Do different time horizons in the volatility of the US stock market significantly affect the China ETF market?

ABSTRACT

This article investigates the linear, nonlinear and time-varying Granger causality between different time horizons in the volatility of US stock market and the China Exchange-Traded Fund (ETF) market. We find evidence of linear causality from the US stock market to the China ETF market, with a bilateral nonlinear causal relationship in the longer term. Bootstrap rolling causality analysis indicates high rejection rates of a noncausal relationship running from the US stock market to the China EFT market. The causality linkage from the China ETF market to the US stock market was determined to be time-horizon-dependent, and the null hypothesis rejection rate of non-Granger causality increased in the longer term.

KEYWORDS:

• Linear and nonlinear causality
• rolling window causality
• US stock market
• China ETF market
• time horizon in volatility

JEL CLASSIFICATION:

• C22
• G11
• G15

Acknowledgements

The authors thank the editor and the anonymous reviewer for suggestions and comments. We are also grateful to Cees Diks and Jonathan B. Hill for offering the codes on the nonlinear and time-varying causality test. All errors are ours.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ As a common practice, we assume that and for presentation purposes, .

² Due to limited space, this article does not include the detailed models defined by Diks and Panchenko (2006) and Hill (2007). Please refer to their papers for further details on their approaches.

³ The Rolling Window Bivariate Causality Tests adopted here are based on 500 bootstrap replications, and the code is available at: http://www.unc.edu/~jbhill/software.htm.

⁴ Available at http://www.cboe.com.

⁵ The causality between the two residual series of the VAR model represents nonlinear predictive power by removing the linear predictive power with the VAR model (Hiemstra and Jones 1994). The lag lengths of the VAR are based on the Schwarz information criteria.

⁶ Empirical evidence indicated that the GJR-GARCH (1,1) model offers a good estimate for the fluctuation of a financial series (Jiang, Mo, and Nie, forthcoming).

⁷ The lag lengths of the VAR are based on the Schwarz information criteria.

Additional information

Funding

This work is supported by the Natural Science Foundation of Guangdong Province [grant no. 2015A030310444], the Social Sciences Funding Program of Guangdong Province [grant no. GD14XYJ03] and the Fundamental Research Funds for the Central Universities [grant no. 15JNQM009].