Determinations of corporate earnings forecast accuracy: Taiwan market experience

Abstract

Individual investors are actively involved in stock market and are making investment decision based on publicly available and nonproprietary information, such as corporate earnings forecasts from the management and the financial analyst. Also, the management forecast is another important index investors might use. To examine the accuracy of the earnings forecasts, the following test methodology have been conducted. Multiple regression models are used to examine the effect of six factors: firm size, market volatility, trading volume turnover, corporate earnings variances, type of industry, and experience. If the two-sample groups are related, Wilcoxon two-sample test will be used to determine the relative earnings forecast accuracy. The results indicate that firm size has no effect on management forecast, voluntary management forecast, mandatory management forecast, and analysts' forecast. There are some indications that forecasting accuracy is affected by market ups and downs. The results also reveal that relative accuracy of earnings forecasts is not a function of trading volume turnover. However, management's earnings forecast and analysts' forecasts are sensitive to earnings variances. Readers are well advised and referred to the chapter appendix for methodological issues such as sample selection, variable definition, regression model, and Wilcoxon two-sample test. ​