Search

Momentum strategies are receiving increasing attention in the field of finance due to their potential to generate superior returns. Momentum investing is based on the premise that assets that have exhibited strong performance in the recent past are likely to continue to perform well in the near future. This thesis explores the different momentum strategies in diverse markets and asset classes, such as the US equity market utilizing industry portfolios, multi-asset funds, and developed countries. Our study involves a comprehensive analysis of various momentum strategies, including relative momentum and absolute momentum, as well as more recently extended strategies such as dual and triple momentum. In this study, we investigated the optimal parameters for various momentum trading strategies. Our findings suggest that there is no single set of optimal parameters for past price performance and the number of chosen best/worst performing assets that is universally applicable across all datasets and strategies. The study examines momentum strategies using three distinct datasets. The long-only dual momentum strategy has been identified as the optimal approach for the dataset of US industry portfolios, exhibiting a statistically significant Sharpe ratio. The multi-asset fund dataset reveals that the optimal strategy is the triple momentum strategy, which presents the highest statistically significant Sharpe ratio. Similarly, the dataset of developed countries presents the triple momentum strategy as the optimal strategy with the highest Sharpe ratio and the highest statistically significant alpha values from both the Capital Asset Pricing Model and the Fama-French-Carhart Four-Factor Model. Overall, the study highlights the importance of considering various performance and risk measures to determine the optimal momentum strategy.

The breakthrough Black-Scholes (BS) model predicts a horizontal line when plotting the implied volatility (IV) against the strike price. However, empirical studies uncovered that the implied volatility derived from option market prices in the BS model varies with strike prices and time to maturity, leading to the identification of three stylized facts that the BS model fails to explain. First, the IV curves exhibit a smile/smirk pattern, with an upward-sloping term structure for at-the-money options. Second, option prices tend to reflect higher implied volatility compared to the realized volatility of asset returns. Third, the negative skewness implied by options prices is greater in absolute terms compared with the realized skewness. Consequently, numerous sophisticated models have been developed to address these stylized facts. Nevertheless, traditional models often fall short of fully explaining all aspects of these phenomena. This study introduces a novel approach to the utility indifference model by incorporating behavioral utility functions to provide a more accurate representation of these anomalies. To evaluate the model’s performance, the standard function used in expected utility theory and behavioral utility functions are tested under both normal and Normal Inverse Gaussian (NIG) distributions. The findings indicate that the conventional utility function fails to capture the observed smirk patterns. In contrast, the behavioral utility function generates the IV smirks that closely align with empirical shapes, even under the normal distribution. These results highlight the effectiveness of the utility indifference model with behavioral utility functions in explaining these stylized facts that standard models struggle to reproduce.

This thesis examines the effect Norway’s Government Pension Fund Global (GPFG) has on the companies it excludes from the portfolio. The data sample consists of 25 companies excluded based on environmental damage and unacceptable greenhouse gas (GHG) emissions. An event study has been conducted to examine whether the exclusion from the fund has a negative impact on the companies’ stock prices. This makes it possible to examine the abnormal returns, and whether they are affected by the exclusion. The main findings and conclusion are that the exclusions do not negatively influence the excluded firms. The influence of exclusions does not result in statistically significant negative abnormal returns, and thus we fail to reject the null hypothesis. These findings are inconsistent with the majority of earlier studies, which found that the exclusion had a statistically significant negative impact on the excluded companies. This may be because of the differences in the data selection, and information leakage, or it might be explained by the fact that there are always willing buyers. One explanation may also be that investors do not regard sustainability as important and essential for their investments.

The emergence of artificial neural networks has given us some of the most impressive technological tools. Inspired by the human brain, these networks consist of interconnected artificial neurons that can detect patterns invisible to the human eye. These qualities have caught the attention of investors seeking ways to beat the market. In this thesis, we explore how artificial neural networks can be used to construct an active trading strategy and evaluate the strategy's performance against two benchmark strategies. Two stock indices were used to train neural networks using the lagged return as input to predict the market state. By using the networks' predictions, an active trading strategy was constructed. To evaluate the network-based strategy, we test if the Sharpe ratio differs significantly from the Sharpe ratio of a simple moving average strategy and the buy and hold strategy. Additionally, we estimate the alpha in the capital asset pricing model. The results show that the network strategy performs similar to the benchmark strategies in terms of Sharpe ratio and fails to generate significant alphas. Overall, our results contribute to the previous literature seeking to apply neural networks to finance and should serve as a reminder of the shortcoming of financial data for machine learning and the importance of statistical testing.

Several studies have presented evidence for higher returns in stock markets from the beginning of November to the end of April compared to the rest of the year. This phenomenon is well known as the Sell-in-May effect. The main focus of this thesis is to investigate whether the Sellin-May effect still exists in financial markets and whether the power of the effect deviates between time and markets. Additionally, we investigate discrepancies in risk and return between the summer and winter months, as well as the January effect, in order to examine whether these explanations could help us understand the existence of the Sell-in-May effect. In an effort to examine if the investors could exploit and profit from the market anomaly, we developed and simulated a trading strategy based on the Sell-in-May effect and performed various statistical tests against the Buy-and-Hold benchmark strategy. The output from the regression model showed evidence of an existing Sell-in-May effect. Neither the January effect nor differences in risk proved to be viable explanations for the existence of the effect. Further results indicated that the Sell-in-May strategy outperformed the Buy-and-Hold strategy in most scenarios, indicating that investors could exploit and profit from the market anomaly.