Article ID: | iaor2006189 |
Country: | Netherlands |
Volume: | 163 |
Issue: | 1 |
Start Page Number: | 242 |
End Page Number: | 252 |
Publication Date: | May 2005 |
Journal: | European Journal of Operational Research |
Authors: | Holden Ken, Thompson John, Ruangrit Yuphin |
Keywords: | stock market |
This paper systematically examines daily returns of the Thai Stock Market Index to determine whether there is evidence of calendar effects due to the day of the week, the month of the year, days before and after holidays and within-month effects. Particular attention is given to the evidence of these anomalies prior to, during and after the Asian crisis. Most of the previous empirical evidence examines each of these effects in isolation. The approach adopted within this paper is to start with a general model, which incorporates a range of different anomalies before estimating the conditional volatility models. Daily data from 3rd January 1995 to 29th December 2000 are used for estimation providing 1473 observations. The final models are tested by comparing their forecast performance for 2001 (244 observations) with various models, which do not include calendar effects. The conclusion is that the inclusion of calendar effects improves the forecast accuracy.