Impermanent Loss: The Cost Hiding Inside Crypto Liquidity Pools
A decentralized exchange pays liquidity providers a share of every trading fee, a pitch that sounds like straightforward passive income. What the pitch usually omits is impermanent loss, a mechanical cost built into how these pools function that can exceed every dollar of fees collected, particularly when the two pooled assets move apart in price.
The core principle
Impermanent loss is the reduction in value a liquidity provider experiences, relative to simply holding the same two assets outside a pool, when the prices of those two pooled assets diverge from the ratio at which they were originally deposited. It arises specifically from how automated market makers (AMMs), the mechanism behind most decentralized crypto exchanges, maintain liquidity: rather than matching individual buyers and sellers the way a traditional order book does, an AMM pool holds two assets and uses a fixed mathematical formula, most commonly x times y = k, where x and y are the quantities of each asset in the pool and k is a constant, to determine prices for every trade against the pool.
As traders buy one asset from the pool and sell the other into it to profit from any price gap between the pool and the broader market, the pool's own holdings automatically rebalance: it sells more of whichever asset is rising in price and accumulates more of whichever asset is falling, precisely the opposite of what a holder who simply kept both assets untouched would experience. This is a structural, mathematical consequence of the constant-product formula, not a bug or a fee, and it happens continuously as arbitrage traders correct the pool's pricing to match the outside market.
The loss is called "impermanent" because if the two assets' prices return to their original ratio before the liquidity provider withdraws, the loss fully reverses and disappears. The name is also, in my experience, the single most misleading piece of terminology in decentralized finance, because for any provider who withdraws while prices remain diverged, which is most providers most of the time, the loss is entirely permanent and entirely real, locked in the moment the liquidity is removed.
How the math works
Example 1: a fifty percent price move. An investor deposits $5,000 of Token A and $5,000 of a stablecoin into a pool, for a total initial deposit of $10,000, at a moment when Token A trades at $50, meaning the deposit represents 100 Token A and 5,000 stablecoins. Token A then doubles to $100. Simply holding the original assets outside the pool would now be worth (100 Token A x $100) + $5,000 stablecoin = $10,000 + $5,000 = $15,000. Inside the constant-product pool, the standard impermanent loss formula for a 2x price change works out to a loss of approximately 5.7% relative to holding, using the formula loss % = [2 x sqrt(price ratio) / (1 + price ratio)] − 1, where the price ratio here is 2. Plugging in: 2 x sqrt(2) / (1 + 2) = 2 x 1.4142 / 3 = 2.8284 / 3 = 0.9428, so the pooled position is worth 0.9428 x $15,000 ≈ $14,142, an impermanent loss of roughly $15,000 − $14,142 = $858 compared to simply holding, before counting any trading fees earned along the way.
Example 2: comparing the loss against fees collected. Continuing the scenario above, suppose the pool generates enough trading volume that the liquidity provider earns $400 in accumulated fees over the period the price moved from $50 to $100. Net of fees, the provider's actual result relative to holding is −$858 + $400 = −$458, still a loss, despite the pool having generated real fee income throughout. Only if fee income had exceeded $858 over that period would the pool position have outperformed simply holding both assets untouched, illustrating why a sharp, sustained price divergence can outrun months of accumulated trading fees in a single move.
How it shows up in real portfolios
The risk is largest in pools pairing two assets with low price correlation, a volatile token paired against a stablecoin being the clearest example, since any significant price move in the volatile asset produces meaningful divergence by definition. Pools pairing two assets that tend to move together, two large-cap cryptocurrencies during a broad rally, or two stablecoins pegged to the same currency, experience far smaller impermanent loss because the price ratio between the pooled assets stays closer to constant.
A common real-world pattern involves an investor drawn in by an advertised annual percentage yield on a liquidity pool, often quoted prominently and without adjustment for impermanent loss, only to find months later that the dollar value of their withdrawn position is lower than what simply holding the two original assets would have produced, despite having collected fee income the entire time. The advertised yield describes fee income only; it says nothing about the separate, simultaneous mechanical loss from divergence.
A second scenario, more favorable to the liquidity provider, involves stablecoin-to-stablecoin pools during periods when both assets hold their peg closely; because the price ratio barely moves, impermanent loss stays minimal, and the fees collected represent a much larger share of realized return, which is why experienced participants in this space concentrate much of their activity in correlated-asset pools rather than volatile pairs, when the goal is closer to steady income than speculation.
Some newer AMM designs attempt to reduce impermanent loss directly, using concentrated liquidity ranges that let a provider specify a narrower price band over which their capital is active, or alternative curve formulas better suited to assets expected to trade near a fixed ratio, such as two different stablecoins. These designs can reduce, but do not eliminate, the underlying mechanical exposure to divergence, and they typically introduce their own added complexity, since a concentrated liquidity position stops earning any fees at all if the price moves outside the specified range entirely, trading one risk for another rather than removing risk altogether.
Actionable breakdown
- Before providing liquidity to a pool, check:
- How correlated the two pooled assets' prices have historically been.
- The pool's actual trading volume and realistic fee income.
- The impermanent loss at a plausible range of price divergence scenarios.
- Whether the underlying protocol itself carries smart contract risk.
- Watch for these red flags:
- A prominently advertised APY with no mention of impermanent loss.
- A pool pairing a highly volatile token against a stablecoin.
- Assuming "impermanent" means the loss will reliably reverse.
- Treating liquidity provision as passive income rather than active risk-taking.
- Favor pools of correlated assets when the goal is steady fee income.
- Model worst-case divergence before depositing, not after.
- Treat crypto liquidity provision as a speculative activity, sized accordingly.
A further complication for anyone attempting to evaluate a pool before committing capital is that both future trading volume and future price divergence are unknown at the time of the deposit decision, meaning the entire calculation described above is necessarily a forward-looking estimate built on assumptions, not a guaranteed outcome. A pool that historically generated strong fee income during a calm, range-bound period can behave very differently once a sharp, sustained price move begins, precisely the environment in which impermanent loss grows largest and fee income, ironically, often fails to keep pace, since sustained directional moves can also coincide with reduced two-way trading volume compared to a period of stable, back-and-forth price action.
Common pitfalls
- Assuming the word "impermanent" means the loss reliably reverses, when it becomes fully permanent and realized the moment liquidity is withdrawn while prices remain diverged.
- Evaluating a pool purely on its advertised fee yield without separately modeling the mechanical impermanent loss that the same price action generates.
- Underestimating how quickly a single sharp price move can outweigh months of accumulated trading fees, as shown in the worked example above.
- Providing liquidity to volatile, uncorrelated asset pairs while expecting the stability of a savings-like return.
Related concepts
For the broader digital asset category this concept applies within, see the guide on crypto. For the underlying asset class whose price swings drive the loss, see cryptocurrency. For the general concept of return demanded for taking on unusual risk, see illiquidity premium, a related but distinct compensation-for-risk idea.
The bottom line
Impermanent loss is a mathematical certainty whenever pooled assets diverge in price, so any liquidity pool decision needs to weigh that mechanical cost against fee income directly, not evaluate the advertised yield in isolation.