Abstract This article discusses various types of time series processes for accounting earnings, based on results of previous studies. Basically, evidence regarding the time series properties is important since it provides some guidance as to the process generating accounting signals. Three accounting processes are used and discussed here: the mean reverting process, autoregression, and moving average process. According to the author, a pure mean reverting process can be used for generating accounting signals if: events affecting the firm had no future period impact or events had a future period impact and the accounting system recognized the future period effects in the current periods earnings. With the statistical properties of the three processes as a foundation, the author examines findings of several previous studies. The purpose of the time series analysis in this study is to distinguish between these two competing explanations of the observed moving average properties of the series of earnings signals.
Larry L. Lookabill (Fri,) studied this question.