Working Papers in Economics,
University of Bergen, Department of Economics

Abstract: I introduce a new method to incentivise the elicitation of belief distributions, the Random Partitions method. With this method, an agent’s payoff not only depends on the realised state and the elicited distribution, but also on a randomised two-level partitioning of the state-space. The method creates a binary lottery payoff structure where reports closer to an agent’s true belief distribution generate a higher probability to earn a high payout. The randomisation of the state-space partitioning ensures that the agent is incentivised to report correctly across the entire belief distribution. I compare the introduced Random Partitions method with both the well known Quadratic Scoring Rule, and a method based on the Becker-DeGroot-Marschak procedure and argue that the Random Partitions method gives substantially stronger truth-telling incentives to agents in situations where there are many states/bins.

Keywords: Belief elicitation; randomized state/event Space; partitioning; proper scoring rules.

JEL-codes: D82; D83

23 pages, January 6, 2020

Full text files

WP%2001-20.pdf  Full text

Download statistics

• University of Bergen, Department of Economics
• Home page for this series

Questions (including download problems) about the papers in this series should be directed to Kjell Erik Lommerud ()
Report other problems with accessing this service to Sune Karlsson ().

RePEc:hhs:bergec:2020_001This page generated on 2024-10-27 22:37:47.