A risk-sensitive momentum approach to stock selection

Abstract

One of the main implications of Lo’s Adaptive Markets Hypothesis (2004, 2012, 2017) is that returns of virtually all assets can change over time. We present a local linear trend smoothing method by which this phenomenon can be captured empirically. Moreover, we introduce two localised, amended goodnessof- fit indicators capable of capturing both the direction and the continuity of recently observed price trends. Our related empirical investigation is based on a sample of 30 German blue-chip stock price series observed over a period of more than 16 years. Its results indicate that the use of these indicators as a stock-screening device can be a more useful means of identifying stocks with a superior risk/return profile than applying a conventional momentum strategy. The validity of this finding is underscored by statistical significance tests based on a Moving Blocks Bootstrap procedure.

Keywords

• Adaptive Markets
• Local Least Squares smoothing
• Moving Blocks Bootstrap