On this page
Amihud Illiquidity Measure: Price Impact from Daily Volume
The Amihud illiquidity measure, known as ILLIQ, estimates how much a stock's price moves per dollar of trading. A higher ILLIQ means a stock is less liquid, and historically, less liquid stocks have earned higher average returns.
Key Takeaways
- ILLIQ is the daily average of the absolute return divided by dollar volume, requiring only standard OHLCV data to compute.
- Amihud's 1964-1997 NYSE study showed ILLIQ significantly predicts expected returns, confirming an illiquidity risk premium.
- ILLIQ is a historical average over a window, not a forecast for the next large order; execution desks need a calibrated impact model instead.
- Factor investors add an ILLIQ tilt to earn the liquidity premium, but transaction costs in the least-liquid names can easily offset the premium.
Key Takeaways
- ILLIQ is the daily average of the absolute return divided by dollar volume, requiring only standard OHLCV data to compute.
- Amihud's 1964-1997 NYSE study showed ILLIQ significantly predicts expected returns, confirming an illiquidity risk premium.
- ILLIQ is a historical average over a window, not a forecast for the next large order; execution desks need a calibrated impact model instead.
- Factor investors add an ILLIQ tilt to earn the liquidity premium, but transaction costs in the least-liquid names can easily offset the premium.
What It Is
Yakov Amihud introduced the measure in his 2002 paper "Illiquidity and Stock Returns" in the Journal of Financial Markets. The idea is to build a liquidity proxy from data every investor already has: daily returns and daily dollar volume. No tick data, no order book, no trade signs.
ILLIQ is a rough but durable price-impact estimate. Across NYSE stocks from 1964 to 1997, Amihud showed that ILLIQ has a positive and significant effect on expected returns. Investors demand compensation for holding names that are costly to trade.
The Intuition
If moving a stock's price by 1 percent takes $100 million of volume, it is liquid. If the same 1 percent swing happens on $2 million, it is not. ILLIQ compresses that relationship into one number per day, then averages across days.
The measure is popular because it works on long historical samples where high-frequency data is unavailable or unreliable. It has been used in asset pricing tests, portfolio construction, and corporate-finance research on the liquidity premium. It is not a substitute for an order book, but as a cross-sectional ranking tool it is hard to beat for its simplicity.
How It Works
The core formula, averaged over D days in a period (often a month or a year):
ILLIQ_i = (1 / D) * sum over t of ( |R_it| / DVOL_it )
Where:
R_it = return of stock i on day t
DVOL_it = dollar trading volume of stock i on day t
D = number of trading days in the period
The units are return per dollar. Most papers scale by 1,000,000 so the numbers are readable. Because volume appears in the denominator, stocks with thick trading get small ILLIQ values, and thin names get large ones. The absolute value of return treats up moves and down moves symmetrically, since liquidity cost should not depend on direction.
Amihud shows ILLIQ is highly correlated with more expensive measures like the effective spread and Kyle's lambda, which is why it survives as a proxy when better data is missing.
Worked Example
Consider a small-cap stock with five trading days. Returns and dollar volumes are:
| Day | Return | Dollar volume |
|---|---|---|
| 1 | +1.2% | $3,000,000 |
| 2 | -0.8% | $2,500,000 |
| 3 | +0.5% | $4,000,000 |
| 4 | -2.1% | $2,000,000 |
| 5 | +0.9% | $3,500,000 |
Daily ratios |R| / DVOL (in returns per dollar):
0.012 / 3,000,000 = 4.00e-9
0.008 / 2,500,000 = 3.20e-9
0.005 / 4,000,000 = 1.25e-9
0.021 / 2,000,000 = 1.05e-8
0.009 / 3,500,000 = 2.57e-9
Average: approximately 4.32e-9, or 4.32 when scaled by 1,000,000.
Compare this to a large-cap with ILLIQ near 0.05 on the same scale. The small-cap is about 85 times less liquid by this proxy. A portfolio manager would expect materially larger market-impact costs and, on average, a higher required return for holding it.
Common Mistakes
-
Comparing ILLIQ across markets without care. ILLIQ depends on return and volume conventions, scaling factors, and currency. A US equity ILLIQ is not directly comparable to a London or Tokyo number unless both are constructed identically.
-
Treating ILLIQ as a live market-impact model. ILLIQ is a historical average. It describes typical conditions over a window, not what will happen to your next large order. For execution sizing, use a calibrated impact model instead.
-
Forgetting the stock-split and dividend adjustments. Raw returns must be total-return, split-adjusted. Otherwise a single corporate action can blow up the numerator or denominator for a day and contaminate the monthly average.
-
Ignoring the zero-volume problem. Some names have zero-volume days. Dividing by zero is meaningless. Most implementations drop those days, winsorize extreme ratios, or aggregate over longer windows to cushion the issue.
-
Reading the illiquidity premium as a free lunch. The fact that illiquid stocks earned higher returns historically does not mean tilting toward them is costless. Those same stocks are harder and more expensive to trade, especially when you need to exit.
Frequently Asked Questions
Q: What is the Amihud illiquidity measure in simple terms? ILLIQ is the daily average of absolute return divided by dollar volume, giving a number that rises when a stock moves a lot on thin trading and falls when a stock barely moves despite heavy volume.
Q: How does the Amihud illiquidity measure affect investment decisions? Factor investors use it to tilt toward higher-ILLIQ stocks to capture the historical illiquidity premium, and risk managers use it to estimate which holdings will be hardest and most expensive to exit during a stress event.
Q: What is a real-world example of the Amihud illiquidity measure? In Amihud's original 1964-1997 NYSE study, a small-cap stock might show ILLIQ of 4.32 versus a large-cap at 0.05 on the same scale, an 85x difference that predicts significantly higher required return for holding the less liquid name.
Q: How can investors use the Amihud illiquidity measure? Sort the investable universe by ILLIQ each month, tilt the portfolio toward higher-ILLIQ deciles, but cap position sizes in the most illiquid names to avoid paying more in transaction costs on entry and exit than the premium earns.
Q: How is the Amihud illiquidity measure different from Kyle's lambda? Lambda requires trade-level tick data and regression estimation. ILLIQ uses only daily OHLCV data available going back decades, making it practical for long cross-sectional asset pricing studies where tick data is unavailable.
Sources
- Amihud, Y. (2002). "Illiquidity and Stock Returns: Cross-Section and Time-Series Effects." Journal of Financial Markets, 5(1), 31-56. https://www.cis.upenn.edu/~mkearns/finread/amihud.pdf
- Amihud, Y. (2020). "Illiquidity and Stock Returns: A Revisit." https://cfr.ivo-welch.info/published/papers/amihud2020illiquidity.pdf
- Odegaard, B.A. "The Amihud Illiquidity Estimator" (lecture notes). https://ba-odegaard.no/teach/notes/liquidity_estimators/amihud_estimator/amihud_lectures.pdf
- Lou, X. and Shu, T. "Why is the Amihud (2002) Illiquidity Measure Priced?" AEA Conference Paper. https://www.aeaweb.org/conference/2015/retrieve.php?pdfid=7476&tk=ESaTTBEz
Disclaimer
This article is educational content only and is not financial advice. Nothing here is a recommendation to buy, sell, or hold any security. Consult a licensed advisor before making investment decisions.