Leversens
← Back to algorithm

Regime indicator

A factor-level momentum filter applied to the strategy itself. The circuit-breaker that turns a quality long-only portfolio into a defensively timed exposure.

FULL exposure
100%

Strategy's trailing 3-month return is positive. Fully invested in the top quality + momentum candidates.

HALF exposure
50%

Trailing 3-month return turns negative. The other 50% sits in cash. Reactivates as soon as the trend recovers.

Why factor-level, not market-level?

Most "trend-following" filters look at the broad market index. This one looks at the strategy's own performance. The reason: quality-momentum may underperform in a market that is rallying (e.g. the junk rallies of 2009, 2020, 2023). Market-level filters would keep you fully invested in those exact moments where the strategy is losing relative ground.

Watching the factor's own drawdown is a sharper signal. It catches regime shifts within the strategy before they become deep drawdowns.

Empirical impact (2007-2026)

MetricWithout regimeWith regime
Annualized return14.1%13.7%
Annualized volatility18.5%16.1%
Sharpe ratio0.650.73
Max drawdown−38.7%−31.0%

Trading off ~40 bps of return per year for a 770 bps reduction in maximum drawdown and a meaningful Sharpe improvement.

A note on naming

We avoid calling this "market timing." The signal is endogenous, non-discretionary, and rule-based. It's closer in spirit to Moskowitz, Ooi and Pedersen's (2012) "Time-Series Momentum", applied to a single factor rather than asset classes.