<![CDATA[This study intends to provide initial evidence of the randomness of the time-series behaviors of annual earnings of the companies in JSX. It, specifically, aims at examining the problems about: first, whether or not the time-series behaviors of annual earnings are random (following random walk model), second, whether or not Box-Jenkins models (ARIMA) are relevant to describing the time-series behaviors of annual earnings, and third, whether or not the parameters estimated in the ARIMA models are the same as one would expect for a random walk modelThe result of this research shows that the time-series behaviors of annual earnings are random (following random walk models) and can be relevantly and significantly modelled by Box-Jenkins models. This is indicated from the significant parameters of ARIMA models: either autoregressive, moving-average, or both of them. The parameters of those models identified, however, are just the same as one would expect for a random walk model.]]>
Copyrights © 2013
