Andreas Thomann
Welcome! I am a lecturer at the Institute of Financial Services Zug (IFZ @ Lucerne University of Applied Sciences and Arts). I hold a PhD in Numerical Finance from the University of Zurich.
My research interests are quantitative finance, asset allocation, hedge fund and trading strategies as well as data science.
Publications:
Factor-based tactical bond allocation and interest rate risk management (Journal of Investment Strategies, 2019, Vol. 8, Issue 3, pp. 49-79)
Abstract: This paper offers two composite bond market factor investment strategies each for the Swiss bond market and for the global sovereign bond market. These composite factor strategies can be useful tools when making tactical asset allocation decisions between bonds and cash, and they can act as a base for the duration debate. As such, the output of our bond market factors can guide tactical interest rate views and therefore interest rate risk management. To construct these composite factors, we use four economically meaningful individual factors. Following an investment strategy based on a composite bond market factor, constructed as the equally weighted average of individual components, we are able to outperform cash as well as the static buy-and-hold strategy with regard to the Sharpe ratio, annualized standard deviation and maximum drawdown. Testing the composite and individual factors on their performance during periods of historical rising interest rates, we observe improved drawdown results compared with holding the underlying asset passively.
Multi-asset scenario building for trend-following trading strategies (Annals of Operations Research, S.I.: Recent Developments in Financial Modeling and Risk Management, 2020, Epub ahead of print.)
Abstract: This paper presents a new method for improving the performance of trend-following trading strategies. This new approach improves the inherent problem of trend-following strategies, which is their lagging signals. We simulate alternative price paths of financial assets using a modification of a distribution-free, semi-parametric approach that combines a GARCH-type process with historical simulation. These simulated price paths are used to construct and optimize trend-following trading strategies. The study is conducted in a multi-asset environment. Our empirical results demonstrate the superior performance for the MSCI World and Standard & Poor's 500 indices in terms of Sharpe ratio and maximum drawdown compared to widely applied trend-following trading strategies. The results are robust to variations in input specifications, such as tested time and look-back period, number of simulated price paths and price steps per simulation but also in terms of trading strategy calibration and market positioning (long only, long--short, short-only).
Is Trading Indicator Performance Robust? Evidence from Scenario Building (Journal of Investment Strategies, 2020, Vol. 9, Issue 1, pp. 1-22)
Abstract: This paper challenges widely applied trading indicators in their ability to generate reasonable signals. Assessing the robustness of a trading strategy is often done by either using a subsample of the data or applying the strategy on another asset's price series. Using the same configuration on another asset can be beneficial to show the underlying economic soundness of the trading strategy. Different statistical properties of the data series however can lead the developer to reject the strategy since it performs worse on the alternative data set. In this study we use a semi-parametric scenario building approach to simulate artificial price series based on the observed price characteristics. Our price simulations provide a backtesting environment to test trading strategies on simulated prices series with comparable properties as the observed price series. In addition to testing the indicators on the observed price series, we are able to test the indicators on a large set of simulated prices which describe the empirical distribution function of the specific asset. This provides an additional performance assessment and allows to test the trading indicators for robustness on a large set of artificially created price series with similar characteristics as the observed price series. We find that some momentum and trend-following trading indicators are robust whereas others, independent of their calibration, fail to deliver stable performance.