Allocating Capital vs Allocating Risk

Give two strategies $50m each, call it an even split, and walk away — that's capital allocation, and it's a trap. If strategy A runs at 20% annualized volatility and strategy B runs at 4%, the $50m/$50m split isn't even at all: strategy A is contributing roughly five times as much risk to the combined portfolio as strategy B, because risk scales with volatility, not with dollars deployed. A portfolio built this way is, in practice, almost entirely a bet on strategy A, with strategy B along for the ride as a rounding error. Risk allocation fixes this by sizing each strategy so it contributes a chosen share of portfolio risk, which usually means giving the low-volatility strategy much more capital, often with leverage, so its risk contribution matches its intended weight.

The arithmetic

A strategy's risk contribution is approximately its capital allocation times its volatility (ignoring correlation with the rest of the portfolio for a moment): risk contribution $\approx w \times \sigma$. To make two strategies contribute equal risk, their capital weights need to be inversely proportional to their volatilities: $w_A / w_B = \sigma_B / \sigma_A$.

Worked example. Strategy A (say, a directional macro strategy) runs at 18% annualized volatility; strategy B (a market-neutral equity strategy) runs at 3% annualized volatility. To give them equal risk contribution, the capital weights need to satisfy $w_A / w_B = 3/18 = 1/6$ — strategy B needs six times the capital of strategy A for the same risk. On a $700m risk-allocated portfolio split this way, strategy A gets $100m and strategy B gets $600m. Since strategy B is a market-neutral strategy, $600m of capital deployed at, say, 3x gross leverage means $1.8bn of gross notional traded — a specific, large operational number that a pure capital allocation would never have surfaced, because $600m "feels like" a much bigger bet than $100m even though the two are risk-equivalent by design.

equal risk A: $100m B: $600m risk: balanced

Why this matters beyond the arithmetic

Risk-based allocation forces an honest conversation about leverage that capital-based allocation hides. Strategy B in the example above needs real leverage and real operational capacity — financing lines, prime broker relationships, margin — to deploy $600m, and if the platform can't actually supply that leverage, the "equal risk" plan on paper doesn't translate into equal risk in the book. This is the same logic behind Risk Parity applied one level up, from asset classes to entire strategies, and it's the standard framework multi-strategy platforms use to size pods (see The Multi-Strategy Platform Model): a pod's risk budget, not its capital, is the number that's actually managed day to day.

What erodes it

Volatility is estimated from a finite, noisy history, so a strategy's "true" risk contribution drifts as its recent volatility regime changes — a market-neutral strategy that's been quiet for two years can suddenly spike in volatility during a stress event, and if the capital allocation was sized to its old, low volatility, its risk contribution jumps well above its intended share exactly when the portfolio can least afford it. Risk allocation also assumes the volatility and correlation estimates feeding the sizing formula are trustworthy, which is a much stronger assumption in calm markets than in the ones that matter most.