Thursday, November 26, 2020

The November momentum factor crash - Regime and crowd changes will shock


For the second time this year, the momentum factor experienced a performance crash. Quick reversals in market expectations will lead to these negative factor events. This is especially the case for momentum which buy in-favor price moves and sells out-of-favor securities. You are trading with the herd and when it changes, it can get ugly. 

Performance is worse because there is a loss on both the long and short positions. The long in-favor stocks lose and the short out-of-favor gain so there is little protection for investors. This event can happen even if there seems to be only modest changes in the overall market.

The folks at PremiaLabs in their weekly report show the impact of the post-election crash. The same impact was found by Venn in their weekly factor performance report. Investors should be aware that momentum factors usually do not have any risk management to stop these shocks. Additionally, the construction of the factor matters. If there are no restrictions in sector, industries, or correlation as well as infrequent rebalancing, there will be greater impact  on the created momentum factor. There is no one universal definition of the momentum factor.

At the same time as the momentum shock occurred, there were gains in size and value factors. The small and out-of-favor firms gained on news of a COVID vaccine and macro expectations for improved growth. The old story driven by continued lockdown was reversing, albeit this may change again as positive COVID cases explode to the upside. We may see more of these reversals in the coming weeks. 

Like other shocks, factors correlations moved to extremes. Focusing on one factor isolates risks, but also accentuate return extremes when a regime or crowd switches.


Difference in 2020 factor returns can be seen with the return difference between momentum and value. The gain in cumulative momentum versus value returns was over 60% only to see short-term declines of 15+% and 10% respectively in June and November. Investors have to be cognizant of the rationale for why one factor may outperform another and what will be the scenario for reversal. 


No comments: