Can Market Regimes Really be Timed with Historical Volatility?

21 Pages Posted: 23 Apr 2021 Last revised: 24 Apr 2021

See all articles by Richard McGee

Richard McGee

University College Dublin (UCD) - Michael Smurfit Graduate School of Business

Date Written: April 22, 2021

Abstract

Recent research findings suggest long-term investor utility benefits through scaling expected returns by recent realized volatility. We test for utility gains to volatility timing using a utility regime-based methodology to classify investor-specific market investment regimes based solely on recent realized volatility levels. Under this framework we find limited informational content in using recent realized volatility to forecast utility regimes for the market index. To reconcile our findings we replicate work by Moreira and Muir (2017) and find that their reported Sharpe ratio gains through volatility-managing the US market factor do not appear to be statistically significant. We find that their scheme under-performs buy and hold in terms of Sharpe ratio over 30 of the 70 twenty year sub-periods in our sample (58 out of 70 for an un-leveraged investor). Furthermore, the historical out-performance of volatility management for the market index is highly sensitive to the timing of re-balancing within a month, suggesting that the strategy may not be robust to the precise timing of key market events relative to volatility changes. Strategy adopters should be aware that this timing is not guaranteed to line up favorably over future investment periods.

Keywords: Volatility Timing

JEL Classification: G11

Suggested Citation

McGee, Richard, Can Market Regimes Really be Timed with Historical Volatility? (April 22, 2021). Available at SSRN: https://ssrn.com/abstract=3832340 or http://dx.doi.org/10.2139/ssrn.3832340

Richard McGee (Contact Author)

University College Dublin (UCD) - Michael Smurfit Graduate School of Business ( email )

Blackrock, Co. Dublin
Ireland

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
1,061
Abstract Views
2,398
Rank
43,765
PlumX Metrics