Momentum strategies and market timing: Characteristics of cross-sectional and time-series momentum strategies in stock market

No Thumbnail Available
Journal Title
Journal ISSN
Volume Title
School of Business | Master's thesis
Ask about the availability of the thesis by sending email to the Aalto University Learning Centre oppimiskeskus@aalto.fi
Date
2018
Major/Subject
Mcode
Degree programme
Finance
Language
en
Pages
63 + 6
Series
Abstract
PURPOSE OF THE STUDY Focus of this thesis is to increase the understanding of momentum strategies, including traditional cross-sectional momentum as well as recently discovered time-series momentum, in the context of stock markets. More precisely, I provide a throughout analysis on the differences between the performance of the strategies concentrating on two of the most common momentum caveats: momentum crashes and poor Januaries. This analysis will help to understand momentum strategies better as well as provide tools for investors to optimize the use of momentum strategies. DATA AND METHODOLOGY My sample consist of U.S. non-micro-cap stocks from 1927 to 2016. I construct portfolios for cross-sectional momentum strategy utilizing the past relative performance of the stocks to sort them in winners and losers, and for time-series momentum strategy using the stocks own past performance (positive or negative excess return) to define winner and loser status. Both strategies use zero-cost equally weighted portfolios long in winners and short in losers. I run a series of statistical tests for both strategies focusing on general performance, bear market characteristics and January performance to compare their differences. In addition, I test if value-destroying Januaries are predictable. Based on this, I present an improved momentum strategy that utilizes optimal market timing based on past stock and market returns. RESULTS I find evidence of time-series momentum anomaly being present in U.S. stock market. Moreover, I find that time-series momentum strategy is able to avoid the severe momentum crashes of which cross-sectional momentum suffers from. Neither of the strategies is immune to January losses, but I find evidence of past market returns as a predictor for the disastrous Januaries. I present an improved momentum strategy that utilizes these findings. This strategy is able to nearly double the Sharpe ratio of conventional momentum strategies as well as generate high abnormal returns without an exposure to momentum crashes or poor Januaries. The performance of the strategy is robust throughout the sample period, against a benchmark momentum enhancing method, including transaction costs as well as internationally.
Description
Thesis advisor
Nyberg, Peter
Keywords
cross-sectional momentum, time-series momentum, market anomalies, market efficiency
Other note
Citation