Illiquid Stocks Must Pay You More: Amihud and Mendelson's Liquidity Premium

Microstructure and asset pricing used to live in separate buildings. Asset pricing people worried about expected returns, risk, and betas. Microstructure people worried about spreads, dealers and order flow, and were politely regarded as plumbers.

Amihud and Mendelson knocked down the wall in 1986 with an argument so simple it is almost annoying: if trading a stock costs you money, you will not hold it unless it pays you extra. Therefore the bid-ask spread, a microstructure quantity, must show up in expected returns, an asset pricing quantity.

The problem: the return you earn is not the return you keep

Every asset pricing model up to that point quietly assumed you could trade for free. The CAPM says a stock's expected return depends on its market risk, full stop. Nowhere does it ask what it costs to get in and out.

But it does cost something. You buy at the ask and sell at the bid. On an illiquid small-cap with a wide spread, that round trip might cost several percent of your money. On a mega-cap it might cost a few basis points. That is not a rounding error, it is a wedge between the return the stock generates and the return you actually take home.

So the question becomes: how does the market price that wedge?

The key idea via analogy: the house with a huge estate agent's fee

Imagine two houses, identical in every way: same street, same rent, same condition. The only difference is that selling House A costs you one percent in fees, while selling House B costs you ten percent.

Nobody would pay the same price for both. House B has to be cheaper to compensate you for the exit cost. And a cheaper price for the same rental income means a higher yield. That is the entire theory in one image: illiquid assets trade at a discount, and a discount is just another word for a higher expected return.

But Amihud and Mendelson did something cleverer than stopping there, and this is the part worth understanding.

Who ends up owning House B? Not a house-flipper. Someone who plans to move every year would be crucified by the ten percent exit fee. House B is bought by someone who plans to live there for thirty years, because they only pay the fee once and can amortize it over a long holding period.

This is the paper's real insight, and it is a beautiful one. Investors have different expected holding periods, and the market sorts them.

• Short-horizon investors, who churn, gravitate to liquid assets where the toll is small.
• Long-horizon investors, who buy and forget, are the natural owners of illiquid assets, because they can spread the cost over many years.

This sorting is what economists call clientele formation, and it has a striking consequence. Because the illiquid assets end up in the hands of patient investors who suffer the spread least, the required compensation for illiquidity does not rise in a straight line with the spread. It rises, but at a decreasing rate. Doubling the spread does not double the required extra return, because the asset simply migrates to a clientele that cares less. This concave shape is the paper's signature theoretical prediction, and Amihud and Mendelson found it in the data.

Why it mattered

• It made liquidity a priced characteristic, not a nuisance. Before this paper, the spread was a friction to be minimized. After it, the spread was a source of expected return, and holding illiquid assets became a legitimate, harvestable strategy, if you have the horizon for it. The modern "illiquidity premium" that private equity, private credit and small-cap funds all invoke is a direct descendant.
• It handed microstructure a seat at the asset pricing table. The paper is the reason "liquidity" appears in serious cross-sectional return studies at all, and it set the stage for Amihud's own 2002 illiquidity measure and for Pastor and Stambaugh's liquidity risk factor.
• It reframed corporate finance decisions. If illiquidity raises a firm's cost of capital, then a company has a direct financial incentive to make its own stock easier to trade: list on a better exchange, tighten the spread, improve disclosure. That claim, "liquidity is worth real money to the issuer," changed how firms think about market structure, and Amihud and Mendelson pushed it hard in later work.
• It gave patient capital a coherent edge. The clientele argument is the cleanest theoretical justification for long-horizon investing that microstructure has produced: your edge is not that you are smarter, it is that the toll booth charges you less per year than it charges the person next to you.

The honest limitations

• Untangling liquidity from size is genuinely hard. Illiquid stocks are overwhelmingly small stocks, and small stocks were already known to have higher average returns. Separating "the liquidity premium" from "the size premium" is a statistical nightmare, and the follow-up literature has argued about it for four decades without full resolution. Some of the original effect may simply be the size effect wearing a different hat.
• The spread is only one dimension of liquidity. A stock can have a tight quoted spread and still be a disaster to trade if the quote is only good for 100 shares. Depth, resilience, and price impact all matter, and the paper's single-number view of liquidity misses them. Later measures (Amihud 2002, Kyle's lambda, Pastor and Stambaugh) exist precisely because "the spread" is too crude.
• The model treats liquidity as fixed and known. In reality, liquidity changes, and it changes worst exactly when you need it most, evaporating in crises. That is a risk, not just a cost, and this paper does not price it. Pastor and Stambaugh's contribution was to argue that the risk of liquidity drying up is itself a priced factor, which is a different and arguably deeper claim.
• Trading costs have collapsed since 1986. Spreads in liquid US equities are now a tiny fraction of what they were when the paper was written. The mechanism is intact, but the magnitudes in the original tests belong to a vanished world, and much of the modern action has moved to genuinely illiquid corners: small caps, corporate bonds, private markets.

The one-line takeaway

Amihud and Mendelson argued that assets that are costly to trade must offer higher returns to compensate, and that because patient investors naturally end up owning the illiquid stuff, the required compensation grows with the spread but at a decreasing rate.