| Popis: |
Analyst recommendations are one of the types of information whose appearance on the market can have an influence on security prices. In this paper, I study the impact of analyst recommendations on stocks listed on the WIG20 Index, using event-study methodology and linear regression models. The dataset contains 576 absolute recommendations published from the1st January 2012 to the 1st of September 2015 by various analyst houses. The prefatory study researches price reaction to positive, neutral, and negative recommendations separately. Subsequently, to check if investor reaction depends on a change in the level of recommendation, corresponding research is repeated for events clustered in nine groups defined in terms of possible level changes. Linear regression models with categorical variables are used in search of additional factors affecting investor reactions. Changes in the level of recommendation, size of the company, and reputation of brokerage house represent explanatory variables. Preliminary results point out that the direction of investor reaction is generally consistent with the information contained in the recommendation, and that the reaction of the market seems to be stronger in the case of positive events than in the case of negative ones. The analysis of recommendation changes reflects more-detailed dependents. In particular, the interpretation of a neutral recommendation depends strongly on the level of the previous recommendation. If it represents growth from SELL or REDUCE, the reaction is positive, while in the case of drop from ACCUMULATE or BUY, it leads to negative abnormal returns. This relationship is additionally confirmed by results from the linear regression models. The models show the size of the firm as a significant factor that has an influence on the reaction to a recommendation: the smaller the firm, the stronger the reaction. |
