A Bunch of Non-random Walks Down Wall Street: Investing During Up- and Downswings

Per Sloth Aagaard

Student thesis: Master thesis


The aim of this paper is to find a way to exploit booms and busts in stock markets to get a return which is higher than the return for a buy-and-hold strategy. The question that gives rise to this goal is whether stock prices follow short run trends during financial booms and busts. The thesis therefore start out by discussing the relevance of historical stock prices. It finds that historical information is fully reflected in prices hence historical prices cannot be used to predict future prices according to proponents of the Efficient Market Hypothesis (Fama 1980) and (Malkiel 1975). Opponents of EMH argue that portfolios of stocks exhibit serial correlation in weekly returns which means that historical stock prices can be used to predict future stock prices (MacKinlay 1999). The strongest evidence of autocorrelation is found for Nordic stock markets. It is therefore chosen to develop a New Model based on a filter strategy on portfolios of Nordic stocks to exploit potential autocorrelation in returns. Financial insights are hence uncovered to give guidance on how to find stocks for a portfolio and how to assign weights. It is decided to use a Naïve portfolio model because it is not prone to estimation errors. Each portfolio should include 18 stocks to secure adequate diversification. The filter strategy is defined and tested in-sample and it is chosen to use an Y-filter of 25% and an X-filter of 16.67%. The empirical analysis tests the New Model on Danish, Norwegian and Swedish stocks and find that the expected return of the New Model is higher than the expected return of a buy-and-hold strategy for small cap stocks. Contrary, it finds that the average return is lower for the NM for portfolios of mixed, large and mid cap stocks.

EducationsMSc in Applied Economics and Finance, (Graduate Programme) Final Thesis
Publication date2018
Number of pages100
SupervisorsPeter Belling