International transmission of stock market movements: a wavelet analysis

Abstract

This study investigates the international transmission mechanism of stock market movements via wavelet analysis. While GARCH-type models have mainly been used in most recent research in this area, the discrete wavelet decomposition is used in this study to propose a new methodology for investigating the dynamics and the potential interaction in international stock markets. Using the data on daily stock indices from the USA and Korea, strong evidence is found for price as well as volatility spillover effects from the developed stock market to the emerging market, but not vice versa.

Acknowledgements

I would like to thank G.H. Lee for helpful conversations and suggestions on an earlier draft. Financial support from the Korea Research Foundation (grant number: 2000-003-C00173) and the Brain Korea 21 Project in 2001 are also gratefully acknowledged.

Notes

¹ Recent studies on this area include, among others, Eun and Shim (1989), Hamao et al. (1990), Ng et al. (1991), Lin et al. (1994), Karolyi (1995), Kim and Rogers (1995), Booth et al. (1997), and Ng (2000).

² For comparison, an alternative choice for the basic wavelet such as the ‘symmlet’ is also tried, but the results are not much affected.

³ The 693rd observation indicates the KOSPI return on 1 December 1997.

⁴ For a given calendar day, the trading hours of the Korean market precede those of the US by 14 hours. Thus, if the Korean stock market is influenced by any developments in the US market, the former would not be able to respond to a US stock on the same day; instead, it would respond to the US stock with a one day lag.

⁵ As the Korean stock market is closed before the US market opens for a given calendar day, spillover effects from the former to the latter, if any, could occur in the same day.