Quantile Time-Frequency Higher-Moment Risk Spillovers and Multilayer Networks: Bridging Climate Risks to Carbon, Energy, and Metals Markets

Technical Summary: Quantile Time-Frequency Higher-Moment Risk Spillovers and Multilayer Networks

Problem Statement
Existing literature has extensively explored the spillover effects of climate risks (physical and transition) on the returns and volatility of carbon, energy, and metal markets. However, a significant gap remains in the systematic comprehension of these interactions when considering higher-moment risks (skewness and kurtosis), which capture asymmetry, tail risks, and extreme volatility. Furthermore, traditional single-layer network models fail to fully capture the heterogeneity of information transmission across different market conditions (bull, bear, normal) and time-frequency domains. This study addresses these gaps by investigating the quantile time-frequency connections among climate physical risk (PRI), climate transition risk (TRI), and the return, volatility, skewness, and kurtosis risks of the carbon, energy, and metal markets.

Methodology
The study employs a multi-stage methodological framework:

1. Higher-Moment Risk Measurement: The Generalized Autoregressive Conditional Heteroskedasticity with Skewness and Kurtosis (GJRSK) model is utilized to quantify the conditional volatility, skewness, and kurtosis of the target markets.
2. Quantile Time-Frequency Connectedness: The analysis integrates the Quantile Vector Autoregressive (QVAR) model with the Baruník and Kehlík (BK) frequency connectedness method. This allows for the examination of spillover effects across different quantiles (0.1, 0.5, 0.9) and frequency bands (short-term: 1–22 days; long-term: >22 days).
3. Multilayer Network Construction: A 12-layer information network is constructed, encompassing return, volatility, skewness, and kurtosis layers under three market states (extreme downside, normal, extreme upside).
4. Network Metrics:
   • System-level: Global Efficiency (GE) and Average Connectedness Strength (ACS) measure information exchange efficiency and risk intensity. Inter-layer correlations are assessed via Edge Overlapping, Deg-deg Pearson/Spearman coefficients, and Shortest Path Similarity.
   • Market-level: Average overlapping in-strength, out-strength, and net-strength characterize the connectivity of individual nodes.
5. Data: The sample covers daily data from July 1, 2015, to December 29, 2023, comprising PRI and TRI indices, ICE-EUA carbon futures, Brent crude oil, natural gas, the S&P Global Clean Energy Index, and futures for eight metals (Gold, Silver, Copper, Aluminum, Zinc, Lead, Nickel, Tin).

Key Contributions

• Holistic Systemic Analysis: Unlike previous studies that examined climate risks and specific markets in isolation, this research integrates climate risks with a unified carbon-energy-metal system, including both traditional and clean energy, as well as precious and non-ferrous metals.
• Methodological Integration: The study combines higher-moment modeling with quantile time-frequency analysis. This dual approach enables the identification of short-term and long-term spillover systems under extreme negative and positive impacts, providing insights into the intertemporal dependence structure of tails.
• Multilayer Network Framework: By constructing a 12-layer network, the study moves beyond single-layer return/volatility analyses. It simultaneously presents multiple risk metrics (return, volatility, skewness, kurtosis) across varying market states, offering a more granular view of complex financial system interconnectivity.

Key Results

• Significance of Higher-Moments: While the magnitude of higher-moment (skewness and kurtosis) spillovers is lower than that of returns and volatility, they are statistically significant and exhibit time-varying characteristics.
• Time-Frequency Dynamics:
  • Return, skewness, and kurtosis spillovers are primarily driven by short-term dynamics (high-frequency).
  • Volatility spillovers are concentrated largely in the long run (low-frequency).
  • Correlations between climate risks and markets are significantly stronger in extreme market scenarios compared to normal conditions.
• Role Reversals and Net Spillovers:
  • Metal Markets: Generally occupy a leading position as information transmitters (spillover senders) in the system, particularly for lower-moment risks. However, their role shifts to being net receivers during extreme market downturns for higher-moment risks.
  • Climate Risks: Under normal conditions, PRI and TRI act primarily as spillover receivers. However, during market downturns (bear states), they shift to become net transmitters of long-term risks.
  • Extreme Events: Events like the COVID-19 pandemic and the Russia-Ukraine conflict trigger rapid intensification of spillover effects and can reverse the direction of net spillovers (e.g., carbon markets becoming net contributors during the pandemic).
• Network Topology:
  • The contagion speed of higher-moment risks is faster than that of lower-moment risks.
  • In upward (bull) market states, information exchange between climate risks and markets slows down, despite tighter connections.
  • Structural configurations differ significantly across layers; for instance, the kurtosis network at the 0.9 quantile shares fewer edges with other layers, indicating low structural similarity.

Significance and Claims
The paper claims that its findings provide a more comprehensive understanding of the linkage mechanisms between climate risks and financial markets. By demonstrating that risk contagion networks based on higher-moment risks offer more favorable conditions for contagion than those based solely on lower moments, the study argues that existing literature focusing only on returns and volatility cannot fully reflect actual risk transmission dynamics.

The results offer specific value for:

• Investors: By highlighting the need for a "multi-moment + multi-scenario" risk management framework and identifying metal markets (e.g., copper, nickel) as key risk early warning indicators.
• Policymakers: By revealing the asymmetry of spillover effects across market states, suggesting the need for differentiated strategies (e.g., monitoring oil volatility in bull markets and preventing climate risk transmission in bear markets).
• Regulators: By emphasizing the necessity of cross-market collaborative supervision and the establishment of real-time monitoring platforms for skewness and kurtosis risks to address current supervision blind spots.