Definition
Expected utility theory is the dominant framework in economics for decisions under uncertainty — the idea that a rational decision-maker weighs each possible outcome by both its probability and its perceived value, and chooses the option with the highest weighted total.
Why it matters
Expected utility theory has been the dominant analytical framework for decisions under uncertainty since von Neumann and Morgenstern in 1944. Its structural form — taking an expected value across the population of possible outcomes — embeds the implicit assumption that the population average is what an individual should care about, which is the assumption ergodicity economics scrutinizes for non-ergodic problems. Naming the framework explicitly is what makes the structural critique available rather than implicit.
How it works
An expected-utility maximizer evaluates a choice by listing each possible outcome, assigning each outcome a utility score that captures how much the decision-maker values it, weighting each utility score by the probability of the outcome, and summing the weighted scores. The utility scoring captures risk aversion: a risk-averse decision-maker scores certain outcomes higher than actuarially equivalent gambles by curving the utility scale so that gaining a dollar from a low base counts for more than gaining a dollar from a high base. The weighted-sum operation — the expectation — is an ensemble average over the population of possible outcomes; the individual is implicitly treated as facing a population of parallel selves and averaging the result across them. In ergodic systems, the expected-utility evaluation also describes what a single agent experiences over time. In non-ergodic systems it does not, and the standard expected-utility framework can recommend choices that maximize the cross-sectional average while underperforming the long-run outcome for the individual making the choice.
In practice
For an individual evaluating a financial decision through expected-utility advice — a Monte Carlo retirement projection, a portfolio optimization, an insurance purchase analysis — the framework's recommendations rest on the implicit assumption that the population average is what the individual should care about. The practical move is to ask whether the variable being averaged is ergodic for the individual's time horizon; if not, the expected-utility recommendation can systematically diverge from what would maximize the individual's long-run outcome. This is not a criticism of expected utility as a logical framework — it is internally consistent — but a recognition that the framework's standard application to single-realization problems imports an ergodicity assumption that may not hold.
In the Longevity Standard Framework
Expected utility theory is the core analytical framework in ergodicity economics that the Peters program critiques, and is the framework whose ensemble-average orientation the Longevity Standard framework deliberately moves away from in favor of an individual-path treatment. The cost-of-income framework operates on the individual's specific arrangement and planning horizon and produces realized value as a per-individual metric rather than an expected-utility figure aggregated over possible outcomes.
Related terms
- Log utility
- Kelly criterion
- Ergodicity
- Non-ergodic system
- Wealth trajectory
- Path dependency
- St. Petersburg paradox
- Ensemble average