This study adopts the autoregressive conditional jump intensity (ARJI) model proposed by Chan and Maheu [J. Business Econ. Stat. 20 (2002) 377–389] to investigate the impact of news on SIMEX-Nikkei 225 and CME-Nikkei 225 (regards it as the twins). Empirical results demonstrate that the twins were captured by responses to various events; moreover, the twins have distinct jump intensity and risk. Finally, this investigation evaluates the lead–lag relationship between returns and jump behavior by the Granger causality test. Returns are based on unidirectional causality from two futures (the twins) to spot and feedback causality between the twins. Jump intensity reveal feedback causality between spot and the CME-Nikkei 225 and unidirectional causality from the CME-Nikkei 225 to in SIMEX-Nikkei 225.
Physica A: Statistical Mechanics and Its Applications 377(1), pp.199-210