Dynamic Building of Dividend Stock Portfolios Using Copulas

This master's thesis studies the advantages of copula-based multivariate models application to the problem of high dividend portfolio optimization in the Russian stock market. Alternative optimization methods that compete with the use of copulas in the framework of hypotheses, are naive approach, the application of the Markowitz portfolio theory, as well as the market portfolio investing in the form of the Moscow Exchange Index and its modification of the total income. Based on an extensive list of references, electronic publications of leading journals, as well as data from financial terminals Thomson Reuters Eikon, Bloomberg, the author describes in detail the procedure of high-dividend stock portfolio optimizing using ARMA and GJR-GARCH models for estimating the parameters of marginal distributions and D-Vine constructions of paired copulas to simulate the joint distribution. Technically, the work was executed in the RStudio programming environment, data preprocessing was performed in MS Office Excel. The volume of the master's thesis is 70 pages, containing 3 figures and 4 tables, when writing the work 41 sources of literature were used. Keywords: copulas, ARMA, GJR-GARCH, CVaR, optimization, investment portfolio, dividend stocks, dividends. The master's thesis includes an introduction, three chapters, conclusion, and a list of sources used.