Prediction of Analysts Earnings Forecast Errors on Russian Stock Market

The ability to predict what mistakes analysts make in their forecasts for the EPS of companies on the Russian market and which determinants have an effect on this mistakes is actual both for the internal use of any investor and for building strategies on companies' stocks around the date of financial statement publication. A number of studies demonstrate a positively significant anomalous return, when the actual EPS values ​​turned out to be higher than the predicted ones, negatively significant return for the opposite situation, while errors close to zero do not affect the behavior of company stocks. Most often, both in developed and emerging markets, analysts are over optimistic, what results negative errors in the date of financial statement publication of analyzed companies. However, often the company's management tries to give as much information to the market as possible in order to avoid sharp shift in stock prices. For correct forecasting, it is necessary to understand what characteristics of the analyst, the company itself and the market can affect the resulting errors of analysts on the Russian market. In this paper, a completely new model was applied to existing studies on this topic - generalized ordered logistic regression, which allows to take into account what factors influence the change in the probability of analytics errors in certain groups (group 1 - errors statistically different from 0; group 2 - errors are statistically greater than 0, positive surprises, group 3 - errors are statistically less than 0, negative surprises). According to the results of this work, it was revealed that analysts are extremely optimistic about the companies on the Russian market (the average forecast EPS error for the period 2011-2018 between fact and forecast, divided by the actual value, was -0.401). In addition, it was found on a sample of 32 companies (172 unique analysts) and annual data of last forecasts for 2010-2017 that such factors as: analyst confidence, the number of companies covered by the analyst, the analyst's "name", the number of days between the publication of the latest forecast and the annual reporting of the company; the size of the company, the company's debt leverage, the liquidity of the company's shares, the book-to-market ratio, the frequency of news, the share of insiders in the property of the company, the age of the company, the company's loss for the last reporting year; information asymmetry on the market have a significant impact on type of the analyst's error . In addition, the market timing theory was tested.