Paper: arXiv 2604.25406

Authors: Ying-Hui Shao, Yan-Hong Yang, Yun Zhang

Abstract

Connectedness measures quantify aggregate risk spillovers but obscure the local interaction patterns that generate systemic risk. We develop a motif-based framework that first extracts multiscale backbones from quantile connectedness networks and then identifies directed triadic motifs whose frequencies exceed randomization baselines. To distinguish how assets’ sectoral identities shape local spillover structures, we introduce colored motifs under sector partitions of increasing granularity. Using orbit positions that capture each node’s structural role within directed triadic motifs, we construct portfolio strategies that exploit an asset’s place in the spillover architecture. Applying the framework to 39 commodity and equity futures across lower, median, and upper conditional quantiles, we find that motif-based portfolios outperform minimum correlation and minimum connectedness benchmarks on risk-adjusted returns. We further show that in tail networks, assets with greater orbit-position diversity tend to act as net spillover transmitters rather than receivers, establishing positional diversity as a tail-specific marker of systemic influence. These findings demonstrate that local triadic topology carries portfolio-relevant information that aggregate connectedness measures miss.

Complexity vs Empirical Score

  • Math Complexity: 7.5/10
  • Empirical Rigor: 6.0/10 (capped from 8.0: transaction costs not modelled; see penalty flags)
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper introduces a novel motif-based network decomposition framework combining quantile connectedness with graph-theoretic triadic motifs and colored network analysis. It delivers strong empirical validation using a 39-asset commodity and equity futures dataset, demonstrating outperformance in portfolio backtests compared to traditional risk benchmarks.

Research Flowchart

  flowchart TD
    A["Research Goal:<br>Extract local risk spillover topology<br>for portfolio optimization"] --> B["Data Inputs:<br>39 Commodity & Equity Futures<br>(Lower, Median, Upper Quantiles)"]
    
    B --> C["Network Extraction & Partitioning:<br>Extract multiscale backbones &<br>apply colored motifs (sector partitions)"]
    
    C --> D["Structural Role Mapping:<br>Identify directed triadic motifs &<br>calculate node orbit positions"]
    
    D --> E["Outcomes:<br>1. Motif portfolios outperform benchmarks<br>2. Orbit diversity signals net spillover transmitters"]