Russia’s financial markets have been an uncharted area when it comes to exploring the performance of investment strategies based on modern portfolio theory. In this thesis, we focus on the country’s stock market and study whether profitable investments can be made while at the same time taking uncertainties, risks, and dependencies into account. We also pay particular interest in tactical asset allocation. The benefit of this approach is that we can utilize time series forecasting methods to produce trading signals in addition to optimization methods.

We use two datasets in our empirical applications. The first one consists of nine sectoral indices covering the period from 2008 to 2017, and the other includes altogether 42 stocks listed on the Moscow Exchange covering the years 2011 – 2017. The strategies considered have been divided into five sections. In the first part, we study classical and robust mean-risk portfolios and the modeling of transaction costs. We find that the expected return should be maximized per unit expected shortfall while simultaneously requiring that each asset contributes equally to the portfolio’s tail risk. Secondly, we show that using robust covariance estimators can improve the risk-adjusted returns of minimum variance portfolios. Thirdly, we note that robust optimization techniques are best suited for conservative investors due to the low volatility allocations they produce.

In the second part, we employ statistical factor models to estimate higher-order comoments and demonstrate the benefit of the proposed method in constructing risk-optimal and expected utility-maximizing portfolios. In the third part, we utilize the Almgren–Chriss framework and sort the expected returns according to the assumed momentum anomaly. We discover that this method produces stable allocations performing exceptionally well in the market upturn. In the fourth part, we show that forecasts produced by VECM and GARCH models can be used profitably in optimizations based on the Black–Litterman, copula opinion pooling, and entropy pooling models. In the final part, we develop a wealth protection strategy capable of timing market changes thanks to the return predictions based on an ARIMA model.

Therefore, it can be stated that it has been possible to make safe and profitable investments in the Russian stock market even when reasonable transaction costs have been taken into account. We also argue that market inefficiencies could have been exploited by structuring statistical arbitrage and other tactical asset allocation-related strategies.