What is the Square Root of Time Rule?

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

Square Root of Time Rule: Volatility scales with the square root of the horizon while returns scale linearly, so a daily standard deviation becomes an annual one by multiplying by the square root of the number of periods. The rule holds only for independent, identically distributed returns, and breaks in exactly the markets where the scaled number is most needed.

What the Square Root of Time Rule Actually Measures

The square root of time rule converts a volatility measured over one horizon into a volatility over another. A daily standard deviation of 1 percent becomes an annual figure of roughly 16 percent; a one-day value-at-risk becomes a ten-day regulatory number by multiplying by 3.16.

The rule exists because variance, not standard deviation, is the quantity that accumulates. Over independent periods variances add, so the standard deviation — being a square root — grows with the square root of elapsed time. Everything else in risk reporting follows from that single asymmetry.

Its importance is structural rather than technical. Volatility figures are quoted on one horizon and consumed on another: implied volatility is annualised, position limits are daily, regulatory capital is ten-day, and investment committees think in quarters. Every one of those translations passes through this rule, usually without comment.

How It Works

For returns that are independent and identically distributed, the variance over $T$ periods is the sum of $T$ single-period variances:

$$\sigma^2_T = T \cdot \sigma^2_1$$

Taking the square root gives the scaling law:

$$\sigma_T = \sigma_1 \cdot \sqrt{T}$$

Expected return, by contrast, is additive in time rather than in variance, so it scales linearly:

$$\mu_T = T \cdot \mu_1$$

That difference in exponent is the most consequential fact in the rule. Because return grows with $T$ and risk with $\sqrt{T}$, the ratio between them grows with $\sqrt{T}$ as well. A strategy with a barely detectable daily edge can have a respectable annual Sharpe ratio, and the same arithmetic annualises a Sharpe ratio by multiplying it by the square root of the number of periods per year.

Scaling factors between common horizons

Trading-day conventions for US equities. The daily-to-annual factor of 15.87 is the constant behind the desk shorthand known as the Rule of 16.

Worked Example

A portfolio delivers a daily standard deviation of 0.95 percent. Annualising:

$$\sigma_{\text{annual}} = 0.95% \times \sqrt{252} \approx 15.1%$$

The same input scales downward for a regulatory report. A one-day 99 percent value-at-risk of USD 4.2 million becomes a ten-day figure under the Basel scaling convention:

$$\text{VaR}_{10d} = \text{USD } 4.2\text{m} \times \sqrt{10} \approx \text{USD } 13.3\text{m}$$

And it scales upward for a committee that thinks in quarters. At 15.1 percent annualised, a quarterly one-sigma move is 15.1 / 2 = 7.6 percent — the number an allocator should have in mind before reading a quarterly drawdown as evidence of anything.

One portfolio, four horizons

Frequencies assume normally distributed, independent returns. Empirical markets deliver 2σ periods considerably more often, which is the first limitation below.

When It Applies (and Limitations)

Independence is the whole assumption. Variances add only when periods are uncorrelated. Any serial dependence in returns breaks the additivity, and the direction of the error depends on the sign of that dependence.

Trending markets understate risk. With positive autocorrelation, moves reinforce one another and multi-period variance exceeds the sum of single-period variances. Scaling a daily number up will then report less risk than the position actually carries — the error runs in the dangerous direction precisely during momentum-driven dislocations.

Mean-reverting markets overstate it. With negative autocorrelation, moves partially cancel and the scaled figure is too high. Range-bound rates and spread-trading books are the usual cases.

Volatility is not constant. Realised volatility clusters: quiet periods follow quiet periods and violent ones follow violent ones. Under a GARCH-type process the correct multi-period scaling depends on where current volatility sits relative to its long-run level, and the square-root rule is only accurate when the two coincide. Diebold and co-authors showed that naive scaling under such dynamics can be badly wrong in both directions.

Fat tails do not scale the same way. The rule scales a standard deviation, not a quantile of a fat-tailed distribution. Applying it to a value-at-risk number implicitly assumes the shape of the distribution is preserved across horizons, which it is not when jumps are present. Under jump-diffusion the scaled figure systematically understates tail risk at longer horizons.

Drift is ignored. The rule scales risk only. Over horizons long enough for expected return to matter, a risk figure scaled by $\sqrt{T}$ and a return figure scaled by $T$ must be handled separately — collapsing them into a single "annualised" statement conflates two different exponents.

Why It Matters for Institutional Investors

Regulatory capital. The Basel framework's ten-day horizon is in practice produced by scaling a one-day internal model by the square root of ten. The convention is explicit, so its assumptions become the institution's assumptions whether or not anyone re-examines them.

Cross-horizon comparison. A hedge fund reporting monthly volatility, a futures desk reporting daily and an options book quoting annualised implied volatility are describing risk in three languages. The rule is the dictionary, and mixing the conventions silently — annualising a monthly number with the daily factor, for instance — produces errors of several hundred percent that survive into board packs.

Option pricing and term structure. Black-Scholes takes a single annualised volatility and scales it to the option's maturity by exactly this rule. Observed implied volatility term structures depart from flat scaling, and the size of that departure is itself a tradable signal — the entire premise of calendar spreads and term-structure positioning.

Performance statistics. Sharpe ratios are annualised by multiplying by the square root of the observation frequency. Because the numerator scales linearly and the denominator by the square root, higher-frequency strategies show flattering annualised ratios from small per-period edges, and the statistic becomes harder to interpret the further it is scaled.

Risk budgeting. Limits set daily and reviewed quarterly are related by a factor of roughly eight. Governance that fixes both without reconciling them either constrains nothing at one horizon or paralyses trading at the other.

References

1. Diebold, F. X., Hickman, A., Inoue, A., & Schuermann, T. (1998). Scale Models. Risk, 11, 104–107.
2. Danielsson, J., & Zigrand, J.-P. (2006). On Time-Scaling of Risk and the Square-Root-of-Time Rule. Journal of Banking & Finance, 30(10), 2701–2713.
3. Basel Committee on Banking Supervision. Minimum Capital Requirements for Market Risk. (https://www.bis.org/bcbs/publ/d457.htm)
4. Sinclair, E. (2013). Volatility Trading (2nd ed.). Wiley.