You deposit two tokens into a liquidity pool, collect trading fees for weeks, then withdraw and find you have less money than if you had simply held the two tokens in your wallet. That gap is impermanent loss, and it is the single most misunderstood risk in decentralized finance.
This guide explains impermanent loss in plain English: why it happens, exactly how big it gets as prices move, and the hard data on how often it wipes out the fees you were chasing. If you want to provide liquidity without walking into that trap, it pays to learn the mechanics inside a structured crypto trading course that covers DeFi risk before you commit capital.
- Impermanent loss is the gap between providing liquidity and just holding — it appears whenever the two pooled prices diverge.
- The loss is fixed by math: a 2× price move costs 5.7%, a 4× move costs 20% — regardless of which token moves.
- A landmark study found 49.5% of Uniswap v3 liquidity providers lost money once impermanent loss was netted against fees.
- Stablecoin and correlated pairs carry near-zero impermanent loss; volatile pairs carry the most.
- “Impermanent” is misleading — the loss becomes permanent the moment you withdraw at a changed price ratio.
What is impermanent loss?
Impermanent loss is the difference in value between depositing tokens in an automated market maker liquidity pool and simply holding those same tokens in your wallet. It appears when the prices of the two pooled assets move apart. The bigger the divergence, the larger the loss — and it exists even while your position is earning fees.
It is called “impermanent” because the loss shrinks back to zero if prices return to where you deposited. In practice they rarely do, so most liquidity providers realize the loss when they withdraw. A better mental model is divergence loss: the cost of holding a portfolio that a robot constantly rebalances against you.
Why does impermanent loss happen?
The cause is the pricing formula itself. Most decentralized exchanges run on a constant-product automated market maker, where the two token balances in a pool must always satisfy x × y = k. That single rule sets the price, and it forces the pool to trade against every move.
Here is the catch: when one token's market price rises, arbitrage traders buy it out of your pool until the pool price matches the outside market. The pool sells the winner and accumulates more of the loser to keep x × y = k intact. You end up holding more of the asset that fell and less of the asset that rose — the exact opposite of what you would want.
The traders doing this are not doing you a favor. They are capturing the price gap between your pool and the wider market as near risk-free profit, and their gain is funded by the rebalancing that leaves you holding more of the weaker asset. In effect, liquidity providers subsidize arbitrageurs in return for the fees that ordinary swappers pay into the pool.
A holder captures 100% of a token's upside. A liquidity provider captures only part of it, because the AMM has been selling that token the whole way up. That structural drag is impermanent loss, and no amount of clever timing removes it — it is baked into the constant-product curve. The same pooling mechanics power the DEXs covered in our guide to how centralized and decentralized exchanges differ.
How much does impermanent loss actually cost?
The loss is not a vague risk — it is a precise function of how far prices diverge. For a standard 50/50 pool, impermanent loss equals 2 × √r / (1 + r) − 1, where r is the ratio of the new price to the old price. Plug in the numbers and the schedule looks like this.
Impermanent loss vs holding, by price move of one asset
Source: Uniswap v2 documentation, “Understanding Returns”; Speedrun Ethereum, 2026. Constant-product 50/50 pool.
Two features matter here. First, the loss is symmetric: a token doubling (2×) and a token halving (0.5×) both cost you 5.7%, because impermanent loss depends on the ratio of the move, not its direction. Second, it accelerates — a 5× move costs 25.5%, more than a quarter of your capital versus holding.
What this means for you: if you expect one asset in a pair to run hard, providing liquidity with it is a bet against your own thesis. You would keep more by simply holding the token you think will rise.
A worked example: a $10,000 ETH/USDC position
Say you deposit $10,000 split evenly — 2 ETH at $2,500 plus $5,000 in USDC. If ETH then doubles to $5,000, arbitrage rebalances the pool until you hold roughly 1.41 ETH and about $7,071 in USDC, worth around $14,142 in total.
Had you simply held the original 2 ETH and $5,000 in cash, you would be sitting on $10,000 of ETH plus $5,000 — $15,000. That $858 gap is your impermanent loss, and it is exactly the 5.7% the formula predicts for a 2× move. Every fee you earned while the price ran is what you hope will close it.
Flip the scenario and the symmetry holds. If ETH instead halved to $1,250, the pool would rebalance you into more ETH and less cash, and you would again trail a simple holder by 5.7%. Up or down, the pool leans you the wrong way — that is the price of quoting a two-sided market around the clock.
Do trading fees make up for impermanent loss?
This is the question that decides whether liquidity provision is worth it, and the honest answer is often no. Fees are your compensation for taking on impermanent loss. In many pools, they do not cover it.
The most-cited evidence comes from a November 2021 study by Topaz Blue and the Bancor Protocol, which analyzed more than 17,000 wallets across 17 Uniswap v3 pools — about 43% of the protocol's total value locked at the time.
Source: Topaz Blue & Bancor Protocol, 2021 (17 Uniswap v3 pools, ~43% of TVL, 5 May–20 September 2021).
Read those numbers slowly. Across the sample, impermanent loss of roughly $260M outran $199M in fees, and 80% of the pools studied lost more to divergence than they earned. Nearly half of all providers would have been richer doing nothing but holding.
Uniswap v3 made this sharper still. By letting providers concentrate capital in a narrow price band, it lifts fee income while the price stays in range — but it magnifies impermanent loss the instant price leaves that band. Research on v3 by Bancor and IntoTheBlock found concentrated positions carried materially higher divergence risk than the old full-range model. Higher yield, higher loss: the trade-off never vanishes, it only changes shape.
When impermanent loss becomes permanent
The word “impermanent” lulls people into thinking the loss cannot hurt them. It can. The loss only reverses if prices return exactly to your entry ratio. The instant you withdraw at any other ratio, the paper loss is realized — permanently. Add gas costs and a mispriced exit, and “impermanent” becomes a very real line on your balance sheet.
Which liquidity pools have the least impermanent loss?
Because impermanent loss is driven entirely by price divergence, the safest pools are the ones whose two assets barely move apart. That single principle ranks every pool you will ever see.
| Pool type | Example | Price divergence | Impermanent loss risk |
|---|---|---|---|
| Stablecoin – stablecoin | USDC / USDT | Almost none | Near-zero |
| Correlated assets | ETH / stETH | Very low | Minimal |
| Two blue-chips | ETH / BTC | Moderate | Moderate |
| Blue-chip – volatile alt | ETH / new token | High | High |
| Stable – volatile alt | USDC / micro-cap | Very high | Highest |
Source: Phemex Academy, 2026; constant-product AMM math.
The trade-off is real: the pools with the lowest impermanent loss also tend to pay the lowest fees, because they attract the least volatile trading. Stablecoin pools are the closest thing DeFi has to a low-risk yield — and even they carry smart-contract and depeg risk that has nothing to do with impermanent loss.
There is no free lunch hiding in this table. Chasing the fat fees of a volatile alt-coin pool means accepting exactly the divergence that produced the losses in the Uniswap v3 data above. Matching the pool to your goal — steady income versus directional exposure — matters far more than the headline yield figure any dashboard flashes at you.
How to reduce impermanent loss
You cannot delete impermanent loss from a constant-product pool, but you can manage how much of it you are exposed to. The practical levers:
- Favor correlated or stablecoin pairs when your goal is yield, not directional exposure — USDC/USDT and ETH/stETH keep divergence tiny.
- Never pool an asset you are bullish on. If you expect it to run, hold it; pooling forces the AMM to sell it into strength.
- Check fees against the loss schedule. Compare the pool's fee APR to the impermanent loss implied by its historical volatility before you deposit.
- Treat concentrated liquidity as active work. Uniswap v3-style ranges can multiply impermanent loss when price leaves your band, so they reward tight, hands-on management, not set-and-forget.
- Do not lean on “IL protection.” Some protocols once offered programs (Bancor v2.1 built toward full coverage after ~100 days) that absorb the loss from reserves, but those have been paused under market stress. Protection is only as solid as the treasury behind it.
If your real aim is predictable on-chain income rather than market-making, compare pooling against validator rewards in our breakdown of how crypto staking yield actually works, and read what DeFi is and how its building blocks fit together before you supply capital to any pool.
Context matters too. Even after a sharp cooldown, DeFi held roughly $72 billion in total value locked in 2026, and decentralized exchanges processed on the order of $4.2 trillion in trailing-12-month volume — so impermanent loss is not a niche edge case. It is quietly reshaping real returns for a large, active pool of capital.
Frequently asked questions
Trading and providing liquidity involve substantial risk of loss and are not suitable for every investor. Crypto assets are highly volatile and regulation varies by country. This article is educational content, not investment advice.
.jpg)
.jpg)
