| Summary: | 碩士 === 國立中興大學 === 財務金融研究所 === 92 === It is always the research focus that to what extent does earning announcement influence the market and how the market responds. And lots of theories and empirical models are constantly come up with. However, there is still no definite answer to the effect of earning announcement. Furthermore, whether industry factor has significant impact on the earning announcement, the past study results are not consistent. As a consequence, we would pay more attention to the two topics by cataloging electronic industry, financial industry and sewing industry to see if industry difference has significant effect and if earning announcement is informative.
This essay is trying to solve the bias and inconsistency caused by the traditional Ordinary Least Square (OLS) by using Two Stage Least Square (2SLS) and Three Stage Least Square (3SLS) according to the simultaneous- equations by Tung (2002). We also use traditional OLS method as a comparison and the empirical result shows simultaneous equations indeed decrease the bias and inconsistency when estimating the parameters.
Through our empirical study, we can reach the conclusions. First, the industry factor affects significantly, whether OLS or simultaneous-equations are adopted. To put it differently, the extent of earning announcement effect varies in different industries and industry factor is an important variable we would never neglect when dealing with the information effect of earning announcement. Second, it is more appropriate to use simultaneous-equations rather than traditional OLS in our study. Because the values of estimated parameters apparently increase, it proves that simultaneous-equations can solve the problems of bias and inconsistency. Third, although the earning announcements released by the three industries all have got information contents, the effects differ. Earning announcements released by electronic industry and sewing industry increase the degree of information asymmetry and financial industry otherwise.
|
|---|
