Intraday price information flows between the CSI300 and futures market: an application of wavelet analysis

Attributions of changes in price leadership involve more market structure variables, such as investor structures and market openness, and are left for future studies.

This study focuses on causality tests. Second or higher-order moment modeling of nonlinearities is left for future studies.

Gençay et al. (2002) and Ramsey (2002) explained how wavelet analysis could be utilized in economic and finance. Several studies in these two areas that use wavelet analysis include but are not limited to: Ramsey and Zhang (1997)—foreign exchanges, Davidson et al. (1998)—commodity price behavior, Ramsey and Lampart (1998a, b)—decomposition of the relationship between expenditure and income, Pan and Wang (1998)—stock market inefficiency, Lin and Stevenson (2001)—causality between the equity spot and futures market, Gençay et al. (2003a, b)—systematic risk in an asset pricing model, Kim and In (2003)—the relationship between financial variables and real economic activity, Almasri and Shukur (2003)—causality between public expenditure and income, Dalkir (2004)—causality between money supply and income, Lee (2004)—causality among international stock markets, Kim and In (2005a)—the relationship between stock returns and inflation, Kim and In (2005b)—the Sharpe ratio, In and Kim (2006a)—hedge ratios and the relationship between the stock and futures market, Kim and In (2007)—the relationship between changes in stocks prices and bond yields in the G7 countries, and Alzahrani et al. (2014)—causality between oil spot and futures prices.

Readers are referred to Hou and Li (2013) and Yang et al. (2012) for more institutional backgrounds of the CSI300 and futures. Different platforms may provide different price information.

Unless stated otherwise, I will refer to “log prices” as “prices” hereafter.

The ADF, PP, and KPSS test results are available upon request.

Detailed numerical results are available upon request.

The null hypothesis “\(\beta _{1}+\beta _{2}=0\)” is failed to be rejected in Yang et al. (2012) based on the sample from April 16, 2010, to July 30, 2010.

Other measurement of relative contributions of different markets to information flows in the long run includes Gonzalo and Granger’s (1995) common factor weight model and Hasbrouck’s (1995) information share model. Schwarz and Szakmary’s (1994) approach can be derived from Gonzalo and Granger’s (1995) framework and has similar qualitative results with Hasbrouck‘s (1995) method (Theissen 2002; Bohl et al. 2011).

Detailed numerical results of the BDS test are available upon request.

The choice of a VAR specification is based on the fact that residuals are stationary. Lags are selected by the Bayesian information criterion (BIC). Detailed numerical results are available upon request.

The ADF, PP, and KPSS test results are available upon request.

Detailed numerical results of the BDS test are available upon request.

Detailed numerical results not reported to save space are available upon request.

See appendix in Hiemstra and Jones (1994) for a detailed derivation of the variance.