Decreasing Marginal Value of Information Under Symmetric Loss

Authors:
Sushil Bikhchandani;John W. Mamer
Affiliations:
Anderson School of Management, University of California, Los Angeles, Los Angeles, California 90095;Anderson School of Management, University of California, Los Angeles, Los Angeles, California 90095
Venue:
Decision Analysis
Year:
2013

Citing 4
Cited 1

Optimal Sequential Exploration: A Binary Learning Model

Decision Analysis
Some Comparisons among Quadratic, Spherical, and Logarithmic Scoring Rules

Decision Analysis
The Value of Information and Intensity of Preference

Decision Analysis
Paradoxes in Learning and the Marginal Value of Information

Decision Analysis

From the Editor---Optimal Betting, Reducing Unnecessary Mammography in Breast Cancer Diagnosis, Product Line Design, and Value of Information

Decision Analysis

Quantified Score

Hi-index	0.00

Visualization

Abstract

We investigate conditions under which the value of additional information is diminishing in a setting where the decision maker has access to multiple, identically-distributed, information signals. The signals are assumed to be independent conditional on an unknown payoff-relevant parameter. The decision maker minimizes a quadratic loss function. Quadratic losses arise in quality control, scoring rules, and other applications. We characterize two concepts of diminishing marginal value of information. The first is an ex ante concept, before any information is observed, and the second is an ex post concept, after observation of previous information signals. The former concept is useful for ex ante information acquisition decisions and the latter for sequential information acquisition.