This thesis investigates how mean-variance asset allocation can benefit from factors in stock returns on the U.S. stock market. The performance of several portfolios formed upon a factor-based meanvariance analysis will thus be evaluated and compared to the performance of an equally weighted portfolio. Based on this objective, the study first establishes a theoretical foundation for the identification of factors and the implementation of the empirical analysis. Having presented the theoretical foundation, a review of the literature within the field of factor models is conducted to investigate how factors in stock returns can be explained, and which factors have been found to explain variability in stock returns. Particular attention is paid to four well-established factor models within the literature. These include the factor models of Sharpe (1963), Fama & French (1993, 2015) and Carhart (1997). Afterwards, the thesis presents a framework for the implementation of a factor-based mean-variance analysis. The framework is an attractive alternative to the traditional mean-variance analysis, as problems of singular variance-covariance matrices are not encountered in our implementation. The framework describes an unconstrained and a constrained solution to the mean-variance optimization procedure. Moreover, the framework is assessed in an econometric context. The factor-based mean-variance analysis allows us to form portfolios on a dataset of stocks included in the S&P 500 Index. This provides the basis for the empirical analysis, which involves portfolio backtesting and performance evaluation from January 1979 to January 2019. The results are that the portfolios formed upon the factor-based mean-variance analysis do not outperform the equally weighted portfolio over the full evaluation period. However, separating the full evaluation period into several decades shows evidence for better performance of the factor-based portfolios relative to the equally weighted portfolio. Several aspects of the thesis will form the basis of a discussion including causations for our findings, data biases and practical applicability. Along with the results of the empirical analysis, the discussion will lead to proposals of topics for further research.
|Educations||MSc in Finance and Investments, (Graduate Programme) Final ThesisMSc in Business Administration and Management Science, (Graduate Programme) Final Thesis|
|Number of pages||134|