I wrote my thesis on this subject. Slow active allocation rules like selling 2x leveraged for 1x fund when hitting x volatility percentile and buying back in when we drop below the volatility percentile can GREATLY reduce downside risk.
OOS I managed to have near identical return to 2x's return, but shrunk the MaxDD from 78% to 52%
It's actually being reviewed currently, not sure how much info I could/should share online prior to it being validated by my thesis supervisors, however if there's anything specific you'd like to know I can tell you. My literature review basically delves into the mechanics that you highlighted in your post.
I can probably DM you the core content of the methodology/empirical results/implications, if you're really interested. But main takeaway : monthly signal (yes really, super slow), realized 21-day annualized vol percentile as the signal, very simple switching rules (but they are still a design choice, thereby introducing some bias at least), 5Y/1Y IS/OOS periods, 2005-2026 OOS (so we don't observe the effects of the dot com bubble, also a form of bias, though we do see GFC), and yeah as I said, the results :
- Annualized return of ≈23%
- MaxDD of 54%
- Max recovery time of 2.2 years (massive improvement over the 6.4 Max Recovery of static 2x ETF)
--> So same annualized return but stronger Calmar ratio than 2x ETF, which in my OPINION (not a fact) is waaaay more important than Sharpe or even Sortino ratio. People (or me, at least) don't care about their investments going up and down sequentially, people care about DRAWDOWNS. A 78% MaxDD is a death sentence for most investors.
I wrote my thesis on this subject. Slow active allocation rules like selling 2x leveraged for 1x fund when hitting x volatility percentile and buying back in when we drop below the volatility percentile can GREATLY reduce downside risk.
OOS I managed to have near identical return to 2x's return, but shrunk the MaxDD from 78% to 52%
Interesting. Can you share your thesis here?
It's actually being reviewed currently, not sure how much info I could/should share online prior to it being validated by my thesis supervisors, however if there's anything specific you'd like to know I can tell you. My literature review basically delves into the mechanics that you highlighted in your post.
I can probably DM you the core content of the methodology/empirical results/implications, if you're really interested. But main takeaway : monthly signal (yes really, super slow), realized 21-day annualized vol percentile as the signal, very simple switching rules (but they are still a design choice, thereby introducing some bias at least), 5Y/1Y IS/OOS periods, 2005-2026 OOS (so we don't observe the effects of the dot com bubble, also a form of bias, though we do see GFC), and yeah as I said, the results :
- Annualized return of ≈23%
- MaxDD of 54%
- Max recovery time of 2.2 years (massive improvement over the 6.4 Max Recovery of static 2x ETF)
--> So same annualized return but stronger Calmar ratio than 2x ETF, which in my OPINION (not a fact) is waaaay more important than Sharpe or even Sortino ratio. People (or me, at least) don't care about their investments going up and down sequentially, people care about DRAWDOWNS. A 78% MaxDD is a death sentence for most investors.
Great breakdown of volatility drag.
This risk is exactly why I manage my portfolio through a strict dual mandate to balance growth against decay.
I anchor my core in VT for unconstrained global equity exposure, while systematically deploying a covered call overlay like PAYG.
Relying on options premiums to extract consistent cash flow mitigates the exact performance drag you highlighted here.
Leverage can be a powerful tool… if you can stomach the drawdowns.
https://www.aqr.com/-/media/AQR/Documents/Insights/Journal-Article/Leverage-Aversion-and-Risk-Parity.pdf
Each to there own methods, I encourage everyone to juice up and hold LEFT I have 3 strategies
Quarterly entry’s
80/20
And a timing strategy.
All three are smashing it