Cross-asset time-series momentum strategy: A new perspective

Abstract

We propose a new investment strategy, the improved cross-asset time-series momentum (I-XTSM) strategy, to predict future returns and make investments. Using data on 25 investment portfolios and common commodities for the period from January 1990 to December 2023, we find that the I-XTSM strategy increases profitability substantially in the stock market and avoids momentum collapse effectively. We also document that its profitability is driven by the predictive power of the industrial metal assets' past signals. Even after considering transaction costs and market exposure, the I-XTSM demonstrates superior performance and explains the excess profits of other momentum strategies.