DOES MOMENTUM TRADING REALLY WORK?

After my last post concerning my momentum signals: I was asked by a skeptic why I thought Momentum trading works even though some studies (like the original one by Jegadesh/Titman) have shown the returns eventually can reverse.

My response was that some studies look at Momentum returns (i) agnostic of the current point in the cycle (without testing performance in distinct macro regimes) and/or these studies (ii) use raw momentum scores and weightings instead of more sophisticated methods.

To prove these two points (and that my signals outperform): I decided to test the raw momentum approach vs my bespoke signals on a longer time period. In the test I do not allow shorts nor leverage and I compare the weekly rebalance performance with fees to a buy/hold strategy.

The Momentum factor is thought by many to not do well in bear markets - my tests show this is the case for a ‘raw’ momentum approach, but my bespoke signals show a different story – they provided bear market outperformance. Why? Because they set the ‘losers’ to low/zero weight.

Pic 1 (upper left) – shows the ‘raw’ momentum approach (small caps) works in some years sporadically and needs to be switched (tilted) on/off frequently (However, a different rebalance frequency may yield different results admittedly)

Pic 2 (upper right) – shows that my bespoke momentum signals (small caps) provide consistent outperformance in all market regimes

Pic 3 (below) – Shows a more advanced proprietary model – which uses momentum and volatility and has only seen data before 2010, nonetheless is able to strongly outperform from 2010-2024 (small caps)