Joint-outcome prediction markets for climate risks

RESEARCH ARTICLE Joint-outcome prediction markets for climate risks Mark S. RoulstonID1,2, Kim Kaivanto2,3* 1 Economics and Policy Institute, Land, Environment, University of Exeter, Exeter, United Kingdom, 2 CRUCIAL: Climate Risk and Uncertainty Collective Intelligence Aggregation Laboratory, United Kingdom, 3 Lancaster University Management School, Lancaster University, Lancaster, United Kingdom * k.kaivanto@lancaster.ac.uk Abstract Predicting future climate requires the integration of knowledge and expertise from a wide range of disciplines. Predictions must account for climate-change mitigation policies which may depend on climate predictions. This interdependency, or “circularity”, means that cli- mate predictions must be conditioned on emissions of greenhouse gases (GHGs). Long- range forecasts also suffer from information asymmetry because users cannot use track records to judge the skill of providers. The problems of aggregation, circularity, and informa- tion asymmetry can be addressed using prediction markets with joint-outcome spaces, allowing simultaneous forecasts of GHG concentrations and temperature. The viability of prediction markets with highly granular, joint-outcome spaces was tested with markets for monthly UK rainfall and temperature. The experiments demonstrate these markets can aggregate the judgments of experts with relevant expertise, and suggest similarly structured markets, with longer horizons, could provide a mechanism to produce credible forecasts of climate-related risks for policy making, planning, and risk disclosure. Introduction

Forecasting climate change is an important and multi-faceted prediction problem. Predicting climate given concentrations of greenhouse gases (GHGs) requires combining expertise in meteorology, oceanography, and other physical sciences; predicting what concentrations of GHGs might be in the future extends the problem into the realms of demographics, econom- ics, policy, and technological innovation. The cornerstone of climate forecasting is coupled general circulation models (CGCMs). These models simulate the dynamics and thermodynamics of the atmosphere and ocean. They are relatives of the numerical weather prediction models used to produce short to medium range (1 to 14 day ahead) weather forecasts. The main differences are that climate models include a dynamical ocean component, and they are run at a lower resolution because they are used to produce simulations lasting decades rather than weeks. State-of-the-art CGCMs, such as the U.K. Met Office’s HadGEM3 models, have horizontal resolutions of approximately 60 km [1]. CGCMs require large amounts of computing power and historically have been the pre- serve of major institutions and government agencies. The Coupled Model Intercomparison PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone.0309164 August 30, 2024 1 / 14 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Roulston MS, Kaivanto K (2024) Joint- outcome prediction markets for climate risks. PLoS ONE 19(8): e0309164. https://doi.org/10.1371/ journal.pone.0309164 Editor: Robin Haunschild, Max Planck Institute for

Solid State Research, GERMANY Received: March 22, 2023 Accepted: August 1, 2024 Published: August 30, 2024 Peer Review History: PLOS recognizes the benefits of transparency in the peer review process; therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pone.0309164 Copyright: © 2024 Roulston, Kaivanto. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: https://dataverse. harvard.edu/dataset.xhtml?persistentId=doi:10. 7910/DVN/797RXY. Funding: The experimental markets were run while MR was employed by Winton Group. Winton also provided the cash incentives for institutions