Short-Run Behavior of Stock Returns : Speed of Adjustment and its Contributing Factors in the Jakarta Stock Exchange

This study examines the price formation process, noise level, non-trading and component of bid-ask spreads of individual shares listed on the Jakarta Stock Exchange for the period from 2000 to 2004. The price formation process is estimated using speed of adjustment based on the simple partial adjustment method proposed by Amihud and Mendelson (1987) that was later extended into ARMA(1,1) estimation by Theobald and Yallup (2004). The results are consistent with studies for the U.S. and European markets that find short term underreaction in security returns for most stocks. We further find that large companies lead small companies in price adjustment to new information. Using the intervalling properties to approximate the time to adjustment, we find that large stocks need only one day to fully adjust to new information, while medium and small stocks need three and five days respectively. The predominant factor contributing to the speed for adjustment for stocks listed on the Jakarta Stock Exchange in the ARMA(1,1) model is the MA(1) component reflecting significant noise in the price formation process. We assume that noise contains two microstructure components, non-trading and bid-ask bounce. This study finds that the role of non-trading is too small to be reliably justified as the source of negative autocorrelation, resulting in overreaction to the speed of adjustment. Further evidence revealed that the role of bid-ask spread is significant in determining the sign of the autocorrelation coefficient. The decomposition of bid-ask spreads disclosed that the size of the adverse selection spread cost component is negatively correlated with the speed of adjustment