This report adopts an event study approach which was introduced by [Brown & Warner (1985) and (1980), Fama, Fisher, Jensen & Roll (1969) and Ball & Brown (1968)]. The Mean Adjust Return model, which is consistent with the Capital Asset Pricing Model used by the above authors, is implemented in this report. This model examines whether or not the returns on the sample securities in day "0" are statistically significantly different from the returns on securities in the time period surrounding the event.
The report analyses daily announcements sent via the Signal "G" wire to the Australian Stock Exchange through the SIRCA Project. The announcement data analyzed in this report covers the period 1 July 1994 to 30 June 1996. The Oracle database and SAS statistical package was used to filter the test data and normalize it. A number of properties were required to be present in the announcen1ent data before it could be accepted for further tests. The selection criteria required announcements to be classified by the ASX as Market Sensitive and announcements of companies were selected where there was no other announcements twenty days before and ten days after the selected event date. After the filteration process, 1347 companies met the selection criteria out of the initial 5000 companies covering a period of 505 trading days.
The volume traded, value and price change data used in this report was normalized to achieve average volume traded, value and price change over a hundred and twenty days and twenty one days prior to the event date. Results of all the three tests show that there is abnormal performance on the day of the announcement and the day after. However, volume and value traded is abnormally high on the day prior to the event date, but not as high as the event day and the day after. These findings are similar to those previously found by [Beaver (1968), Petell & Wolfson (1984), Bamber (1986), Smirlock & Starks (1988) and Atkins & Basu (1995)].
Abnormal trading patterns before the event date is due to investors having access to private information and effects of these announcements are significant on at least the first day or t\VO after the announcements is made, this is probably due to thin trading in the market. [Karpoff (1986)] also suggests that volume traded and value increases caused by informational event persist after the event period which is consist with the existing empirical evidence and suggests that market do not immediately clear all orders or that investors have demands to re-contract.
Price change data used in the analysis also show abnormal changes in price on the day of the event. This finding is consistent with the EMH theory (price changes are directly associated with information arrival) and with the empirical literature in this area.
Seasonality is present in the data set, such as, in the day of the week data, the number of announcements released on Monday is at its low which starts picking during the other days of the week and by the end of the week, Friday, it reaches its low again. There is also some evidence of seasonality in the month of the year data and the time of the day data, both of these like the day of the week data, revealed a U-shape pattern, which was also examined by [Atkins and Basu (1995)].