[Paper Review] Explosive Volatility: A Model of Financial Contagion

Nicholas G. Polson, James G. Scott|arXiv (Cornell University)|Oct 26, 2011

Financial Risk and Volatility Modeling24 references6 citations

TL;DR

This paper proposes a financial contagion model that explains explosive volatility through time-varying sensitivities of global assets to shocks in aggregate volatility, using a factor model with global and regional risk factors. It shows that high sensitivity to these shocks strongly predicts financial crises and that the evolving correlation between volatility and returns aligns with major historical events.

ABSTRACT

This paper proposes a model of nancial contagion that accounts for explosive, as excess correlation in the residuals from a factor model incorporating global and regional market risk factors. Some of this excess correlation can be explained by quantifying the impact of shocks to aggregate volatility in the cross-section of expected returns|but only, it turns out, if one is extremely careful in accounting for the explosive nature of these shocks. We show that global markets have time-varying cross-sectional sensitivities to these shocks, and that high sensitivities strongly predict periods of nancial crisis. Moreover, the pattern of temporal changes in correlation structure between volatility and returns is readily interpretable in terms of the major events of the periods in question.

Motivation & Objective

• To explain the emergence of explosive volatility in financial markets through a structured model of contagion.
• To account for excess correlation in asset returns beyond what standard factor models predict.
• To investigate how shocks to aggregate volatility drive changes in cross-sectional return correlations.
• To identify time-varying sensitivities of global assets to volatility shocks as predictors of financial crises.
• To interpret shifts in the correlation structure between volatility and returns in light of major financial events.

Proposed method

• The model employs a factor model incorporating global and regional market risk factors to decompose asset returns.
• It isolates residual excess correlation in returns after accounting for these common factors.
• The analysis quantifies the impact of shocks to aggregate volatility on the cross-section of expected returns.
• Time-varying cross-sectional sensitivities of assets to volatility shocks are estimated and monitored over time.
• The model tracks temporal changes in the correlation between volatility shocks and asset returns.
• Historical financial events are used to interpret and validate observed patterns in the correlation dynamics.

Experimental results

Research questions

• RQ1 How do shocks to aggregate volatility contribute to explosive volatility and excess correlation in financial returns?
• RQ2 To what extent can time-varying sensitivities of assets to volatility shocks predict financial crises?
• RQ3 How do the dynamic correlations between volatility and returns align with major historical financial events?
• RQ4 What role do global and regional risk factors play in amplifying or dampening volatility contagion?
• RQ5 Why is careful modeling of the explosive nature of volatility shocks essential for capturing excess correlation?

Key findings

• Excess correlation in asset returns is significantly driven by shocks to aggregate volatility, particularly when the explosive nature of these shocks is properly accounted for.
• Global markets exhibit time-varying cross-sectional sensitivities to volatility shocks, which are strong predictors of subsequent financial crises.
• Periods of high sensitivity to volatility shocks are consistently followed by increased systemic risk and market stress.
• The temporal evolution of the correlation between volatility and returns aligns meaningfully with known financial crises and major market events.
• Standard factor models alone fail to capture the full extent of excess correlation, which is only revealed when explosive volatility dynamics are explicitly modeled.
• The model demonstrates that volatility shocks are not uniformly distributed but exhibit clustering and asymmetry across asset classes and regions.