Explaining Subsequent Trading Activity Using Wavelet Time-Scale Analysis : International Evidence

This study seeks to examine the behaviour of subsequent trading activity in international stock markets. We investigate whether stock returns can explain subsequent trading using wavelet time-scaled returns and volume. A wavelet decomposition of trading volume over time-scales of up to 32 days is conducted. This trading volume is then regressed on recent and decomposed returns. Furthermore, we consider both the linear and non-linear regressions and the wavelet-variance estimator. We find only a small amount of evidence that recent returns predict subsequent trading, but stronger evidence is observed at longer horizons. Notably, there is greater evidence of a statistical relation over the investment horizons of [8-16] and [16-32] days. Interestingly, this relation has a negative sign for the developed markets considered, but is positive for two emerging markets. It also tends to be economically stronger during bull periods and is robust to the crisis period. We also find that stock market volatility tends to be significantly correlated with recent trading and at the longest timescale [16-32] day. The results in this paper should enable participants in the stock market as well as regulators to better understand the interrelation between returns and trading activity and the changes over different time horizons