Combining Combined Forecasts: a Network Approach

Marcos R. Fernandes

DOI: https://doi.org/10.48550/arXiv.2406.13749

2024-06-20

Abstract:This study investigates the practice of experts aggregating forecasts before informing a decision-maker. The significance of this subject extends to various contexts where experts inform their assessments to a decision-maker following discussions with peers. My findings show that, irrespective of the information structure, aggregation rules introduce no bias to decision-making in expected terms. Nevertheless, the concern revolves around variance. In situations where experts are equally precise, and pair-wise correlation of forecasts is the same across all pairs of experts, the network structure plays a pivotal role in decision-making variance. For classical structures, I show that star networks exhibit the highest variance, contrasting with $d$-regular networks that achieve zero variance, emphasizing their efficiency. Additionally, by employing the Poisson random graph model under the assumptions of a large network size and a small connection probability, the results indicate that both the expected Network Bias and its variance converge to zero as the network size becomes sufficiently large. These insights enhance the understanding of decision-making under different information, network structures and aggregation rules. They enrich the literature on combining forecasts by exploring the effects of prior network communication on decision-making.

Theoretical Economics