Contributions published at Quantitative Finance (Erich Walter Farkas)

Selling put options can be lucrative; however, the returns tend to exhibit a strong negative skewness. Moreover, the seller may have liquidity issues during the holding period, especially when margin requirements become too large. Existing hedging techniques often overlook potential liquidity problems during the holding period, focusing solely on terminal losses. To address this limitation, we present a novel risk management approach by reformulating the closing time of the short position as an optimal stopping problem. To find the solutions, we decompose the holding period into a sequence of binary stopping decisions, which naturally fit into the reinforcement learning framework. Multiple deep reinforcement learning algorithms, namely Deep Q-Learning, Rainbow, and Synchronous Advantage Actor-Critic, are employed to identify the optimal times for closing the position. Our training framework introduces a new reward function that enables the agents to maximize each option’s profit and enhance its Sharpe ratio. In a simulated environment with nontrivial optimal stopping solutions, we demonstrate the e↵ectiveness of the algorithms and our training setup. Furthermore, we apply these algorithms to market data; specifically, SPY put option data from 2005 to 2022. During this analysis, we encounter a significant imbalance in the training data between paths with negative and positive returns, making it challenging for the algorithms to learn an optimal solution. Consequently, we propose several approaches to tackle this issue in future research. Overall, our work presents a promising approach to address liquidity concerns during option selling strategies, and our findings contribute to the advancement of reinforcement learning techniques in the financial domain.

We investigate the stock return predictive power at the high-frequency level after statistically significant overnight jumps conditioned on prevailing stock and market level news sentiment. We provide evidence that sentiment variables as well as the jump direction explain variation in intraday returns following a jump event and document the effect over the trading day. We identify overnight jumps through highfrequency based jump tests and calculate our sentiment variables from the Thomson Reuters News Analytics dataset. We document our findings for S&P 500 constituents from 2004 to 2021. In the case of positive jumps, we document a stronger overreaction behaviour to both, the direction of the jump and to the prevailing news sentiment, whilst for negative jumps, we can only document a reversal behaviour relating to the direction of the jump. In addition, the paper presents a trading strategy based on the observed phenomena. The strategy exhibits no correlation to the market portfolio, exhibits tail hedging characteristics, whilst maintaining a positive drift component. Key words: stock return predictability, statistical jumps, private information, news sentiment, nontrading hour information, market level sentiment, company-specific sentiment

null

This bachelor’s thesis presents an examination of factors, such as the factors introduced by Fama and French (2014), as well as Carhart (1997). Additional factors are built using financial ratios. The research is done on US stocks from 1990 to 2021. The factor’s ability to explain asset returns is observed by doing two simple linear regressions. To visualize interrelations, a correlation matrix is analyzed. From the factors that are already established in existing literature, the excess market return, profitability, investment and momentum factors are good at explaining returns. Besides, factors built using one of these ratios have a significant influence: Price to earnings, free cash flow to operating cash flow, or interest coverage. Further, a factor model is constructed based on financially turbulent times. That is, taking the factors that had the best performance during drawdowns. When conducting a multi-linear regression and a test by Gibbons et al. (1989), it appears that this multi-factor model does not capture the expected returns accurately, therefore it is not using its full potential efficiency. This shows that when constructing a new factor model, the criteria for the factor choice should be extended.

null

This paper presents a portfolio construction approach that combines the hierarchical clustering of a large asset universe with the stock price momentum. On the one hand, investing in high-momentum stocks stabilizes portfolio performance across economic regimes and enhances risk-adjusted returns. On the other hand, hierarchical clustering of a high-dimensional asset universe ensures sparse diversification and mitigates the problems of increased drawdowns and large turnovers typically present in momentum portfolios. Moreover, the proposed portfolio construction approach avoids the covariance matrix inversion. An out-of-sample backtest on a non-survivorship-biased dataset of international stocks shows that hierarchical-momentum portfolios achieve substantially improved cumulative and risk-adjusted portfolio returns as well as decreased portfolio drawdowns compared to the model-free benchmarks net of transaction costs. Furthermore, we demonstrate that the unique characteristics of the hierarchical-momentum portfolios arise due to both dimensionality reduction via clustering and momentum-based stock selection.

In this thesis, the impact of recent armed conflicts on worldwide stock market indices is examined. The aim is to provide additional support to the findings of Schneider and Troeger (2006). This study focuses on two current conflicts, the Russian-Ukraine conflict and the ChinaTaiwan conflict. Using a one-year rolling GARCH(1,1) model, the reactions of 14 stock market indices to key events in these conflicts are analyzed. The findings indicate that conflict-related news tends to frequently impact North American markets, with European markets showing strong reactions to the developments of the Russia-Ukraine conflict. Asian markets appear to be less sensitive to geopolitical signals. While international stock markets are generally adversely affected by conflict situations, certain industry indices manage to thrive. Overall, the findings of this thesis support the idea of rational expectations in recent armed conflicts.

In this thesis, we examine 25 zero-investment long-short strategies derived from factors constructed using price, market capitalization, and trading volume of cryptocurrency assets to identify significant alpha generation. We employ various parsimonious factor models, incorporating market return, size, momentum, and market cycle to explain the observed alphas. Our findings reveal that a model encompassing all four factors can account for six out of the seven significant strategies. Additionally, we utilize a Fama-MacBeth Regression to demonstrate that the risk premia associated with market, size, and momentum exposures are not only significant over time but also exhibit greater strength and volatility during bullish market cycles.

This thesis analyses the impact of currency hedging on international bond portfolios from the point of view of a Swiss investor. Four of the most important regions including their currency were included in this analysis, namely the US Dollar, Euro, Pound Sterling, and Japanese Yen. The focus of the thesis is on the mitigation of risk stemming from the currency exposure and the comparison of different hedging approaches. The leading question of this thesis is whether a Cash-Flow-Matched hedging approach is superior to the benchmark used by most index providers or to no hedging at all. All hedging approaches use currency forward contracts as hedging instruments. The results show that the benchmark is the best hedging method among these three. Superiority is not only shown in terms of the mean return but also in terms of higher moments. The monthly structure of hedging and reporting forms the base of the success. The results are robust across all regions and bond-maturity buckets, showing that currency hedging is absolutely crucial in the setting of international bond portfolios. Additionally, more active hedging strategies were carefully introduced and compared to the three main approaches. An effort was undertaken to stay as close as possible to the philosophy of risk management. Overall, the results are less robust, pointing towards a slightly speculative nature of the introduced hedging strategies. Nonetheless, promising results were reported with rather simple quantitative trading rules, especially for the US region. A simple momentum approach is superior to all other approaches in terms of mean return. The development of the currency exchange rates including the recent turbulence underlines the importance of currency hedging. Throughout the considered period, the Swiss Franc was very strong compared to the currencies of the other regions, generally playing in the favour of hedging. The analysis is strongly affected by the point of view as well, which is tailored to an institutional Swiss investor such as a Swiss pension fund.

This paper introduces a non-Gaussian dynamic currency hedging strategy for globally diversified investors with ambiguity. Assuming that ambiguity of a typical investor can be measured from market data, we associate it to non-Gaussianity of financial asset returns and compute an optimal ambiguity-adjusted mean-variance (dynamic) currency allocation. Next, we extend the filtered historical simulation method to numerically optimize an arbitrary risk measure, such as the expected shortfall. The out-of-sample backtest results show that the derived non-Gaussian dynamic currency hedging strategy outperforms the benchmarks of constant hedging and dynamic hedging with Gaussianity for all base currencies and net of transaction costs.

Volatility plays a key role in various areas of the financial market. It is used as an input in risk- management and derivatives pricing models, and it is traded to express views on the future realized volatility. This paper employs a MIDAS regression to predict and analyze potential predictors of realized volatility. It shows that predictability is improved significantly by including macroeconomic and financial variables as predictors. Further, trading strategies that use the models’ predictions to trade volatility outperform the Unconditional strategy by reducing risk at the correct time.

This thesis adds to the literature of option return predictability by leveraging supervised classification methods that have been left untouched until now. Using 766,524 option-month observations of S&P500 call options between 1996 and 2021, I show that nonlinear machine learning algorithms also outperform linear ones in classifying option returns. Further, I find that the class that captures the highest positive option returns is hard to predict, which is why long-short portfolios that avoid this class generate the statistically significant and economically highest risk-adjusted returns. Expected volatility and the uncertainty about it seem to play a role in explaining the long-short returns of the portfolio that additionally avoids the negative tail. Across all machine learning models, option features seem to be by far the most important, but stock features should still be included. Finally, I show that it can be beneficial to include a transformer in the model stack as it belongs to the models that perform the best.

This thesis studies the behaviour of three types of Stop Losses - Point based, Percent based & Trailing Based. To simulate different trading strategies, five Technical Indicators are used to generate buy and sell signals. To capture a range of risk appetites, 3 risk reward ratios are used to fix the take profit thresholds across different stop loss values. The simulations are performed on the E Mini S&P 500 Futures and the results are analyzed based on 4 different metrics. Finally, we answer the following questions: is stop loss a useful technique in the intraday horizon? Which stop loss technique is better suited for this? Does an optimal stop loss exist?

null

null

The diversification gain (DG) measures the reduction in risk-based capital by bundling risks in a portfolio rather than holding them individually. We use the Value-at-Risk (VaR), Expected-Shortfall (ES), and Expectile to estimate the DG for Frechet distributed risks with different tail indices based on Monte Carlo simulations. This leads to the following results: The DG is lower for heavier tails. However, the DG is not significantly affected by the tail for dependent risks with finite second moment. The DG is higher for a higher number of risks, but with marginally decreasing effects. For dependent risks, the DG does not increase from a certain number of risks. Exposure limits significantly increase the DG for heavy-tailed risks. Lower limits lead to a higher DG. Higher dependencies decrease the DG. The behaviour of the DG for the ES and Expectile are similar. We observe several risk conditions where the VaR is increased by diversification. But the coherent risk measure ES and the Expectile always show positive diversification gains for risks with finite expectation. The VaR increase due to diversification is a measurement error rather than a risk increase. We also address the advantages of using the diversification gain by Burgi, Dacorogna, and Iles (2008) as a diversification measure. It allows for comparison across different distributions and is economically intuitive. The results of the DG approach are also compared to the expected utility framework. We also explore the conditions for negative diversification effects in the expected utility approach of Ibragimov, Jaffee, and Walden (2009).

The objective of this thesis is to examine viable return-generating trading strategies in the high frequency foreign exchange space using shifts in news sentiment momentum as signals. More specifically, this thesis backtests various moving average crossover strategies of hourly Euro-specific news sentiment using the TRNA dataset with over 70’000 observations. The crossover events serve as entry (exit) signals to assume long (short) positions in the EUR/USD exchange rate. In line with the findings of prior literature, an OLS regression analysis reaffirms the existence of a predictive relationship between Euro-specific news sentiment and the EUR/USD exchange rate. A backtesting exercise evaluates both simple and exponential moving average crossover strategies for three different combinations of crossover periods (2-10-hour, 2-20-hour, and 2-50-hour) with varying levels of transaction costs. The moving average crossover approach reduces the number of transactions and consequently brings down transaction costs. The results show that both sets of strategies outperform the buy-and-hold benchmark strategy and yield consistent and positive returns for the 2-20-hour and the 2-50-hour crossover periods, even after taking transaction costs into account.

Im Moment gibt es mit Kryptowahrungen einen grossen Boom bei Investoren. Dabei ist es wichtig zu wissen, ob uberhaupt Gewinn abgeschopft werden kann und falls ja, welche Strategie langfristig am erfolgversprechendsten ist. Diese Bachelorthesis gibt zu Beginn eine Einfuhrung in die Kryptowahrungswelt. Zudem gibt es einen Einblick wo und inwiefern Kryptowahrungen im nationalen und internationalen Zahlungsverkehr zur Anwendung kommen konnte und ob die Kryptowahrungen eine Chance haben die Fiat-Wahrungen zu bedrohen. Der Fokus dieser Arbeit liegt bei der Anwendung der Momentum sowie Buy and Hold Strategie im Jahr 2021. Das Ziel ist, herauszufinden welche dieser zwei Strategien eine hohere Rendite abschopft und mit wie viel Risiko dieser Gewinn verbunden ist. Obschon mit der Momentum Strategie einen hoheren Gewinn erzielt wurde, ist die Volatilitat bei der Momentum Strategie auch hoher.

null

Classic option pricing models are known to perform moderately well in standard market conditions. However, if this is not given, their performance is greatly reduced. By divid- ing the Covid-19 crisis into three periods (pre-, mid-, and post-crisis), we measure the pricing error of four classic option pricing models (Black-Scholes-Merton model (1973), Cox-Ross-Rubinstein (binomial) model (1979), (crude) Monte Carlo Simulation (1977), and Heston model (1993)) in contrasting market situations. We first state an empirical proof of the Cox-Ross-Rubinstein model and Monte Carlo Simulation’s convergence to the Black-Scholes-Merton model as proclaimed by Cox, Ross, and Rubinstein (1979) and Boyle (1977). Then, we compare the performance of the Black-Scholes-Merton and Heston model, where the results show, with an agreement to preceding studies, that the former is significantly outperformed in both in-sample and out-of-sample testing. Both models display an increase in error in the mid-period, during which we observe high volatility and many out-of-the-money options. The Black-Scholes-Merton model overprices options in all periods, whereas the Heston model only shows a tendency of overvaluing during the mid-period.