My doctoral thesis provides evidence of two different types of trading strategies. The first type is based οη market neutral trading strategies under the methodology of pair trading strategies. The second part is on rotation strategies according to sign forecasting specifications and explores the probability of profitable market and volatility timing. The thesis is comprised by three chapters. The first chapter is dedicated to Exchange Traded Funds, ETFs. Ι am presenting an extended literature review on the topic. The review tries to capture every aspect of ETFs that is of concern for the academic community. Moreover, Ι am presenting the mechanism of ETFs and the pros and cons that are inherent in an ETF structure. Ιη addition, Ι am discussing active ETFs. Οη the 4th of March 2008, the Securities and Exchange Commission approved the listing and trading of Active Exchange Traded Funds in the US market. This decision opens up a new era on asset management. Ι am trying to identify the most appealing issues from this new decision. Ι am analysing the similarities and the differences with passive ETFs and conventional mutual funds and the obstacles that arise from the inception of active ETFs. The second chapter is dedicated to pair trading strategies. Gatev, Goetzmann and Rouwenhourt (2006) applied a trading algorithm based οη the concept of mean reverting returns. Prices of two assets that move together in the long run and diverge in the short term will revert to their equilibrium. An alternative definition for the pair trading strategies is that of a relative value statistical arbitrage methodology. Engleberg, Gao and Jagannathan (2009) examined pair trading methodology and tried to explain the factors behind the profitability. The contribution of my work is the implementation of a modification of pair trading investment strategy and the examination of the profitability and the motives that create the profitability in the contest of ETFs. Ι implement different estimations for each separate step of the formulation of the strategies in order to examine and find an “optimal” algorithm. Ι then conduct different tests to check the robustness of my methodology. Ιη the next step, Ι check the pattern of profitability based on several tests based on the segmentations according to market capitalization, emerging and developed markets. The second part involves the empirical evidence of pair trading portfolios according to risk profile. Ι incorporated Fama and French risk factors to explain for potential patterns behind the profits. The estimations included national and international risk factors on profitability. The most important part is the decomposition of the traded pairs and the examination one by one according to its own risk characteristics. My dataset is constructed by international ETFs which is the tradable version of country indices. In that concept, Ι research in each separate pair its own variables and Ι test the factors that affect profitability. Among the extended research all over pair trading strategies, this research provides the following contributions. 1. It is the first time that ETFs are used in pairs trading. 2. International evidence on pair trading with easily accessible instruments. 3. Pair trading profitability outperforms S&P500. 4. The US and international Fama and French risk factors are insufficient to explain pair trading international profitability. The third chapter is dedicated to volatility and market timing strategies. Ι examined a new methodology that assesses the economic and statistical significance of market and volatility timing according to a novel forecasting specification. My methodology combines the dynamics of time-varying expected returns and volatility timing and several thresholds derived by expected returns and variability. The specification is incorporating forecasting sign ability. The forecast estimations are incorporated to create trading rules and the formation of portfolios. The trading rules, then, are incorporated to the allocation decision. In every decision, we allocate the total wealth to one asset. In every transaction, we rotate between the two assets. The methodology is based on a pairwise asset evaluation. I test for the patterns behind volatility timing, and for the day of the week effect. The results indicate that under specific assumptions market and volatility timing can lead to profitable trading strategies. The selection of the specification appear to be sensitive between past returns and volatility which confirms the initial conception of the cross interaction between time varying expected returns and variation. Comparing the performance of the rotation portfolios based on forecasts using different model selection criteria, the rotation trading is performing the highest final wealth, when there is not a clear domination between expected return and variation. Applying the methodology under different days of the week, I can differentiate from the literature in means of the performance with rotation trading to exhibits the most statically and economic significant excess returns οη Monday. The next test examines if different levels of volatility generate correct sign predictions. The empirical analysis shows that there is not clear dependence between returns and level of volatility. Empirical evidence appear to be sensitive about the selection of trading specification which confirms the motivation of the research of cross interactions between time varying expected returns and variation. Rotation trading outperformed the market in means of final performance and risk levels as represented by the maximum drawdown indicator. My thesis makes a distinct contribution in the area of active asset management and asset allocation methods. It explores in depth two different trading strategies in the context of a relatively new financial tool, the Exchange Traded Fund (ETF), and is to the best of my knowledge one of the few existing works that address the issue of ETF profitability in a relatively thorough manner, always in the context of active trading.
Measuring Stock Market Investor Sentiment
