Vol Targeting (Multi-Asset)
Target a stable portfolio volatility by scaling exposure to a diversified multi-asset basket inversely to realized vol.
Thesis (edge)
By scaling exposure so that portfolio volatility stays near a target (e.g. 10%), we avoid over-leveraging in calm markets and reduce exposure when vol spikes, improving risk-adjusted returns.
Where it works (regimes)
Works across regimes by design; smooths equity drawdowns when vol rises. Can underperform in steady bull markets (reduced exposure).
Signals
- ( \sigma_{20} ): 20-day realized vol of the portfolio (or proxy).
- Scale = ( \sigma_{\mathrm{target}} / \sigma_{20} ), capped.
Portfolio construction
Start with equal-weight or risk-parity basket. Apply vol scaling to total notional. Rebalance daily or weekly.
Risk model
Tail: vol spikes faster than we can de-lever. Use cap on leverage and fast vol estimator (e.g. 5d) as circuit breaker.
Costs & implementation
Turnover increases when vol moves. Use liquid instruments. Consider half-life for vol (e.g. 20d) to avoid over-trading.
Failure modes
Over-leveraging in low vol; lag in vol estimate; single-asset dominance in basket.
Our Notes & Suggestions
Shrink vol estimate (blend with long-term average) to avoid extreme scaling. Test with 10d and 20d windows. Cap leverage at 2x.
Our Notes & Suggestions
See the "Our Notes" subsection in the body above for practical guidance, gotchas, and best practices. Always validate regime assumptions and transaction cost assumptions before scaling.
Implementation Checklist
- Compute rolling 20d realized vol for basket
- Define target vol (e.g. 10%)
- Scale notional = target_vol / realized_vol
- Cap max leverage (e.g. 2x)