Trading strategies with market impact constraints – A case study at SEB

Abstract: Purpose of the master thesis is to develop an algorithmic trading model for a single asset, which is well functioning on the Scandinavian market. Deliverables: The main focus has been towards creating an indicating model to quantify market impact, cost of trading and model an efficient trading frontier. Methodology: Initially we focused on research papers and literature within our specific field trying to get a deeper knowledge in related theories. Further we had meetings with our project initiators to understand their ambitions with the project. This helped us to formulate the direction of our thesis. After broad theoretical studies we started to evaluate different models that are used to quantify market impact and efficient trading. We understood that there is not a general model that is used to quantify these values. At this point we were able to decide on certain models to approach and which simplifications that had to be done. We decided to use the theoretic framework for a market impact model created by US researchers. Finally we compiled all research and data into our master thesis. We also had several meetings with experienced individuals from the stock market to discuss the validity of our models that we had developed. In the end we delivered a clear theoretical framework and a model that is simple to use. Conclusions: Based on cutting edge research in financial trading theory combined with discussions with traders on the Scandinavian market a pre trade analysing application has been created. The application is today usable to facilitate the decision-making process when minimizing cost at a desired risk level for a large stock order. The application provides the user with cost and risk estimates. Additional information on how to liquidate the initial portfolio and anticipated market participation is supplied to the user. The information provided by the application is realistic according to experienced professionals. However manual trading can not be completely replaced. Several orders are of such kind that algorithms are not capable of executing these. Further a good trader takes much information into consideration when trading not possible to capture in algorithms. The conclusion is that traders and algorithms should work side by side to maximize profit. Standardized orders are handled by automatized trading systems allowing traders to focus on more complex orders.