Time-varying causality tests and the application to BRIC and US stock markets

主讲人:陆凤彬(预测科学中心)
时间:2013年11月14日上午9:00   地点:思源楼一层报告厅

摘要:
To detect time-varying causality among financial markets, we propose time-varying Granger causality tests based on Hong (2001) tests and the dynamic correlation estimator from DCC-MGARCH (e.g., DCC-MGARCH Hong test). We apply the proposed tests to examine the dynamic causality between BRIC and US stock markets. The empirical studies show that the causal relationships between BRIC and US stock markets exhibit a time-varying behavior apparently. The bidirectional causal relationships grow stronger on the whole, which means BRIC stock markets become more and more integrated with US stock market, and the bidirectional tests increase rapidly after some financial crises, which indicates the occurrence of financial contagion. In the unidirectional causal relationships, US stock market plays a dominant role in BRIC-US relations before March 24, 2009; thereafter, though US leads Chinese, Russian and Brazilian markets, US no longer dominates India-US relations, which is mainly due to the financial crises in Europe since 2009 and India's financial openness.