Risk spillover between financial markets based on non-parametric quantile Vine-Copula model

Mengwei Wang, Junxiang Lu, Yu Zheng, Weiqin Zhu, Chunyu XieDEVELOPMENT 2025 · DOI

TL;DR

This paper investigates the complex risk spillover dynamics among BRICS financial markets using a novel non-parametric quantile D-Vine Copula model. The findings reveal significant risk contagion, particularly from China to other BRICS nations, with Russia being the most affected, offering crucial insights for global financial risk management.

What Problem Does This Paper Address?

The stability of emerging economies' financial markets, particularly the BRICS nations (China, Brazil, Russia, India, South Africa), is critical for global economic health. Understanding how risks propagate between these markets is essential for comprehending the global financial system and mitigating contagion. Previous studies often relied on parametric models or simpler correlation measures, which may not fully capture the intricate, non-linear, and asymmetric dependencies present in complex financial systems, especially during extreme events. This paper aims to address this gap by providing a more robust and nuanced analysis of risk spillovers within the BRICS context.

Methods

The authors employ a multi-stage approach:

  1. Non-parametric Quantile Regression for Edge Distribution: To accurately model the marginal distributions of stock returns for each BRICS country, they use a non-parametric quantile regression method. This approach is favored over parametric methods because it doesn't assume a specific distribution for the data, allowing it to better capture complex data structures and the impact of extreme events. Kernel density estimation is used for smoothing.
  2. D-Vine Copula Model: To capture the complex, multi-dimensional dependencies between the financial markets, a D-Vine Copula model is constructed. Vine Copulas are powerful tools for modeling high-dimensional dependencies by breaking them down into a series of bivariate (pair-copula) relationships. The D-Vine structure is chosen for its ability to represent asymmetric relationships and its intuitive exploration of country-specific correlations.
  3. Generalized Multidimensional Conditional Tail Expectation (CoES): To quantify the intensity of risk spillovers, the paper utilizes the CoES metric. CoES is an extension of CoVaR (Conditional Value at Risk) that focuses on tail risk and offers a more robust measure of extreme risk contagion, especially in multidimensional settings. It measures the expected additional loss in one market given that other markets are experiencing extreme losses.

Datasets

The study uses daily stock market data for the BRICS countries:

  • China: Shanghai Index (000001.SH)
  • Russia: MOEX index (IMOEX.MCX)
  • India: SENSEX Index (SENSEX.bo)
  • South Africa: Johannesburg All Share Index (JALSH.GI)
  • Brazil: Bovespa index (IBOVESPA.GI)

The dataset spans from April 20, 2020, to April 19, 2024, encompassing the COVID-19 period. Data was sourced from the Wind database. The authors ensured synchronization by selecting periods when each country's stock market was open.

Key Insights and Ideas

  • Beyond Simple Correlations: The paper highlights the inadequacy of simple correlation coefficients for understanding financial market interconnectedness, especially during turbulent times.
  • Non-parametric Advantage: The use of non-parametric quantile regression is crucial for accurately modeling the often skewed and fat-tailed distributions of financial returns, particularly when extreme events are of interest.
  • D-Vine for Complexity: The D-Vine Copula model is a key innovation, allowing for a more granular and flexible representation of how multiple financial markets influence each other, moving beyond simpler bivariate or parametric multivariate models.
  • CoES for Tail Risk: Employing CoES provides a more precise measure of risk contagion, focusing on the impact of extreme market movements rather than just average behavior.

Key Results

The empirical analysis reveals significant risk spillover effects among BRICS financial markets:

  • China's Dominance: China's financial market exhibits a significant risk spillover effect on other BRICS countries.
  • Russia's Vulnerability: Russia is the most prominently affected by risk spillovers from other BRICS markets.
  • India and South Africa: These countries are also significantly impacted, though to a lesser extent than Russia.
  • Brazil's Asymmetry: The impact on Brazil's market is relatively weak and shows obvious asymmetry.
  • Specific Example: The paper notes that when other BRICS markets experience extreme risk (at a 0.05 confidence level), the Chinese stock market may face an average additional loss. The positive and large value for Russia in the table suggests that when other countries face extreme market conditions, the Russian market tends to decline, indicating instability. This is speculatively linked to geopolitical events like Russia's military action in Ukraine.

Limitations and Future Work

  • Data Limitations: While the paper uses a comprehensive dataset, the authors acknowledge that the analysis is based on stock market data, and other financial instruments (e.g., bonds, commodities) could offer a more complete picture.
  • Model Complexity: The D-Vine Copula model, while powerful, can become computationally intensive with a very large number of variables.
  • Future Directions: The authors suggest that their findings have important theoretical and practical implications for monitoring risk contagion and developing response strategies. Future work could explore the impact of other financial assets, extend the analysis to a broader set of emerging markets, and investigate the effectiveness of specific policy interventions in mitigating risk spillovers.

Summary updated

Abstract

Brics (China, Brazil, Russia, India, South Africa) as the world's emerging economies, the stability of their financial markets has an important impact on the global economy. This paper intends to study the correlation structure and risk contagion effect among the financial markets of the BRICS countries. Based on the stock market, an important indicator of the financial markets of each country, the paper uses the non-parametric quantile model to fit the marginal distribution of the returns of each country, constructs the D-Vine Copula model, explores the correlation structure among the financial markets of each country, and uses CoES to describe the risk spillover intensity. The empirical results show that China's financial market has significant risk spillover effect on other BRICS countries, and the impact on Russia is the most prominent, followed by India and South Africa, and the impact on Brazil is relatively weak and presents obvious asymmetry. These results have important theoretical value and practical significance for monitoring the risk contagion effect in BRICS countries.
BibTeX
@inproceedings{wang2025risk,
  author = {Mengwei Wang and Junxiang Lu and Yu Zheng and Weiqin Zhu and Chunyu Xie},
  title = {{Risk spillover between financial markets based on non-parametric quantile Vine-Copula model}},
  booktitle = {DEVELOPMENT},
  year = {2025},
  doi = {10.47297/wspdecwsp2515-797318.20250901},
}
Figure