Drawdown Control

Investors don't experience volatility, they experience drawdowns, the sinking feeling of watching an account fall from its high point. That's the number that triggers redemptions, blows through risk limits, and makes people quit at the bottom (see Drawdown and Max Drawdown). Drawdown control is a family of rules that manage exposure based directly on that feeling: the deeper you are below your peak, the less risk you take, so that an ordinary losing streak can never compound into a ruinous one.

It differs from Vol Targeting, which scales exposure by how choppy markets are. Drawdown control scales by how far you've fallen, a path-dependent signal. You can have low volatility and a deep drawdown (a slow grind down) or high volatility with no drawdown (a violent rally), and the two overlays respond to entirely different things.

The rule

Track your high-water mark $H$, the highest portfolio value reached so far, and your current drawdown $D = (H - V)/H$, where $V$ is today's value. Then set risky-asset exposure as a decreasing function of the drawdown, cutting it to zero once losses hit a maximum tolerable level $D_{\max}$:

$$\text{exposure} = w_0 \cdot \max\!\left(0,\; 1 - \frac{D}{D_{\max}}\right),$$

where $w_0$ is your baseline (full-risk) allocation, $D$ is the current drawdown, and $D_{\max}$ is the drawdown you refuse to exceed. At the peak ($D = 0$) you run full exposure; as the drawdown deepens you scale down linearly; at $D = D_{\max}$ you're flat, in cash, so the floor holds. This is essentially CPPI and Portfolio Insurance measured from a moving high-water mark instead of a fixed floor, the idea behind Time-Invariant Portfolio Protection (TIPP).

Worked example

Baseline allocation $w_0 = 100\%$ risky, maximum drawdown budget $D_{\max} = 20\%$. Start at a peak of $100.

• At the peak, $D = 0$, exposure $= 100\% \times (1 - 0) = 100\%$. Fully invested.
• Down to $90, drawdown $D = 10/100 = 10\%$, exposure $= 100\% \times (1 - 10/20) = \textbf{50\%}$. You've halved risk.
• Down to $95 worth of value but held at 50% risk... suppose it keeps falling to $85: $D = 15\%$, exposure $= 100\% \times (1 - 15/20) = \textbf{25\%}$. Risk keeps shrinking.
• At $80, $D = 20\% = D_{\max}$, exposure $= 0$. You're flat; the drawdown can't get worse.

Now the market recovers. Because you de-risked on the way down, you participate only partially on the way up, at $85 you're back to 25% invested, at $90 to 50%, and you don't return to full risk until you reclaw back near the old peak. That lag is the whole cost of the insurance.

The fundamental trade-off

There's also a subtler cost: because the high-water mark only ratchets up, the strategy is asymmetric. Every new peak raises the bar you must stay above, so after a strong run the overlay becomes increasingly twitchy, small pullbacks from a fresh high now count against your budget. Some implementations soften this by measuring drawdown from a trailing peak (say the highest value in the last year) rather than the all-time high.

Where it fits

Drawdown control, CPPI and Portfolio Insurance, and Vol Targeting are the three standard rule-based risk overlays. Vol-targeting reacts to turbulence, CPPI protects a fixed floor, and drawdown control protects a moving high-water mark. None of them predict anything, they're mechanical responses to what's already happened, which is their strength (no forecasting needed) and their weakness (they always act after the fact). The right one depends on what you're actually trying to protect: a hard capital floor, a volatility budget, or the emotional and business reality that clients flee during deep drawdowns.