---
title: Square Root Law in Scaling Relations
url: https://www.emergentmind.com/topics/square-root-law-srl
type: topic
---

# Square Root Law in Scaling Relations

Searching arXiv for recent and foundational papers on “Square Root Law” across the domains represented in the source material.
{"query":"all:\"square-root law\" OR all:\"square root law\"","max_results":10,"sort_by":"submittedDate","sort_order":"descending"}
{"query":"id:1602.03043 OR id:2411.13965 OR id:2502.16246 OR id:2606.16269 OR id:2607.04280 OR id:2606.24019 OR id:2506.07711 OR id:2502.17906 OR id:2311.11279 OR id:2005.00913 OR id:1405.7657 OR id:2405.13248 OR id:1210.4758","max_results":20,"sort_by":"relevance","sort_order":"descending"}
Square Root Law (SRL) is a generic label for scaling relations in which a response variable grows with the square root of a size, volume, load, or frequency parameter. In current arXiv usage, the term is most developed in market microstructure, where the impact of a metaorder is proportional to the square root of its normalized size, but it also denotes the classical square-root safety rule in staffing, optimal survival rules in Beverton–Holt fisheries, square-root cancellation bounds in Fourier analysis over finite rings, and the Penrose square-root rule in two-tier voting [1602.03043] [2311.11279] [1405.7657] [1210.4758]. The shared structure is a concave exponent \(1/2\); the underlying objects, mechanisms, and interpretations are domain-specific.

## 1. Principal meanings of the term

The phrase “Square Root Law” does not denote a single theorem across disciplines. In the literature represented here, it names several mathematically analogous but substantively distinct regularities.

| Domain | Canonical SRL form | Interpreted quantity |
|---|---|---|
| Market impact | \(I(Q)=Y\,\sigma\,\sqrt{Q/V}\) | Metaorder price impact |
| Staffing | \(s \approx m + z\sqrt{m}\) | Safety capacity under Poisson arrivals |
| Fisheries | \(\gamma^* = 1/\sqrt{\rho}\) | Optimal survival under Beverton–Holt MSY |
| Finite-ring harmonic analysis | \(|\widehat{V}(\psi)| \le C |R|^{-d}|V|^{1/2}\) | Square-root Fourier cancellation |
| Two-tier voting | \(w_i \propto \sqrt{N_i}\) | Fair voting weights under independence |

In market microstructure, SRL is a law of concave execution costs. In staffing, it is a “base plus safety” rule derived from Poisson fluctuations. In fisheries, it links proliferation and optimal survival. In finite-ring harmonic analysis, it is a Salem-type decay estimate. In voting theory, it is a rule for balancing individual influence across constituencies. This suggests that SRL is best understood as a recurrent scaling template rather than a unitary concept.

## 2. Market-impact SRL in equities, futures, and options

In the standard market-microstructure formulation, the SRL states that the average impact of a metaorder of signed volume \(Q\) is proportional to the square root of the traded volume fraction \(Q/V\), scaled by daily volatility \(\sigma\):
\[
I(Q) = Y \,\sigma\, \sqrt{\frac{Q}{V}}.
\]
The corresponding implementation shortfall is
\[
S(Q) = \int_0^Q I(q)\,dq = \frac{2}{3}Y\sigma \sqrt{\frac{Q}{V}}\cdot Q.
\]
Here \(Q\) is signed traded volume, \(V\) is daily traded volume, \(I(Q)\) is the average impact between metaorder decision time and completion, and \(S(Q)\) is the quantity-weighted cost accumulated during execution [1602.03043].

For options, the same structure is written in “volatility space.” The paper "The square-root impact law also holds for option markets" defines metaorder size as net vega \(Q_\nu\), normalizes it by market-wide gross vega \(V_\nu\), and rescales cost by volatility-of-volatility \(\sigma_\sigma\). The option-market implementation shortfall is
\[
S(Q_\nu)=\frac{2}{3}\,Y_{\mathrm{vol}}\,\sigma_\sigma\,\sqrt{\frac{Q_\nu}{V_\nu}}\cdot Q_\nu,
\]
with normalized variables
\[
s := \frac{S}{Q_\nu \sigma_\sigma}, \qquad \phi := \frac{Q_\nu}{V_\nu},
\]
and empirical fit
\[
s = a\,\phi^\delta + b.
\]
In that formulation, \(b\) captures spread costs and/or execution alpha, and \(Y_{\mathrm{vol}}=3a/2\) [1602.03043].

The option dataset was proprietary Capital Fund Management data from August 2013 to January 2016, covering 450,000 metaorders across options on more than 1000 single US stocks. Metaorders were defined as the net vega traded by CFM on a given day on a given underlying, after delta-hedging, aggregated across strikes and maturities for the underlying. The data were split into short term options with maturity \(\le 3\) months and long term options with maturity \(>3\) months, with roughly half of the metaorders in each bucket [1602.03043].

The estimated coefficients were
\[
a_{LT} \approx 0.33,\quad b_{LT} \approx -0.013,\quad a_{ST} \approx 0.40,\quad b_{ST} \approx -0.071,
\]
with exponents from running averages \(\delta_{LT} \approx 0.40\) and \(\delta_{ST} \approx 0.43\). The inferred prefactors were \(Y_{\mathrm{vol}} \approx 0.50\) for long-term options and \(\approx 0.60\) for short-term options, in the same range as the \(Y\) constant for stocks and futures, reported as \(0.5 \to 1\). The fits were adequate for \(\phi\) between \(10^{-3}\) and \(1\), and the paper reported compatible but noisier results for implied volatility of futures contracts on a smaller sample of \(\sim 10{,}000\) metaorders [1602.03043].

## 3. Empirical evidence, universality claims, and contested interpretations

The most direct recent universality claim is the complete Tokyo Stock Exchange survey in "Does the square-root price impact law belong to the strict universal scalings?" The study used complete order lifecycle data for all stocks over eight years, split into 2012-01-04 to 2015-09-18 and 2015-09-24 to 2019-11-02 because of the arrowhead matching-engine change. Liquid stocks were defined as those with more than \(10^5\) metaorders, yielding 942 stocks in Dataset 1, 1,357 in Dataset 2, and 2,299 stock-level datapoints in total. At the trader level, 1,293 active traders satisfied the paper’s density requirements [2411.13965].

After normalizing \(Q \to Q/V\) and \(I \to I/\sigma\), filtering metaorders with execution horizon \(\ge 60\) seconds, and fitting \(I(Q)=cQ^\delta\), the stock-level distribution had mean exponent \(\delta = 0.489 \pm 0.0015\) with standard deviation \(0.071\). A finite-sample null model with exact \(\delta=1/2\) produced \(\bar{\delta}=0.489 \pm 0.0013\) and dispersion \(0.062 \pm 0.0010\), leading to the bias-corrected estimate \(\delta_{\text{unbiased}} = 0.500 \pm 0.0020\). At trader level, the mean was \(\delta^{(i)} = 0.493 \pm 0.0050\) with standard deviation \(0.177\), again close to the null-model dispersion.

Source: https://www.emergentmind.com/topics/square-root-law-srl