A study of analyst forecast reliability in Australia

The purpose of this paper is to determine whether time weighted consensus estimates offer a more effective method for predicting company actual EPS figures than simple mean or median analysis. The study aims to construct a more comprehensive earnings forecast signal using analyst earnings forecasts that have been weighted based on the timeliness of updates. Aimed at extracting valuable information from timely analyst forecasts, the time weighted earnings signal (TWES) methodology allows extracting valuable information from analysts who possess some unique insights about the market and issue their updates more frequently. One would expect the time signal to reflect a more realistic representation of analyst estimate changes and thus be more effective in predicting the companies' reported EPS than the mean and median.