The Halloween Effect : A trick or treat in the Swedish stock market?

The Halloween effect refers to higher stock returns during the period November to April compared to May to October. This is a well-known calendar anomaly that has gained a lot of attention due to the fact that the effect is persistent in the market in spite of the fact that investors are aware of the anomaly today. This evokes questions regarding the efficiency in the markets and the Efficient Market Hypothesis in particular. The main focus of this thesis was to investigate whether the Halloween effect still exists in the Swedish stock market and if the power of the effect deviates between different firm sizes. Furthermore, we examined risk differences between the summer -and the winter months, as well as the January effect in order to find out if these could be possible explanations for the Halloween effect and its existence. A trading strategy based on the Halloween effect was also tested in order to see if investors could use this strategy to outperform a buy and hold strategy. The method that was used to investigate the existence of the Halloween effect was Ordinary Least Squares regression models with dummy variables, standard deviation to ascertain risk-differences between the periods and the Sharpe ratio to determine the risk-adjusted returns of the trading strategies. The results showed that the Halloween effect could be found in all of the examined market-cap indices, and therefore the Efficient Market Hypothesis could be questioned. The Halloween effect turned out to be autonomous from the January effect and the risk measured in standard deviation had no significant difference between the summer -and the winter months, hence, both these possible explanations were rejected. The backtesting showed that the Halloween strategy would perform better than the buy and hold strategy in all indices except from the mid-cap index. The results regarding the Sharpe ratio indicated that the Halloween strategy would be a better strategy to use considering risk-adjusted returns as the Sharpe ratio was higher in all indices