| dc.description.abstract | This Master's thesis addresses how to optimise market moving portfolio liquidation through a theoretical approach to liquidity cost optimisation with an exponential bounce-back model in discrete time. We use numerical methods such as Fixed Point method and Gradient Ascent method, as well as analytical methods like Lagrange Multiplier method and more general mathematical theory. Moreover, we evaluate the investor's investment choices when managing a market moving portfolio with respect to liquidity, market value, sales value, liquidity cost and liquidity cost ratio. A normalised sales value model is used to address liquidation in one, two and three dimensions, where the dimensions represent the number of liquidation steps the investor can utilise in order to liquidate the entire portfolio. The thesis explores investment strategy spaces such as buy, sell and hold, as well as pumping, dumping and short selling. The model and market are investigated from analytical, numerical and naive perspectives based on mathematical and financial theory. On this note, we find that the liquidation cost ratio is constant in one dimension; that the naive analysis seems to be severely insufficient compared to an analysis using numerical methods; and that the sequence of market types does not commute, i.e., a BullFlat market does not behave equal to a FlatBull market. Moreover, we shed light on some perspectives as to why most struggle to read financial markets correctly. | eng |