Boško Šego

Asset allocation and regime switching on Croatian financial market

It has been known for quite some time now that financial markets exhibit changes in regimes over time. A majority of the literature tends to support that financial markets undergo regimes of bull and bear markets. This characteristic should be modeled in a proper way as investors are always interested in beating the market: either by achieving better returns than others, or by minimizing their portfolio risks. There exist many mathematical and statistical models that are used as tools to achieve the mentioned goals. Introducing the regime switching methodology in existing models has proven to be facilitate achieving such goals. Therefore, the objective of this study is to utilize the regime switching methodology on the Croatian financial market to ascertain its usefulness for Croatian investors. Multivariate regime switching and non-switching models were estimated using daily data from the period 2 January 2007 to 31 December 2015. The assumption is that the investor is interested in stock and bond markets. The results from the MGARCH and regime switching MGARCH models are then compared in order to give answers as to whether the respective methodology applied to the Croatian market is useful and how it may benefit investors. Most of the results support the presumption of incorporating this particular methodology in financial modeling for the Croatia markets. This is the first research that applies the regime switching MGARCH methodology in Croatia (including the Balkan region), hence we expect that this will be a significant contribution to existing methodologies in literature.

Dynamic Portfolio Selection on Croatian Financial Markets: MGARCH Approach

Abstract Background: Investors on financial markets are interested in finding trading strategies which could enable them to beat the market. They always look for best possibilities to achieve above-average returns and manage risks successfully. MGARCH methodology (Multivariate Generalized Autoregressive Conditional Heteroskedasticity) makes it possible to model changing risks and return dynamics on financial markets on a daily basis. The results could be used in order to enhance portfolio formation and restructuring over time. Objectives: This study utilizes MGARCH methodology on Croatian financial markets in order to enhance portfolio selection on a daily basis. Methods/Approach: MGARCH methodology is applied to the stock market index CROBEX, the bond market index CROBIS and the kuna/euro exchange rate in order to model the co-movements of returns and risks on a daily basis. The estimation results are then used to form successful portfolios. Results: Results indicate that using MGARCH methodology (the CCC and the DCC model) as guidance when forming and rebalancing a portfolio contributes to less portfolio volatility and greater cumulated returns compared to strategies which do not take this methodology into account. Conclusions: It is advisable to use MGARCH methodology when forming and rebalancing portfolios in terms of portfolio selection.