The thesis tries to explore the performance of an investment strategy that integrates momentum strategies with Black-Litterman model to create portfolios from the universe of Taiwan 50 ETF index. The momentum effects are accounted for by constructing the investor’s view to reflect recent winners that are selected by their historical average returns in top quantiles from the index universe. To achieve robust results from the Mean-Variance portfolio optimization, the sample mean and sample covariance are computed by weekly and monthly return with various data frequencies. Additionally, the performance of the tangency portfolio, minimum variance portfolio, equally weighted portfolio, and equal risk contribution portfolio is compared with the Black-Litterman model as well as the benchmark Taiwan 50 index. The empirical results show that, for a given set of assumptions, the accumulated returns of the Black-Litterman model is significantly higher than Taiwan 50 index but lower than the portfolios in the framework of Markowitz.
