[Paper Review] Invariant Causal Routing for Governing Social Norms in Online Market Economies

Xiangning Yu, Qirui Mi+5 more|arXiv (Cornell University)|Mar 4, 2026

Game Theory and Applications0 citations

TL;DR

The paper introduces Invariant Causal Routing (ICR), a three-stage framework that uses PNS-based causal inference to identify policy–norm routes invariant to distribution shifts, and assembles them into a concise, auditable rule router for stable norm attainment in online markets.

ABSTRACT

Social norms are stable behavioral patterns that emerge endogenously within economic systems through repeated interactions among agents. In online market economies, such norms -- like fair exposure, sustained participation, and balanced reinvestment -- are critical for long-term stability. We aim to understand the causal mechanisms driving these emergent norms and to design principled interventions that can steer them toward desired outcomes. This is challenging because norms arise from countless micro-level interactions that aggregate into macro-level regularities, making causal attribution and policy transferability difficult. To address this, we propose extbf{Invariant Causal Routing (ICR)}, a causal governance framework that identifies policy-norm relations stable across heterogeneous environments. ICR integrates counterfactual reasoning with invariant causal discovery to separate genuine causal effects from spurious correlations and to construct interpretable, auditable policy rules that remain effective under distribution shift. In heterogeneous agent simulations calibrated with real data, ICR yields more stable norms, smaller generalization gaps, and more concise rules than correlation or coverage baselines, demonstrating that causal invariance offers a principled and interpretable foundation for governance.

Motivation & Objective

• Understand how platform policies causally influence the emergence and stabilization of social norms in online markets.
• Identify policy–norm routes that remain invariant under distributional shifts across heterogeneous environments.
• Develop an auditable, interpretable rule-based governance mechanism that generalizes across seeds and contexts.
• Provide causal explanations for why certain interventions succeed or fail across different groups and contexts.

Proposed method

• Stage I uses Probability of Necessity and Sufficiency (PNS) to test if a platform strategy causally enables a group to attain its norm band under a given context.
• Stage II learns a minimal first-match rule router S* that maps contexts to interventions by maximizing causal gains while ensuring broad coverage across initial-condition buckets.
• Stage III attributes key factors by contrasting lever distributions under target versus baseline policies to explain why interventions succeed or fail.
• The framework relies on twin-world paired runs to estimate PNS and a bucketed greedy plus prune procedure to assemble a compact, invariant router.
• Norm attainment is operationalized as macro statistics entering and persisting within a defined tolerance band after a burn-in.

Figure 1 . Social norms in onlineeconomies.

Experimental results

Research questions

• RQ1 RQ1: How do different platform objectives causally influence the emergence and stabilization of social norms across user groups?
• RQ2 RQ2: What is the shortest and most stable invariant causal routing policy that remains effective under distribution shift?
• RQ3 RQ3: What mechanisms explain norm divergence across groups or contexts and which factors are necessary or sufficient for these divergences?

Key findings

• PNS-certified causal routes demonstrate that specific policy switches reliably drive norm attainment across all groups under studied contexts.
• A compact rule router S* generalizes across seeds and initial conditions, achieving strong out-of-distribution robustness.
• Pruning maintains parsimony with negligible loss in causal effectiveness, while correlation-based baselines are more fragile under distribution shift.
• Stage III attribution reveals which platform levers and user responses drive successful norm stabilization under identical initial conditions.
• Invariant causal routing yields smaller generalization gaps and higher robustness than correlation-based or coverage baselines.

Figure 2 . Three-stage framework for discovering causal strategies in social norm formation. Stage I tests causal effects by identifying implicit contracts $(\theta,g\mid\psi)$ where policy $\theta$ enables group $g$ to reach the norm under context $\psi$ . Stage II learns a compact rule router $S^{