Quant Analytics: from scratch to pro · Microstructure, execution and costs · Intermediate
Slippage and market impact: the square-root rule
Large orders move the price against themselves. Empirical studies find that this impact grows roughly with the square root of order size, so doubling the size costs far less than twice as much per share, yet far more than a small trade does.
In this lesson
- Separate the spread, slippage and market impact in the cost of a trade
- Estimate impact with the square-root formula and state its limits
- Show how impact scales when order size or volatility changes
- Turn an impact estimate into a cost in money and in basis points
Read first: Order books, spreads and depth, Volatility explained
A price on a screen is the price for a small order. A large order consumes the best offers, then the next ones, and the market adjusts to the buyer's presence. The result is that the average fill is worse than the price before the order. This cost is real, grows with size, and is often the difference between a method that works on paper and one that works live.
Three layers of cost
- Spread: the cost of crossing from bid to ask, paid even by a tiny order. See order books and spreads.
- Slippage: the gap between the price expected when the order was decided and the average price obtained. It includes the spread, impact and any drift while waiting.
- Market impact: the part of slippage caused by the order's own demand for liquidity. The price moves because you traded.
Impact has a temporary part, which fades once trading stops, and a permanent part, which stays because other participants infer information from the flow. See market impact.
The square-root rule
For orders worked over a day or so, a common finding is that average impact is not proportional to size. It grows more slowly. The usual form is:
impact = Y x sigma x sqrt(Q / V)
Here Q is the order size, V the typical daily volume, sigma the daily volatility, and Y a constant of order one that is estimated for each market. Tóth and co-authors (2011) describe this square-root dependence on data from large orders and offer a theory for it; Almgren and co-authors (2005) fitted a power law close to the three-fifths power of block size to temporary impact; Bouchaud and co-authors (2018) review the evidence. The classic model of Kyle (1985) predicts a linear form, which the data do not support for large orders.
Treat the formula as an empirical regularity, not an exact law. Y varies and the rule is fitted on averages.
Worked example
Daily volatility is 2%. You plan to trade 50,000 shares in a stock that trades 2,000,000 a day, so the order is 2.5% of daily volume. With Y = 1:
sqrt(0.025)is 0.1581.impact = 1 x 0.02 x 0.1581, which is 0.316%, or 31.6 basis points.
On 5,000,000 of stock that is 15,811 in cost. Now make the order four times larger. The square root doubles, so impact is 0.632%. Because there are four times as many shares, the total cost is 8.0 times higher, not 4. Larger orders are worse per share and much worse in total.
A small order of 0.25% of volume has an estimated impact of 0.100%. The spread, not impact, then dominates the cost.
Why a linear model misleads
The chart compares the square-root curve with a straight line fitted through the 2.5% point. The line understates the cost of small trades and overstates the cost of very large ones. The practical lesson is that impact is steep at first: even modest size relative to volume is not free.
Using the rule
- Measure
Qagainst the volume actually available in the window you will trade, not against a year's average. - Use the volatility of that window. Impact rises with volatility.
- Expect the real cost to be higher in stressed markets, when
Vfalls andsigmarises together. - Check the estimate against your own fills before relying on it.
Common mistakes
- Using one fixed slippage figure for every size and every day.
- Treating the formula as precise.
Yis uncertain by a wide margin. - Ignoring that a backtest on prices alone has no impact at all.
- Assuming the cost is only paid on entry. Exits pay it too.
- Applying the rule to very small orders or to instruments with no reliable volume figure.
In practice
Scale expected costs with size and volatility, and keep a record of realised fills to calibrate Y for each market. Analytics on FinUties applies measures like these to the markets it covers, so they are at hand when you read a chart. For schedules that spread a large order over time, see execution and implementation shortfall.
Impact against order size: square root and linear
- square-root rule (Y = 1, 2% volatility)
- linear model through the 2.5% point
Worked numbers
Each result is recomputed from the formula shown.
| What | Formula | Result |
|---|---|---|
| daily volatility | 0.02 | 2% |
| order as a fraction of daily volume | 50000 / 2000000 | 2.5% |
| square root of the fraction | sqrt(part) | 0.1581 |
| estimated impact, Y = 1 | 1 * sigma * root | 0.316% |
| impact in basis points | impact * 10000 | 31.6 basis points |
| impact cost on 5,000,000 of stock | impact * 5000000 | 15,811 |
| impact for four times the size | 1 * sigma * sqrt(4 * part) | 0.632% |
| total cost in money for four times the size | (impact_4x * 4 * 5000000) / cost_money | 8.0 times |
| impact for an order of 0.25% of volume | sigma * sqrt(0.0025) | 0.100% |
Check yourself
Daily volatility is 2% and an order is 2.5% of daily volume. What does the square-root rule with Y = 1 give?
Impact = Y x sigma x sqrt(Q / V) = 1 x 0.02 x sqrt(0.025) = 0.02 x 0.1581 = 0.316%, about 31.6 basis points.
The order is made four times larger. By how much does the impact per share grow?
By sqrt(4) = 2. Impact per share doubles. The total cost in money grows by 4 x 2 = 8 times, because there are four times as many shares.
Why is the formula an estimate and not a law of physics?
Y differs between markets and periods, the fit is on averages over many orders, and the rule describes orders executed over a day or so. For very small or very urgent orders it can be wrong.
A market is twice as volatile on a given day. What happens to estimated impact for the same order?
It doubles, because impact is proportional to sigma. The same order costs more in a volatile market.
A model has a linear impact of 1% for an order of 5% of volume. What does it predict for 20% of volume, against the square-root rule through the same point?
Linear: 4%. Square-root: 1% x sqrt(4) = 2%. Linear models overstate the cost of large orders and understate the cost of small ones.
Key terms
Sources
- Anomalous price impact and the critical nature of liquidity in financial markets. Physical Review X 1, 021006 (Bence Tóth, Yves Lempérière, Cyril Deremble, Joachim de Lataillade, Julien Kockelkoren, Jean-Philippe Bouchaud), 2011. Accessed 9 Nov 2026
- Equity market impact. Risk 18(7), 57-62 (Robert Almgren, Chee Thum, Emmanuel Hauptmann, Hong Li), 2005. Accessed 9 Nov 2026
- Trades, Quotes and Prices: Financial Markets Under the Microscope. Cambridge University Press (Jean-Philippe Bouchaud, Julius Bonart, Jonathan Donier, Martin Gould), 2018. Accessed 9 Nov 2026
- Continuous Auctions and Insider Trading. Econometrica 53(6), 1315-1335 (Albert S. Kyle), 1985. Accessed 9 Nov 2026