What is Volatility Drag?

By EC Assets Research Team · Published 17 August 2026 · Updated 17 August 2026

Volatility Drag: The gap between the average of a return series and the return actually compounded. Because a loss applies to a smaller base than the gain before it, dispersion alone reduces terminal wealth: the shortfall is approximately half the variance, so it grows with the square of volatility.

What Volatility Drag Actually Measures

Volatility drag is the gap between the average of a series of returns and the return the investor actually compounded. A portfolio that gains 30 percent in one year and loses 20 percent in the next has an average annual return of 5 percent and a compounded return of 1.98 percent. Nothing was withdrawn and no fee was charged; the two numbers simply answer different questions about the same two years.

The effect follows from the arithmetic of percentages. A loss is applied to a smaller base than the gain that preceded it, so a symmetric pair of percentage moves does not return the portfolio to where it started. The wider the dispersion of returns, the larger the gap between the average and the outcome — which is why the drag is a function of volatility rather than of direction.

The distinction runs underneath most of the portfolio literature. Expected returns, Sharpe ratios and mean-variance optimisation are stated in arithmetic terms; wealth, drawdown recovery and terminal capital are geometric. Volatility drag is the bridge between the two, and misreading which of them a figure refers to is one of the more expensive category errors in performance analysis.

How It Works

For a series of returns compounded over $n$ periods, the geometric mean is the constant rate that produces the same terminal wealth:

$$1 + g = \left[\prod_{t=1}^{n}(1 + r_t)\right]^{1/n}$$

Jensen's inequality guarantees that this quantity never exceeds the arithmetic mean, and equals it only when every return is identical. The shortfall is approximated by half the variance:

$$g \approx \mu - \frac{\sigma^2}{2}$$

Under a lognormal return distribution with arithmetic mean $\mu$ and volatility $\sigma$, the relationship is exact:

$$1 + g = \frac{1 + \mu}{\sqrt{1 + \dfrac{\sigma^2}{(1 + \mu)^2}}}$$

The half-variance term explains the shape of the effect. Drag rises with the square of volatility, so it is negligible for a short-duration credit book and dominant for a levered or single-name equity position. Doubling volatility quadruples the penalty.

Compounded return at an 8 percent average, by volatility

Exact column uses the lognormal identity above. The approximation is accurate to a few basis points below 15 percent volatility and unusable above 40 percent, where it understates the compounded return by several percentage points.

Two properties are worth reading off the table. First, a portfolio with a positive average return can compound to a negative one: at 50 percent volatility, an 8 percent average produces a loss. Second, the familiar half-variance shortcut is itself an approximation, and it fails in exactly the high-volatility regime where the drag matters most.

Worked Example

An allocator reviews two managers over the same twenty-year window. Both averaged 8 percent a year. Manager A ran at 10 percent volatility, Manager B at 30 percent.

Manager A compounded at 7.54 percent, turning one unit of capital into 4.28. Manager B compounded at 4.06 percent, turning the same unit into 2.22. The naive figure — 8 percent averaged over twenty years — implies 4.66.

Terminal wealth per 1.00 invested, 20 years, 8 percent average return

Terminal wealth is $(1+g)^{20}$ using the exact compounded return. Shortfall is measured against the 4.66 that the arithmetic average appears to promise.

Manager B delivered the same average return and roughly half the wealth. No fee, no tracking error and no timing decision accounts for the difference; it is entirely the cost of dispersion around the mean.

When It Applies (and Limitations)

Drag is not a fee. This is the most common misreading. Nothing is deducted from the portfolio, and there is no counterparty on the other side of the loss. The arithmetic mean and the geometric mean are two different summaries of one return series, and the gap between them is a property of the series, not a payment. A strategy cannot hedge its volatility drag; it can only reduce its volatility, which changes the return series itself.

It requires multiple periods. In a single period there is no compounding and therefore no drag. The effect accumulates only when returns are chained.

The half-variance form is an approximation. It is exact for continuously compounded (log) returns, where the arithmetic mean of log returns is the geometric growth rate. Applied to simple returns it degrades as volatility rises, as the table above shows.

The lognormal identity is a model, not a fact. Real return series are skewed and fat-tailed. For a distribution with meaningful negative skew, the realised compounded return will fall short of what the lognormal formula predicts from the same mean and volatility.

Estimation error propagates. Volatility measured over a finite sample is noisy, and drag scales with its square. A twenty percent estimate that is really twenty-five percent understates the drag by roughly half a percentage point a year — a material error over a fund's life.

Reporting standards already handle it. Composite performance reported under GIPS is time-weighted and geometric by construction, so a published annualised return is a compounded number. The gap appears when annual figures are averaged arithmetically further down the chain — in a summary table, a pitch deck, or a spreadsheet built from a factsheet.

Why It Matters for Institutional Investors

Manager comparison. Two managers with identical expected returns and different volatilities are not equivalent propositions for a compounding pool of capital. The higher-volatility manager needs a higher arithmetic mean to deliver the same terminal wealth, and the required premium is roughly half the difference in variance.

Volatility targeting. Scaling exposure to hold realised volatility near a constant is often defended as a risk-management measure. Its arithmetic justification is narrower and more precise: for a return stream whose mean does not rise proportionally with its volatility, suppressing variance raises the compounded return even when it leaves the average untouched.

Leverage. Applying a constant leverage factor $L$ to a return stream multiplies the mean by $L$ but the variance by $L^2$. The drag on a levered position therefore grows faster than its expected return, at approximately $\tfrac{1}{2}(L^2 - L)\sigma^2$ per period. On a 20 percent volatility index, two-times daily leverage carries roughly 4 percent of annual drag — the arithmetic behind the long-run decay of levered and inverse exchange-traded products.

Diversification and rebalancing. Because drag depends on portfolio volatility rather than on the volatility of the holdings, a rebalanced portfolio compounds faster than the weighted average of its components' compounded returns, even when no component outperforms. This diversification return, documented by Booth and Fama, is volatility drag read in reverse: the portfolio suffers less of it than its parts do.

Drawdown asymmetry. The same arithmetic governs recovery. A 30 percent loss requires a 43 percent gain to break even, and a 50 percent loss requires 100 percent. Drag and drawdown recovery are two views of one property of multiplicative returns.

Position sizing. Growth-optimal sizing follows directly. Maximising the geometric rate rather than the arithmetic one produces an optimal fraction of roughly $\mu/\sigma^2$ — the Kelly result — and explains why growth-oriented capital allocations fall as volatility rises rather than tracking expected return alone.

References

1. Fernholz, R., & Shay, B. (1982). Stochastic Portfolio Theory and Stock Market Equilibrium. Journal of Finance, 37(2), 615–624.
2. Booth, D. G., & Fama, E. F. (1992). Diversification Returns and Asset Contributions. Financial Analysts Journal, 48(3), 26–32.
3. MacLean, L. C., Thorp, E. O., & Ziemba, W. T. (2011). The Kelly Capital Growth Investment Criterion. World Scientific.
4. CFA Institute. Global Investment Performance Standards (GIPS) for Firms. (https://www.cfainstitute.org/en/ethics-standards/codes/gips-standards)