What is Liquidity-Adjusted CAPM?

Introduction

liquidity-Adjusted CAPMis an extension of the standard Capital Asset Pricing Model that asks a simple question:what if the relevant return is not just what a security earns, but what an investor keeps after paying to trade it? That question matters because real portfolios are built in markets where trading is not free, immediacy is scarce, and liquidity often gets worse exactly when investors most want to exit. Standard CAPM abstracts from that friction. Liquidity-Adjusted CAPM puts it back in.

The puzzle is easy to state. Two assets can have the same exposure to market risk, yet investors may still demand a higher expected return from one of them because it is harder to trade, more expensive to unwind, or more likely to become illiquid in bad states of the world. If that is true, then a model that prices only covariance with the market portfolio is incomplete. The point of Liquidity-Adjusted CAPM is to explain how level of liquidityandrisk in liquidity enter expected returns alongside ordinary market risk.

The core idea is this: an asset is less attractive not only when it tends to fall in market downturns, but also when it becomes costly to trade exactly when funding is tight, market liquidity dries up, or investors need cash. That mechanism creates compensation for both expected trading costs and exposure to systematic liquidity shocks. Once you see that, the model is less a strange alternative to CAPM than a correction to something CAPM deliberately ignored.

Why does standard CAPM ignore trading costs and liquidity?

Standard CAPM is built around a clean tradeoff. Investors care about expected return and the covariance of an asset with the market portfolio. In equilibrium, assets that add more nondiversifiable market risk must offer higher expected returns. That logic is powerful, but it assumes that returns are frictionless: if you own an asset, you can buy or sell it without meaningful cost, delay, or price impact.

In actual markets, that assumption is often false. A stock with a wide bid-ask spread, a bond that can only be sold by accepting a large concession, or a currency position that becomes hard to unwind during stress imposes an extra cost on the holder. Even if the asset’s cash flows are unchanged, the investor’s net payoff is lower because some value is lost in trading. And that loss is not merely a one-time nuisance. If liquidity conditions move with the business cycle or with market stress, the cost of exiting the asset is state-dependent. Investors therefore care not only about average trading cost but also about whether liquidity deteriorates in bad times.

This is the first thing that makes Liquidity-Adjusted CAPM click: the object being priced is not just the asset’s payoff stream, but the payoff stream after subtracting trading friction. Once the relevant payoff changes, the relevant risk exposures change too.

That intuition had important predecessors. Amihud and Mendelson showed that illiquid assets, measured for example by wider bid-ask spreads, tend to offer higher expected returns, and that this relation is concave rather than one-for-one. Their mechanism emphasized that investors with longer holding periods can better tolerate trading costs, creating a clientele effect in which less liquid assets are naturally held by longer-horizon investors. Amihud later introduced a practical illiquidity proxy, ILLIQ, based on the ratio of absolute return to dollar trading volume, making long-run empirical liquidity studies feasible even when quote-level microstructure data are unavailable. Liquidity-Adjusted CAPM builds on that tradition but goes further: it does not treat liquidity only as a static characteristic; it treats liquidity risk as a systematic pricing force.

How does Liquidity‑Adjusted CAPM incorporate liquidity alongside market risk?

A good way to think about the model is to separate two channels that standard CAPM merges into one frictionless return.

The first channel is the familiar one: market risk. If an asset tends to have low returns when the market is doing poorly, investors dislike it and require compensation.

The second channel is liquidity. If holding the asset means you should expect to pay more to trade it, that expected cost lowers your effective return and must be offset by a higher gross expected return. Further, if the asset becomes especially illiquid when the market is down or when aggregate market liquidity is poor, investors dislike it for the same reason they dislike downside market beta: it hurts when marginal utility of wealth is high.

So the model says that expected returns rise for assets that are costly to trade on average and for assets whose liquidity is especially bad in bad states. In plain language, investors demand compensation for owning positions that are unpleasant to unwind when conditions are already strained.

This distinction between the levelof illiquidity andcovariation in illiquidity matters. A small-cap stock may be hard to trade on average. That is one reason its expected return might be higher. But another stock might be reasonably liquid in normal times and still be risky if its liquidity collapses exactly when the market crashes. Liquidity-Adjusted CAPM tries to price both features.

What are the core equations and liquidity betas in Liquidity‑Adjusted CAPM?

The formal move in Liquidity-Adjusted CAPM is conceptually simple. Let r denote the asset’s gross return and c denote the illiquidity cost associated with trading it. The investor cares about something like a net return, r - c, rather than just r.

That small subtraction changes the pricing equation in an important way. In ordinary CAPM, expected excess return depends on covariance with the market return. In the liquidity-adjusted version, expected excess return also reflects expected illiquidity cost and covariances involving both returns and liquidity. The exact empirical implementation can differ, but the theoretical structure says an asset can command a higher expected return because:

• its expected trading cost is higher,
• its return is low when market liquidity is poor,
• its own liquidity worsens when the market return is poor,
• or its liquidity worsens when overall market liquidity worsens.

Those are not arbitrary add-ons. They all come from pricing the net payoff of a traded asset in a world where trading conditions are random and correlated across securities.

A useful way to summarize the mechanism is this. Standard CAPM prices covariance between your payoff and the representative investor’s wealth. Liquidity-Adjusted CAPM prices covariance between yourtradablepayoff and the investor’s state, where tradability itself is stochastic. That is the conceptual upgrade.

Some presentations emphasize four liquidity-related betas. The exact naming varies, but the economic content is stable. One beta captures ordinary market risk. The others capture whether the asset’s return and its liquidity co-move with market return and market liquidity. Rather than memorizing labels, it is better to remember the signs investors dislike: low returns in bad markets, high trading costs in bad markets, and especially the combination of both.

Example: how liquidity shocks change expected returns for two similar stocks

Imagine two stocks, A and B. Both have the same standard market beta, so under ordinary CAPM they should have the same expected return. But stock A trades with a narrow spread and deep order flow even during moderate stress. Stock B usually looks fine in calm markets, but when volatility rises and investors rush to de-risk, its spread widens sharply and selling pressure moves the price a lot.

Now think like a portfolio manager who may need to meet redemptions in a downturn. If the market falls and you need cash, selling A is unpleasant but manageable. Selling B is worse in two ways at once. First, B’s price is likely already down with the market. Second, converting B into cash requires paying a larger liquidity cost precisely at that moment. The economic damage is therefore amplified: you lose on mark-to-market value and again on execution.

That double hit is exactly what Liquidity-Adjusted CAPM is trying to capture. Even if A and B have the same market beta in the usual sense, B exposes you to worse states because its tradability deteriorates when you care most about tradability. Investors therefore require a higher expected return to hold B.

This example also explains why liquidity risk often appears during crises rather than in average days. In normal times, measured liquidity shocks may be small and liquidity betas can be hard to estimate precisely. In stress episodes, by contrast, common liquidity dries up, spreads widen, funding becomes scarcer, and the model’s mechanism becomes much more visible.

How do researchers measure liquidity for pricing and tests?

The theory is elegant. The practical difficulty is that liquidity is not one thing.

At the trading level, liquidity can mean a narrow bid-ask spread, low price impact, high depth, fast execution, or resilience after an order imbalance. Different markets reveal these dimensions differently. Equity researchers may use quoted or effective spreads, return reversals, or price-impact measures. Bond researchers may focus on dealer intermediation costs and whether trades are inventory-based or agency-based. In foreign exchange, high-frequency measures are often needed to capture commonality across currency pairs.

Because rich microstructure data are not always available, empirical work often relies on proxies. A classic example is Amihud’s ILLIQ, defined as the average ratio of absolute daily return to daily dollar volume. Intuitively, this asks how much price tends to move per dollar traded. It is a coarse measure, but it is practical and makes long historical samples possible.

Another influential approach comes from Pástor and Stambaugh, who construct an aggregate market liquidity factor from return-volume dynamics and then estimate each asset’s liquidity beta, meaning its sensitivity to innovations in market-wide liquidity. Their later review stresses an important implementation point: liquidity shocks are rare and asymmetric, so liquidity betas are noisy, especially in calm periods. Precision improves when one uses longer windows, portfolio sorts, or samples containing large liquidity events.

This measurement problem is not a side issue. It goes to the heart of why empirical tests of Liquidity-Adjusted CAPM are mixed. The model may be conceptually right while any given liquidity proxy is only an imperfect signal of the friction investors actually care about.

What does the empirical evidence show about liquidity‑adjusted pricing?

The broad empirical message is stronger for a simple claim than for the full model. The simple claim is that illiquid assets tend to earn higher average returns. On that point, the literature is fairly persuasive. Amihud found that less liquid NYSE stocks had higher expected returns in the cross-section, and that expected market illiquidity helped predict ex ante stock excess returns over time. Unexpected increases in illiquidity were associated with lower contemporaneous returns, which fits the idea that sudden liquidity deterioration depresses prices immediately.

Amihud and Mendelson found that bid-ask spreads help explain expected stock returns beyond market beta, and that the return-spread relation is concave. That is important because it suggests liquidity matters in equilibrium, but not in a mechanically linear way. Investors with different holding periods absorb liquidity costs differently, so the market does not price each extra unit of spread identically.

For the full Acharya-Pedersen Liquidity-Adjusted CAPM, the evidence is more nuanced. Supportive studies find that the model explains the cross-section of returns better than standard CAPM in some equity samples and that CAPM-like relations work better for returns net of illiquidity costs. The model performs reasonably on portfolios sorted by liquidity, liquidity variation, and size. But the same work also reports only weak evidence that liquidity risk adds strong explanatory power beyond market risk and the level of liquidity, and it fails to explain the book-to-market effect.

Replication and out-of-sample tests make the picture even more careful. One study reviewing U.S. and Japanese evidence reports only limited support: the weaker one-variable test succeeds in a minority of U.S. regressions, the stronger theory-imposed test succeeds rarely, and the model fails in the Japanese sample considered there. That does not mean liquidity is irrelevant. It means the strict version of the model, with its specific cross-sectional restrictions, is empirically harder to confirm than the general proposition that liquidity matters for returns.

This is a common pattern in asset pricing. The basic economic force can be real while the exact parsimonious model is only an approximation.

Why does liquidity co‑move across assets and become a priced risk factor?

A natural objection is that many liquidity problems look idiosyncratic. A single small stock can be thinly traded for firm-specific reasons. Why should liquidity earn a risk premium if diversification can wash that out?

The answer is commonality. Liquidity often co-moves across assets. During calm periods, many securities are easy to trade at once. During stress periods, many become hard to trade at once. That common component is what makes liquidity risk systematic rather than merely asset-specific.

Research in multiple markets supports this point. Pástor and Stambaugh treat innovations in aggregate stock market liquidity as a priced state variable. FX studies using high-frequency data find strong liquidity commonality across currency pairs and report that liquidity risk is heavily priced in currency returns. Work on funding liquidity provides a mechanism for why this commonality appears: when dealer capital is constrained, margins rise, inventories become costly, and market liquidity deteriorates across many assets simultaneously.

Brunnermeier and Pedersen provide the cleanest intuition here. Market liquidity and funding liquidity can reinforce each other. If dealers or leveraged investors suffer losses, their funding constraints tighten. That reduces their willingness to absorb order flow, which worsens market liquidity. Worse market liquidity then increases losses or margins further, tightening funding again. This spiral is not itself the same as Liquidity-Adjusted CAPM, but it explains why liquidity shocks can become broad, state-dependent, and especially severe in downturns. In other words, it explains where the systematic part of liquidity risk can come from.

How should practitioners use Liquidity‑Adjusted CAPM in portfolio construction and risk budgeting?

In portfolio work, Liquidity-Adjusted CAPM is less often used as a single plug-and-play formula than as a way of organizing thought around expected return, implementation cost, and stress behavior.

If you are comparing assets or strategies with similar conventional betas, the model suggests asking three practical questions. What is the expected cost of trading this position? How does that cost change when market conditions deteriorate? And does the position become hard to exit precisely when the rest of the portfolio is already under pressure? These are portfolio construction questions, not just academic ones.

That perspective affects both security selection and risk budgeting. A strategy that looks attractive on gross expected alpha may look much worse once expected turnover costs and crisis-period illiquidity are incorporated. Similarly, two portfolios with the same expected volatility can differ sharply in liquidity drawdown risk if one relies on assets whose trading costs spike in market stress.

The idea also matters across asset classes. In corporate bonds, March 2020 showed that quoted liquidity and actual immediacy can separate sharply when dealers become unwilling to use balance sheet. In Treasuries during the same period, price dislocations reflected a mix of liquidity stress, demand imbalance, and balance-sheet constraints. In FX, systematic liquidity fell sharply during the global financial crisis and liquidity risk appeared priced in returns. These examples differ in market design, but they all fit the same principle: a portfolio is exposed not just to changes in value, but to changes in the terms on which value can be converted into cash.

What are the main limitations and assumptions behind Liquidity‑Adjusted CAPM?

The most important limitation is measurement. Liquidity is multidimensional, and no single proxy perfectly captures what an investor experiences when trading under stress. A bid-ask spread captures one friction. Price impact captures another. Delay, depth, inventory capacity, and funding conditions matter too. The empirical success or failure of the model partly depends on whether the chosen measure is aligned with the true economic cost.

A second limitation is estimation noise. Liquidity betas are hard to estimate because major liquidity shocks are infrequent. In normal periods, measured innovations in aggregate liquidity may be small, which makes regression slopes unstable. This is why portfolio-level tests often look cleaner than single-security tests, and why crisis samples can make liquidity effects appear much stronger.

A third limitation is that the boundary between liquidity levelandliquidity risk is not always clean in the data. Small firms, for example, are often less liquid, so size can partly proxy for liquidity. Some apparent anomalies may shrink once liquidity measures are included. But this does not mean all anomalies are liquidity in disguise. The book-to-market effect, for example, is not well explained by the basic Liquidity-Adjusted CAPM specification discussed in the evidence above.

There is also a conceptual limitation. CAPM-style models assume a representative-investor equilibrium structure that is useful but stylized. Real markets contain heterogeneous horizons, leverage constraints, mandates, dealer intermediation frictions, and institutional features that can make liquidity premia time-varying and nonlinear. In stress episodes, these nonlinearities matter a great deal. The model gives a disciplined framework, but not a complete crisis microstructure theory.

How does Liquidity‑Adjusted CAPM extend and relate to standard CAPM?

It is best to think of Liquidity-Adjusted CAPM not as a rejection of CAPM’s core logic, but as an extension of it.

CAPM says investors require compensation for bearing systematic risk because they care about payoffs in bad states. Liquidity-Adjusted CAPM says the same logic applies once you notice that bad states involve not only lower prices, but also worse tradability. So the extension keeps the equilibrium intuition of CAPM and changes the payoff being priced from gross return to net return.

That is why the model often feels intuitive to practitioners even when its empirical details are debated. Most investors already know that “paper return” and “realizable return” differ. The model simply insists that asset pricing should start from realizable return.

Conclusion

Liquidity-Adjusted CAPM exists because investors do not live in frictionless markets. They care about what an asset earns, what it costs to trade, and whether those trading conditions collapse in bad times. The model therefore adds two linked ideas to standard CAPM: expected illiquidity is a cost that must be compensated, and exposure to systematic liquidity shocks is a risk that can earn a premium.

The cleanest takeaway to remember is this: an asset can be risky not only because its price falls with the market, but because its tradability fails when the market is stressed. That is the problem Liquidity-Adjusted CAPM was built to solve, and it remains the reason the model is useful even where its empirical fit is imperfect.