aalto1 untyped-item.component.html

Measuring performance of a stochastic dominance based portfolio selection model in Nordic stock markets

Loading...
Thumbnail Image

URL

Journal Title

Journal ISSN

Volume Title

School of Business | Master's thesis

Date

Major/Subject

Mcode

Language

en

Pages

40 + 7

Series

Abstract

Constructing stochastically dominating stock portfolios is attractive, since a stochastically dominating portfolio would mean higher returns without higher risk, lower risk without lower returns, or both higher returns and lower risk simultaneously. There is a lack of studies measuring the out-of-sample performance of stochastic dominance based portfolio selection models, when investing directly in stocks, or in Nordic markets. Therefore, there is a need for this study, which aims to increase knowledge about that topic. The research approach of this study is mainly quantitative. The data used includes 8 years of daily return data of OMX Nordic 40 index and its constituents. 84 different 1-year formation periods and 30-day out-of-sample holding periods are used. The portfolio selection model used is a stochastic dominance based model by Kopa & Post (2015). Empirical findings are that – in this empirical case – returns are higher and risk lower for the optimized portfolios compared to the index. Especially, cumulative returns over all holding periods combined are considerably higher for the optimized portfolios than for the index. Therefore, the optimized portfolios have excellent performance compared to the index. However, the portfolio turnover for this investment strategy is relatively high. Even though the empirical findings are promising, the statistical evidence of these findings being generalizable across samples is mostly weak, with the exception of strong evidence for lower variance of daily returns. Moreover, the high turnover can lead to high transaction costs and taxes in real-world investing, which can lower returns for such stochastic dominance based investments strategies that have high turnover. Stochastic dominance based portfolio selection models might be an effective way to systematically increase returns and lower risks in real-world investing. Yet, more research, testing and development of such models in real-world context is needed. Also, academically, more similar studies are needed to conclusively determine how such models perform in various stock investing contexts.

Description

Thesis advisor

Kuosmanen, Timo
Liesiö, Juuso

Other note

Citation

Endorsement

Review

Supplemented By

Referenced By