In this report, we show that instrument diversification is more important than instrument selection in the context of trend-following.

Relying on Quantica’s generic trend-following strategy applied to a diversified investment universe of 83 liquid futures markets across main asset-classes, we start with highlighting how trend return opportunities across individual instruments and main asset-classes have been varying over time. We then empirically quantify the maximum theoretical trend opportunity set that can be achieved through perfect instrument selection (that means with perfect foresight) over different time periods.

For that purpose, we evaluate the rolling cross-sectional dispersion of trend-following returns between profitable and unprofitable universe constituents over time.

We further quantify, over different time periods, the distribution of the maximum in-sample Sharpe ratio that can be achieved by constructing a trend-following portfolio with a variable number of target constituents from 1 to the original size of the universe.

While we demonstrate a historically high and persistent opportunity set for instrument selection on an in-sample basis, we also show a lack of persistence in the cross-sectional outperformance of individual or group of instruments over time.