Abstract

This paper explored the transmission of returns between the U.S. stock market and four international (Canada, Japan, France, U.K.) stock market indexes from January 1996 to December 2006, using a nonlinear model which focused on threshold effects. The nonlinear, MARS (multivariate adaptive regression spline) model was applied to study this relationship at the general as well as industry specific levels. The nonlinear characteristic of the MARS model was able to provide valuable detailed information with its unique ability to capture and highlight the asymmetric nature of the international financial markets. The findings displayed evidence of asymmetric influence from the previous market from both extreme positive and negative price movements. This paper posits that there are advantages of applying the MARS model to international financial markets, as the results revealed significant information which may be beneficial for managing risk and hedging portfolios.

Key words: Transmission of returns, multivariate adaptive regression spline (MARS), stock market.