# Deciding with Judgment

## TL;DR

This paper introduces a statistical decision rule based on confidence intervals that balances judgmental decisions with statistical risk, applicable to asset allocation, and interprets confidence levels as risk aversion.

## Contribution

It proposes a novel decision rule that combines judgmental and statistical approaches, linking confidence levels to risk aversion and demonstrating its application in asset allocation.

## Key findings

- The decision rule is admissible and guarantees performance not worse than judgmental decisions.
- Confidence levels can be elicited through laboratory experiments involving urns.
- Application to asset allocation shows practical relevance of the decision rule.

## Abstract

A decision maker starts from a judgmental decision and moves to the closest boundary of the confidence interval. This statistical decision rule is admissible and does not perform worse than the judgmental decision with a probability equal to the confidence level, which is interpreted as a coefficient of statistical risk aversion. The confidence level is related to the decision maker's aversion to uncertainty and can be elicited with laboratory experiments using urns a la Ellsberg. The decision rule is applied to a problem of asset allocation for an investor whose judgmental decision is to keep all her wealth in cash.