An Empirical Study On The Accuracy Of Profit Forecast In China 's Securities Analysts

Earnings forecasts is becoming more and more important in China’s securities market, with the maturity of the market, the improvement of information disclosure, and the popularity of the value investing concept. A comparative study on the accuracy of the earnings forecasts by securities analysts and the statistical models have great significance for investors to optimize the parameters of the valuation model; for researchers to find a reliable proxy for investor expectations of earnings per share; and for the government and regulatory authorities in the guidance and supervision of securities analysts. There has long been a debate on whether securities analysts’earnings forecasts are more accurate and a better surrogate for market expectations of earnings than those generated by statistical models in America. In China, only a few researchers compared the accuracy of earnings forecasts by securities analysts and the statistical models. All the previous research are flawed with small samples-the companies they covered are few because the need of history data for parameter estimate, and the time intervals they covered are only one year or two. Also, all the previous research studied the short-term (1year) forecasts of analysts, ignoring the long-term (2-3years) forecasts.This paper re-compared the accuracy of the earnings forecasts by securities analysts and the statistical models using a simple random walk (RW) model, which imposes the least restrictive on data, with all the companies listed in Shanghai and Shenzhen stock market, and the time period of2007-2011, which covered various market state. We find that, in the China market, securities analysts’EPS forecasts are superior to the RW model for forecasting horizons of1-8months, but inferior to the RW model for longer forecasting horizons. The analysts superiority (inferiority) to RW model decreases (increases) as the forecast horizon increases. For smaller firms and younger firms, the analysts’superiority in short forecast horizon is smaller, but decreases slower than for the total sample; while for the forecasts made by star analysts and analysts who worked in the star institutes, the analysts’ superiority in short forecast horizon is larger, but decreases faster. Using the analysts’1-year forecasts to forecast the longer term EPS can beat the RW model.These findings may help us to better use of the EPS forecasts of listed companies, and to better understand the source of the accuracy differences between securities analysts’forecasts and the RW model forecasts.