28 September

Style rotation on the JSE using machine learning

We would like to invite all enthusiasts and researchers to the next Kozminski University Finance Research Seminar Series event! When? September 28th 2022 at 12:15 PM Where? MS Teams Access (remote access). Details will be sent to the e-mail address provided once the registration is completed. The registration form is available below. Abstract: Page, McClelland and Auret (2021) identified short-term style momentum as an accurate predictor of short to medium term future style performance on the JSE. The study found that style momentum is exploitable within a factor rotation framework where several long-only strategies produced significant risk-adjusted Fama-French and Carhart alphas. The study specifically mentioned further avenues of future research, namely the application of ‘more’ advanced techniques for forecasting future style performance. The proposed study is therefore an extension of Page et al (2021) where historical style momentum and residual volatility is applied as features to forecast future style performance within a machine learning framework. The results indicate that machine learning algorithms, specifically gradient boosters, which are adept to mapping non-linear relationships between features and labels, are superior to style momentum when constructing style rotation strategies on the JSE.  

Special guest and speaker: Ph.D. Daniel Page

Daniel Page

Daniel Page is a senior lecturer in finance at the School of Economics and Finance, University of the Witwatersrand.

Daniel’s research focuses on systematic factor anomalies, asset pricing as well as machine learning and data science applications to finance and investment. Daniel is also the head of quantitative strategies as 27four Investment Managers where he is responsible for the management of all passive and smart beta mandates, manager research as well as systematic factor research across global and local equity and fixed income markets.