Simulated Dynamic Asset Allocation In The Presence Of Proportional Transactions Cost:Multi-stage Stochastic Programming

In this work, the researcher aims at creating a system for dynamically allocating assets in a simulated environment where for example a risk adverse investor faces transaction cost with asset dependent returns. This system helps the investor to maximize his expected utility over a particular time frame. The system is based on a practical model which uses an innovative version of the three stages stochastic programming to cope with uncertainty of the market. This model deviates from the mathematical standard on the topic of asset allocation in business transactions, but reaches to the same result and justifies the intuition of practitioners. The system captures the reality of the market transactions. To prove its power, the researcher has created a small application to implement the three stages stochastic model using GAMS and the researcher has connected four independent exercises by changing the transaction cost, risk alertness, length of time used and goals in a dynamic simulated environment. In all the exercises the transaction costs are presented as it is in the real market. The results will then be compared with those from other academic works that have used different settings. They are in line with major theories and other practical applications. This model captures very well the reality of market transactions and also has a practical application. The results are tested for robustness. This system based on an innovative three staged stochastic programming works and turns to be a reliable tool to the investor.