Asymmetric causality in-mean and in-variance among equity markets indexes

• The asymmetric causality in-mean and variance among stock market indexes is our goal.
• The different effects from positive and negative returns is analyzed.
• To avoid the asynchrony in daily data, the indexes are grouped in geographic areas.
• From financial returns, causing shocks are extracted by econometric models according to style facts.
• Despite, the results showed empirical evidence of asymmetry causality relationship.
• For Asian, the SSE (China) and SENSEX (India) are asymmetric causality drivers.
• For American, Dow Jones 100 and STANDARD&POORS 500 are main causality drivers.
• For Europe, DAX, CAC 40, MIB 30 and IBEX 35 are asymmetric causality drivers.

The aim of this study is to research asymmetric causality in-mean and variance among financial markets. The methodology used has several advantages: the estimation is jointly and not by pairs of variables, it identifies whether the causality is asymmetric (different effects from positive and negative returns), and, in the case of bidirectional relationships, it allows us to test whether the effect is the same in both directions. This study provides evidence of causality in-mean and variance among the daily returns of stock markets indexes from January 1, 1997 to December 31, 2014. The indexes are grouped by geographical areas to avoid problems of asynchrony with partially overlapping markets. The main results show that the causalities in-mean and in-variance are asymmetric with a different effect from positive and negative returns. Notably, by geographical areas, the Indian index SENSEX (symmetric) and the Chinese Shanghai Stock Exchange (asymmetric) are causatives in-mean for Asia, The Italian MIB 30 (asymmetric) is causative in-variance for Europe, and the USA DOW JONES 100 (symmetric) and STANDARD&POORS 500 (asymmetric) indexes are causatives in-mean and in-variance for America. Also, the bidirectional relationships between the DOW JONES 100 and STANDARD&POORS 500 in-mean and in-variance have the same effect in both directions.