk, and ambiguity. Expected utility theory converts probabilistic outcomes into comparable certainty equivalents, powering risk management and asset pricing in modern finance.

Motivating Paradoxes

Paradox

St. Petersburg Paradox

Demonstrates the inadequacy of maximizing expected monetary value, motivating utility-based evaluation of lotteries.

Explains why investors demand finite risk premiums despite potentially unbounded payoffs.
Paradox

Bernoulli's Resolution

Introduces diminishing marginal utility of wealth via logarithmic utility, creating finite expected utility values.

Connects risk aversion to asset pricing and insurance demand.
Paradox

Cramér's Square-Root Utility

Illustrates alternative risk-averse functional forms aligned with observed decision behavior.

Supports quadratic approximations in mean-variance analysis.

Bernoulli Expected Utility Resolution

The St. Petersburg lottery has infinite expected payoff but finite expected utility under logarithmic preferences:

E[U(W)]=n=112nalog(W0+2n1) \mathbb{E}[U(W)] = \sum_{n=1}^{\infty} \frac{1}{2^n} a \log\left(W_0 + 2^{n-1}\right)

Concavity of U ensures diminishing marginal utility, rationalizing investors' finite willingness to pay.

Von Neumann–Morgenstern Axioms

  • Completeness

    Decision-makers can compare any two lotteries.

  • Transitivity

    Consistent rankings across lotteries prevent preference cycles.

  • Continuity

    Intermediate lotteries exist between preferred and less-preferred options.

  • Independence

    Preference between lotteries is unaffected by mixing with a third lottery proportionally.

  • Monotonicity

    Higher probability of preferred outcomes enhances lottery attractiveness.

When these axioms hold, a utility function U exists such that lotteries are ranked by expected utility, E[U(X)], providing a consistent numerical representation.
U(p1,,pn;x1,,xn)=i=1npiu(xi) U(p_1,\ldots,p_n; x_1,\ldots,x_n) = \sum_{i=1}^n p_i u(x_i)

Lotteries, Mixtures & Certainty Equivalents

Simple Lottery: A finite set of outcomes with associated probabilities. Example: (\$500, 0.3; \$200, 0.7).

Compound Lottery: Lotteries whose outcomes are themselves lotteries; the reduction of compound lotteries axiom ensures simplification to simple lotteries.

Certainty Equivalent: The guaranteed amount delivering the same utility as a risky lottery. Solved by equating U(CE) = E[U(X)].

Applications in Financial Mathematics

Application

Arrow–Debreu State Pricing

Expected utility justifies pricing contingent claims using state prices, foundational to complete market models and derivative valuation.

Application

Asset Pricing Kernels

Stochastic discount factors emerge from intertemporal expected utility, linking consumption growth to asset returns.

Application

Portfolio Selection

Expected utility maximization yields optimal asset allocations, extending deterministic models to risky environments with multiple states.

Historical Milestones

1738

Daniel Bernoulli proposes logarithmic utility, solving the St. Petersburg paradox.

1944

Von Neumann and Morgenstern publish expected utility axioms in 'Theory of Games and Economic Behavior'.

1953

Arrow and Debreu embed expected utility in general equilibrium models.

1959

Markowitz popularizes mean-variance analysis as an expected utility approximation under quadratic utility or normal returns.

Behavioral Challenges

Empirical evidence reveals deviations from expected utility. The Allais paradox exposes violations of independence, while the Ellsberg paradox highlights ambiguity aversion. These findings inspire alternative frameworks such as prospect theory, rank-dependent utilities, and ambiguity-sensitive models.

Pr(θjx)=Pr(xθj)Pr(θj)kPr(xθk)Pr(θk) \Pr(\theta_j | x) = \frac{\Pr(x | \theta_j) \Pr(\theta_j)}{\sum_k \Pr(x | \theta_k) \Pr(\theta_k)}

Bayesian updating offers a normative benchmark for belief revision—useful when contrasting with ambiguity-averse behavior.

Research Extensions

Risk Management Insights

Expected utility underpins insurance pricing, risk limits, and derivative hedging strategies. Investors compute certainty equivalents and risk premiums to determine whether an uncertain project dominates a safe alternative, aligning decisions with firm risk appetite.

Measure Risk Preferences

Q&A: Expected Utility in Practice

Why does the independence axiom matter for finance?

Independence ensures investor preferences are linear in probabilities, enabling us to price securities by weighting state payoffs with probabilities and marginal utilities—core to expected utility and risk-neutral valuation.

How do Allais and Ellsberg paradoxes challenge expected utility?

The Allais paradox violates independence by showing real-world preference reversals, while Ellsberg demonstrates ambiguity aversion. Both motivate alternative models like prospect theory and ambiguity-sensitive utilities.

When is mean-variance analysis consistent with expected utility?

When returns are jointly normally distributed or investors have quadratic utility, mean-variance optimization mirrors expected utility maximization, providing computational simplicity in practice.

Keep Building Insight
Progress to risk measurement or reinforce stochastic logic with focused practice.