What Goes up Must Not Come Down - Time Series Momentum in Factor Risk Premiums

Abstract

I document significant time variation and predictability in a set of risk factors based on technical indicators. As these indicators are primarily designed to detect trends in asset prices, these findings imply substantial time series momentum in factor risk premiums. Specifically, risk premiums are significantly larger (lower) following recent uptrends (downtrends) in the underlying risk factor. A trend-based dynamic factor strategy, which uses the trend-based signals in order to lever up or down the risk factor, yields annual utility gains of up to 500 basis points, doubles the risk factors' Sharpe ratio, and more than halves their maximum drawdown. In a conditional asset pricing context, employing technical indicators as conditioning information substantially improves a model's explanatory power. Overall, my evidence poses several challenges to risk-based asset pricing theories and seems to be more in line with theories of sentiment and investors initial under- and subsequent overreaction to new information.