| Abstract [eng] |
The master's thesis consists of three main parts: literary analysis, research methodology, and the research and its results. The literary analysis section discusses theoretical aspects of investment portfolio optimization, including the concept, evolution, and importance of an investment portfolio, as well as the diversification and optimization of an investment portfolio. Three main models of investment portfolio optimization are described: modern portfolio theory, post-modern portfolio theory, and the capital asset pricing model. The section also reviews the criticism of investment portfolio optimization in scientific literature, where the main idea is that not everyone unequivocally agrees on the usefulness of investment portfolio optimization in practice, compared to non-optimal portfolios. This led to the goal of the study - to investigate whether optimal investment portfolios are more useful than non-optimal portfolios. The key points of the research methodology section include evaluating the usefulness of optimized investment portfolios by comparing them with two different non-optimal portfolios ("1/n" and "market"), applying three different optimization methods (maximizing the risk-return ratio, maximizing return with fixed risk, and minimizing risk with fixed return), and using eight different risk measurement units (standard deviation, downside deviation, absolute deviation, downside absolute deviation, 95 and 99 percentile value at risk, 95 and 99 percentile conditional value at risk). The study used assets – exchange-traded funds selected from 8 different asset class categories (stocks, bonds, commodities, real estate, mixed assets, preferred stocks, alternative investment instruments), with 4 assets selected in each category. Thirty randomly selected portfolios were created, and unequal variance T-test was applied to compare optimized and non-optimal portfolios. The results of the study show that optimized investment portfolios are not always superior to non-optimal portfolios in all cases. If optimization brought benefits in all cases compared to "1/n" portfolios, it depended on the optimization methods and risk measurement units when compared to "market" portfolios. |
