Balanced Social Learning

Andrew Koh; Ricky Li

Home/Archive/arXiv

arXiv · 2610.25814

Balanced Social Learning

Andrew Koh·Ricky Li

View original source ↗Read source document ↗Export metadata ↓

Abstract

When is social learning efficient if agents have heterogeneous preferences over a set of actions? When myopic agents generate information only about the action they choose, efficiency requires balance such that all high-value actions are discovered. We characterize the value of policies that control the flow of past information to future agents. Under full transparency of past information, the probability of efficient outcomes is approximately one if and only if the prior and preferences are balanced and the log-likelihood ratio distribution satisfies a set of intuitive moment conditions. Noisy bad news is a canonical example. By contrast, approximate efficiency is obtainable on the set of arbitrary adaptive information policies if and only if the prior alone is balanced. Moreover, it suffices to use adaptive evidence policies which permanently and fully disclose subsets of past information. When the prior is imbalanced and approximate efficiency is unattainable, we characterize the value of adaptive information and show that it is uniformly approximated by the value of adaptive evidence as the impact of each individual signal on beliefs vanishes.

Explore related subjects

econ.TH

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Andrew Koh, Ricky Li. 2026-10-19. Balanced Social Learning. https://arxiv.org/abs/2610.25814

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Rationally Inattentive Echo Chambers

Rationally inattentive players allocate limited attention capacities to biased primary sources and to other players as secondary sources to acquire information about an uncertain state. The resulting Poisson attention network stochastically transmits information from primary sources to a recipient either directly or indirectly through other players. We give conditions for the rise of echo-chamber equilibria in which players restrict attention to their own-biased primary source and same-type peers. We characterize the peer attention networks within echo chambers and analyze their comparative statics. Our results explain why modern information environments foster echo-chamber formation, how small differences in attention capacities can magnify into large disparities in allocations, and why regulatory interventions such as altering user visibility on social media or mandating exposure to opposing views may entail unintended consequences.

econ.TH↗

arXiv2026

Trading Frictions in Dynamic Cap-and-Trade Markets

Using 2.8 million EU ETS transactions and compliance records, 2005-2022, we document three facts about the world's largest carbon market: each year, 37% of regulated operators do not trade, and non-trading is persistent. Purchases concentrate in the April surrender month, when returns are systematically high. Operator order flow predicts returns, concentrated among frequent traders. A model with costly market access, limited intermediation, and heterogeneous information rationalizes all three. Because access is endogenous, the frictions interact non-additively, shaping how efficiently the market corrects the externality. Staggering deadlines across four groups lowers the surrender-month premium by 58% without changing the cap.

econ.TH↗

arXiv2026

Effort without Evidence

I study how paternalistic concern for descendants' success affects learning about the need for costly effort. Each generation can use a demanding method or experiment with a less demanding alternative whose effectiveness is uncertain. Parents who underweight descendants' effort costs undervalue the savings that discovery would provide. On an explicit region of priors and preferences, paternalism removes a materially valuable experiment even though the relevant private record is truthfully transmitted. Relative to an aligned equilibrium attaining the material full-record optimum, the paternalistic equilibrium has greater long-run effort and observed success, lower discounted material welfare, and a less informative terminal public signal. Each sender obtains its preferred full-record continuation, so the comparison survives communication deviations within the same physical environment. I also study concealment as an incentive to experiment. A continuous-effort extension establishes that concealment and finite evidence production can coexist in an infinite-horizon equilibrium. Its welfare gain over babbling is distinct from the loss relative to alignment in the main result.

econ.TH↗