This article introduces a discrete false discovery rate (DFRD+/-) method for controlling data snooping. We investigate with DFRD+/- the performance of dynamic portfolios constructed upon over 21,000 technical trading rules on 12 categorical and country-specific markets over the study period 2004-2017. The profitability, persistence and robustness of the technical rules are examined. We note that frontier markets exhibit higher average yield compared to the developed and emerging markets. A cross-validation exercise highlights the importance of frequent rebalancing and the variability of profitability in trading with technical analysis.
Cass Business School, 106 Bunhill Row
106 Bunhill Row, London EC1Y 8TZ, UK
University of Glasgow
Georgios Sermpinis in Professor of Finance at the Adam Smith Business School of the University of Glasgow. He holds a PhD in the Quantitative Finance and Forecasting from Liverpool John Moores University, a MSc in Banking and Finance from Forecasting from Liverpool John Moores University and a BSc (Hons) in Mathematics from the National and Kapodistrian University of Athens.
His mains interests are in the fields of Forecasting, Quantitative Finance, Machine Learning and Operations Research. He has published more than 30 papers in peer-review journals and his research has been cited more than 500 times. Georgios has offer consultancy for major banks such as Santander and Goldman Sachs. He has also co-author a book in Machine Learning and acts as editor for journals such as the Decision Support Systems and the Information Systems and Operational Research. Georgios has also edit special issues for the Quantitative Finance, the Journal of Forecasting, International Journal of Finance and Economics and Annals of Operations Research.