Sampling methods for investment portfolio formulation procedure at increased market volatility

Aim/purpose–In this paper, a market volatility-robust portfolio composition frame-work under the modified Markowitz’s approach with the use of sampling methods is developed in order to improve the allocation efficiency for a portfolio of financial instruments formulation procedure at an increased market volatility.Design/methodology/approach–In order to overcome the risk of not receiving an optimal solution to the portfolio optimization (suboptimal outcomes of attribution of weights in allocation procedures) the developed model, first, implements the rationale that financial markets largely feature two states, i.e., quiescent (non-crisis; low market volatility) periods that are occasionally interspersed with stress (crisis; high market volatility) periods and, second, relies on many input samples of rates of return, either from an empirical distribution or a theoretical distribution (mitigating estimation risk). All computational results are reported for publicly available historical daily data sets on selected Polish blue-chip securities. Findings–Not only did the presented method produce more diversified allocation, but also successfully minimized the unfavorable effects of increased market volatility by providing less risky portfolios in comparison to Newton’s method, typically used for optimization under portfolio theory. Research implications/limitations–The research emphasized that in order to get a more diversified investment portfolio it is crucial to outdo the limitations of a single sample approach (utilized in Markowitz’s model) which may on some occasions be statistically biased. Thus it was proved that sampling methods allow to obtain a less concentrated and volatile allocation which contributes the investment decision-making. However, the current research focused solely on publicly available input data of particular securities. In this manner, an additional analysis can be prepared for other jurisdictions and asset classes. There can also be considered a use of other than variance risk measures.Originality/value/contribution–The suggested framework contributes to existing methods a wide array of quantitative data analysis and simulation tools for composing an unique approach that directly addresses the task of minimizing the adverse implications of increased market volatility that, in consequence, pertains to knowledgeable attributing of investment portfolio proportions of either individual or institutional investors. The prepared method is also proved to hold demanded computational quality and, importantly, the capacity for further development.