TESTING THE CAPM AND THE FAMA-FRENCH 3-FACTOR MODEL ON U.S. HIGH-TECH STOCKS

Abstract: This master’s thesis tests the capital asset pricing model (CAPM) and the Fama-French 3-factor model (FF3FM) for the U.S. high-tech industry. For a total sample of 120 U.S. high-tech companies we run OLS time-series regressions for both models by using return and accounting data from 2002 to 2016. It is found that on average, the CAPM is not sufficient in explaining average excess returns for our sample of U.S. high-tech stocks, indicated by significant abnormal returns in the time-series regressions. However, the FF3FM eliminates the significance of the abnormal returns, or at least lowers the significance of the time-series regressions intercepts. Hence, it is found that the latter model outperforms the traditional CAPM for our sample of U.S. high-tech stocks, indicated by lower significance of the alpha terms as well as increasing adjusted R2 values in the time-series regressions. The higher explanatory power of the FF3FM compared to the CAPM is mainly caused by the high significance of the size factor measured by the SMB (small-minus-big) variable, which confirms a negative size premium for U.S. high-tech stocks. The book-to-market factor represented by the HML (highminus-low) variable does not seem to contribute to explain average excess returns, which is concluded from an insignificant average regression coefficient. Since the FF3FM proves to be an improvement towards the traditional CAPM, it can be recommended to apply the former model as a valuation and decision making tool for U.S. high-tech stocks. However, these results only hold for stable economic periods as the research results show that during economic turmoil both models are not sufficient in explaining the respective average excess returns.