Modifiering av Fama och Frenchs trefaktorsmodell för att estimera avkastning för enskilda aktier - Ett verktyg som simplifierar urvalsprocessen vid aktieköp för småsparare

Abstract

A major stream of capital from private investors is present on the Swedish stockmarket, eventhough the majority of private investors possess limited knowledge about the stockmarket. As a result from this, private investors are exposed to high risks compared to the expected return. This thesis aims to investigate whether theories regarding estimation of return on marketportfolios can be used by private investors to estimate the risk-adjusted return on individual stocks and, thus, facilitate selection of stocks. Two core theories within this area are the three-factor model and the Capital Asset Pricing Model (CAPM). The authors investigated whether the former model can be adjusted to estimate return on individual stocks, to what extent returns can be explained by an adjusted model and if estimated returns corresponds to real returns. The thesis is based on a inductive and deductive method in which secondary data, from 2006 to 2014, was gathered from Yahoo Finance and RetrieverBusiness in order to answer the research questions, by primary using regression analysis. The thesis shows that the three-factor model can be adjusted to estimate return on individual stocks by departing from the three-factor model’s original categorization and instead use continuous variables. Regression analyzes showed that the extent to which an individual stock’s return could be explained, fluctuates greatly. The regression models, for the stocks selected to be analyzed, appeared to not be able to explain return better than CAPM, allthough investment trust companies had considerably higher coefficients of determination. The most probable reason is that these companies acts as portfolios and, thus, contains smaller fluctuations. Furthermore, regression models from a certain time period were analyzed in relation to real values from another time period, which showed that larger deviations were present for larger fluctuations of real values. Since only a few regression models can explain return to the requried degree, the models are not perceived to be applicable directly as a tool for private investors. On the other hand, the model is based on larger amounts of data than the requirements obtained from theories above. To conclude, the authors does, however, find the results useful since the model can be applied on investment trust companies.