Time series momentum: theory and practice

Thumbnail Image
OBrienJJ_PhD2017.pdf(1.76 MB)
Full Text E-thesis
O'Brien, John
Journal Title
Journal ISSN
Volume Title
University College Cork
Published Version
Research Projects
Organizational Units
Journal Issue
Time series momentum (TSM) is a significant component of many investment strategies, both explicitly and implicitly. While academic studies have confirmed long run excess return, other aspects of the strategy have received less attention. This research focuses on performance variations across economic conditions and on the return drivers of TSM and associated investment funds. The performance record is extend back to 1925, confirming long run performance and providing a large sample to analyse its relationship with economic conditions, with a number of links demonstrated. TSM underperforms for periods of up to four years immediately after financial crises, with returns at less than half the level of normal periods. A breakdown in market structure is associated with this. Serial auto-correlation of asset returns, found in all markets at horizons of up to twelve months, is absent in the years following a crisis. Further evidence links the strategy to the business cycle, demonstrating underperformance in periods of recession and high economic uncertainty. A decomposition of returns of individual asset into systematic (macro-economic factor related) and idiosyncratic components shows that TSM generates profits from both components. The exposure of TSM to economic factors can explain part of the excess returns in an efficient market/arbitrage pricing theory framework. The performance of the commodity trading advisor (CTA) sector, closely associated with TSM, is analysed using a sample of 3,419 CTAs. A novel methodology eliminates biases and generates a reliable performance index back to 1987. This exhibits consistent excess returns over the period. Eight different CTA sub-strategies identified, exposed to a variety of risk premia in addition to TSM. Explanatory power is low with less than half of the returns associated with risk exposure. Three of the eight sub-strategies (representing half the funds) show exposure to TSM, each generating a statistically significant Sharpe ratio.
Time series momentum , CTA , Investment , Trend Following , Financial Crises , Strategy
O'Brien, J. 2017. Time series momentum: theory and practice. PhD Thesis, University College Cork.