Realized Variance as a Momentum Crash Signal: Evidence and Limits

Summary

The document asks why realized variance of a momentum strategy might help avoid momentum crashes, particularly when markets rebound sharply after steep declines and the strategy’s short positions rise. It refers to a study that uses the sum of squared momentum-strategy returns as a predictor and questions whether that measure can identify an approaching rebound.

The response says the cited paper offers little theoretical explanation for why the approach works, relying instead on evidence that it performed well historically. It treats the strategy’s apparent improvement as a reason for caution and highlights uncertainty about whether the relationship will persist in future market conditions. The exchange provides no detailed predictive mechanism, implementation rules, or additional empirical results, so it supports skepticism and further evaluation rather than a settled account of how variance forecasts crashes.

Key ideas

• The cited approach uses realized variance of momentum-strategy returns as a predictor.
• Momentum crashes can occur when markets rebound and the strategy’s short leg rises sharply.
• The response finds little theoretical justification for the predictor in the discussed paper.
• Historical performance alone does not establish that the method will work in future crashes.
• The document leaves the predictive mechanism and future reliability unresolved.

Tags

# Realized variance as predictor that improves momentum strategy


# Realized variance as predictor that improves momentum strategy












In the paper "Momentum has its moments" (Pedro Barroso and Pedro Santa-Clara, 2012 - available free from Nova Business School), the authors claim that there is a way to avoid momentum crash (caused by the short leg as the market rebounded following large previous declines, in other words the short leg rise with the market and cause great loses). They using the realized variance computed as the sum of the squared returns of the momentum strategy as a predictor. The question is how and why the realized variance can predict the point where the market start to rebound after large declines?

## Answer by nbbo2 (score 3)

https://quant.stackexchange.com/a/36287

You are right, the authors provide no strong justification for why their method works. They just show that "it would have worked well in the past". But we should be skeptical how well it will work in the future, especially when you consider what a big improvement this simple change makes in the strategy; it seems a little too good to be true. This is a reasonable criticism of the paper. I suppose we will have to wait until the next big "momentum crash" and see what happens...

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.